LIST OF SYMBOLS AND ABBREVIATIONS Symbols B Viscous friction coefficient, N-sec/m e, Back emf, v o Steady state error Applied force, N ooh Opposing force offered by mass of the body,N Opposing force offered by the elasticity of the body, N heh Opposing force offered by the friction of thé body, N Sampling frequency, Hz Conductance, mho zo ‘Transformation or operator i, Armature current, A Field current, A Moment of inertia, ke-m*/rad Complex operator Stiffhess of the spring, N-m / rad Acceleration error constant Back emf constant, V / (rad/sec) Derivative constant or gain Integral constant or gain Gain Margin Motor gain constant RRR RR RR Proportional gain Torque constant, N-m/A nK ‘Torque constant, Nm / A nm Velocity error constant re Armature inductance, Hait cs ERs = z o oS wo - wi” 4 ae 3A AAAS Field inductance, H Mass, kg Maximum overshoot Resonant peak Order of the system Type number Pole of a system Pole of compensator Forward path gain of K* forward path Charge Armature resistance, Q Field resistance, Q Complex variable Dominant pole Applied torque, N-m Electrical time constant Opposing torque due to friction, N-m Delay time Derivative time Field time constant Integral time Opposing torque due to moment of inertia, N-m Opposing torque due to elasticity; N-m Mechanical time constant Peak time Rise time Settling time Normalized bandwidthd6 dt ae dt oS 2 Normalized resonant frequency Armature voltage, V Field voltage, V Displacement, m Zero of a system Zeto of compensator Angular displacement, rad Angular velocity, rad/sec Angular acceleration, rad/sec? Undamped natural frequency, rad/sec Damping ratio Resonant frequency Bandwidth Phase margin Phase crossover frequency Gain crossover frequency Flux, weber Comer frequency Damped frequeney of oscillation Phase angle Frequency of maximum phase lag/lead Maximum lag/lead angle Additional phase lead Angle of asymptotes Angle of departure Angle of arrivalatv o Eigen value Impulse train Standard Input/Output signals c(t) otk) et) (kT) (kT). 1(t) r(k) u(t) a) Response in time domain Response of discrete signal Esror signal : Digital error signal Digital control signal Input in time domain Discrete time input signal Control signal (Analog) Impulse signal Matrices and Vectors A AX System matrix State transition matrix of discrete system Input matrix Output matrix Transmission matrix State transition matrix Identity matrix Jordan matrix Modal matrix or diagonalization matrix Transformation matrix Composite matrix for controllability Composite matrix for observability Input vectorUk) Input vector of discrete time system v Vander monde matrix X(t) State variable vector Xx, Initial condition vector x(k) State vector of discrete time system Y(t) Output vector ¥(k) Output vector of discrete time system a Grammian matrix Transform Operators and Functions ACs) Auxiliary polynomial E(s) Error signal in s-domain Gs) Open loop transfer function G(s)H(s) Loop transfer function H(s) Feedback transfer function £ Laplace transform L Inverse Laplace transform M(s) Closed loop transfer function T(s) Transfer function of the system z Z-transform x Inverse Z-transform Abbreviations BIBO Bounded Input Bounded Output LDS Linear Discrete Time System LTL Linear Time Invariant System ROC Region of convergence ZOH Zero Order HoldCHAPTER 1 MATHEMATICAL MODELS OF CONTROL SYSTEM 1.1 CONTROL SYSEEM Control system thedry evolved as an engineering discipline and due to universality of the principles involved, it is extended to various fields like economy, sociology, biology, medicine, etc. Control theory has played a vital role in the advance of engineering and science. The automatic control has become an integral part of modern manufacturing and industrial processes, For example, numerical control of machine tools in manufacturing industries, controlling pressure, temperature, humidity, viscosity and flow in process industry. When a number of elements or components are connected in a sequence to perform a specific function, the group thus formed is called a system. Ina system when the output quantity is controlled by varying the input quantity, the system is called control system. The output quantity is called-controlied variable or response and input quantity is called command signal or excitation. OPEN LOOP SYSTEM Any physical system which does not automatically correct the variation in its output, is called an open loop system, or control system in which the output quantity has no effect upon the input quantity are called open-loop control system. This means that the output is not fedback to the input for correction. Input_,} Open loop | Output oO 7 system (Plant) rc Fig 1.1 : Open loop system In open loop system the output can be varied by varying the input. But due to external disturbances the system output may change. When the output changes due to disturbances, it is not followed by changes in input to correct the output. In open loop systems the changes in output are corrected by changing the input manually. CLOSED LOOP SYSTEM Controt systems in which the output has an effect upon the input quantity in order to maintain the desired output value are called closed loop systems, Error Detector omtemt c > up Reference _s{Connolier] p| Open loop system| : Input — lant) 1) o Fig 1.2 : Closed loop system.12 The open loop system can be modified as closed loop system by providing a feedback. The provision of feedback automatically corrects the changes in output due to disturbances. Hence the closed loop. system is also called automatic control system. The general block diagram of an automatic control system is shown in fig 1.2. It consists of an error detector, a controller, plant (open loop system) and feedback path elements. The reference signal ( or input signal ) corresponds to desired output. The feedback path elements samples the output and converts it to a signal of same type as that of reference signal. The feedback signal is proportional to output signal and it is fed to the error detector. The error signal generated by the error detector is the difference between reference signal and feedback signal. The controller modifies and amplifies the error signal to produce better control action. The modified error signal is fed to the plant to. correct its output. Advantages of open loop systems 1. The open loop systems are simple and economical. 2. The open loop systems are easier to construct. 3. Generally the open loop systems are stable. Disadvantages of open loop systems 1. The open loop systems are inaccurate and unreliable. 2. The changes in the output due to external disturbances are not corrected automatically. Advantages of closed loop systems 1. The closed loop systems are accurate. 2. The closed loop systems are accurate even in the presence of non-linearities. 3. The sensitivity of the systems may be made smal] to make the system more stable, 4. The closed loop systems are less affected by noise. Disadvantages of closed loop systems 1. The closed loop systems are complex and costly. 2. The feedback in closed loop system may lead to oscillatory response. 3. The feedback reduces the overall gain of the system. 4, Stability is a major problem in closed loop system and more care is needed to design a stable closed loop system. 1.2. EXAMPLES OF CONTROL SYSTEMS TEMPERATURE CONTROL SYSTEM EXAMPLE 1 ‘OPEN. LOOP SYSTEM The electric fumace shown in fig 1.3. is an open loop system. The outputin the systems the desired temperature. The cemperature of the system is raised by heat generated by the heating element. The output temperature depends on the time during which the supply to heater remains ON. ‘The ON and OF F of the supply is governed by the time setting of the relay. The temperature is measured by a sensor, which gives an analog voltage corresponding to the temperature of the furnace. The analog signal is converted to digital signal yan Analog -to - Digital converter (A/D converter).43 C0: Sensor 1 —S | Eleetic (7 AD | Digital urmace Converter; — Interface |-—* Display Lp I ot I 1 | Relay >—? | controi | AC supply Circuit} Heating element Fig 1.3 : Open loop temperature control system. ‘The digital signals given to the digital display device to display the temperature.-In this system ifthere is any changein output temperature then the time setting of the relay is notaltered automatically. CLOSED LOOP SYSTEM ‘The electric furnace shown in fig 1.4 is @ closed loop system. The output of the systems the desired temperature and it depends on the time during which the supply to heater remains ON. Sensor i ' , Digital contro! Electric AD | circuit Fumacelt converter —»memace | iContratier (or) ey | | | CompuierMicro re — x processor 1 ing element f Rel: | ‘lay c 5 Control fe Amplifier te Dia le |__ Reterence input Circuit (Converter) (Desired temperatura) ac Supply Fig 1.4 : Closed loop temperature control system. ‘The switching ON and OFF ofthe relay is controlled by a controller which i a digital system or computer. The desired temperature is input to the system through keyboard or asa signal corresponding to desired temperature via ports. The actual ‘temperature is sensed by sensor and converted to digital signal by the A/D converter. The computer reads the actual temperature and compares with desired temperature. Ifitfinds any difference then it sends signal to switch ON or OFF the relay through D/A converierand amplifier, Thus the system automatically corrects any changes in output. Hence itis a closed loop system EXAMPLE 2 : TRAFFIC CONTROL SYSTEM OPEN LOOP SYSTEM Traffic control by means of traffic signals operated on a time basis constitutes an open-loop control system. The sequence of control signals are based on a time slot given for each signal. The time slots are decided based on a traffic study ‘The system will not measure the density of the traffic before giving the signals. Since the time slot does not changes accordin; to traffic density, the system is open loop system. CLOSED LOOP SYSTEM ‘Traific control system can be made as a closed loop system if the time slots of the signals are decided based on th density of traffic. in closed loop traffic contro! system, the density of the trafficis measured on all the sides and the informatio is fed toa computer . The timings of the control signals are decided by the computer based on the density of traffic . Since th closed loop system dynamically changes the timings, the flow of vehicles will be better than open loop system.Ghapter (Mathematical Models of Control. L4 EXAMPLE 3 : NUMERICAL CONTROL SYSTEM OPEN LOOP SYSTEM Numerical control is a method of controlling the motion of machine components using numbers. Here, the position of work head tool is controlled by the binary information contained in a disk. Se 1 Output Reader >| DIA |—>| Amplifier} —»} Seve | 5 Cutter > FM pulse J motor | "| | Cool ¥ position) { ‘ - Magnetic disk Fig 1.5 : Open loop numerical control system. Amagnetic disk is prepared in binary form representing the desired part P (P is the metal partto be machined). The too! will operate on the desired part P. To start the system, the disk is fed through the reader to the D/A converter. The D/A converter Converts the FM (frequency modulated) output of the reader to a analog signal. Itis amplified and fed to servometer which Positions the cutter on the desired part P. The position of the cutter head is controlled by the angular motion of the servometer. This is an open loop system since no feedback path exists between the output and input. The system positions the too! for a given input command. Any deviation in the desired position is not checked and corrected automatically. CLOSED LOOP SYSTEM Amagnetic disks prepared in binary form representing the desirad partP (Ps the metal partto be machined). To start the system, the disk is loaded in the reader. The controller compares tho frequency modulated input pulse signai with the feedback pulse signal. The controlleris a computer or microprocassor system. The controller carries outmathematical operations on the difference in the pulse signals and generates an error signel, The D/A converter converts the controller output pulse (ertor signal) into an analog signal . The amplified analog signal rotates the servomotar to position the tool onthe job. The position of the cutterhead is controlled according to the input of the servomotor. The transducer attached to the cutterhead converts the motion into an electrical signal. The analog electrical signal is converted to the digital pulse signal by the A/D converter. Then this signal is compared with the input pulse signal. ifthere is any difference between these two, the controller sends a signal to the servomotor to reduce it. Thus the system automatically corrects any deviation in the desired output tool position. An advantage of numerical control is that complex parts can be produced with uniform tolerances atthe maximum milling speed. Tachometer —— Output (Tool | position) 1 i | L OA Servo | | pc Eur? Conttoller —* converter» Amplifier /—o| ere >a pulse ! Magnetic © | ak | Feedback signat [AyD Feedback signal L_“emheck sina | (Digital) (Analog) Fig 1.6 : Closed loop numerical control system.LS EXAMPLE 4: POSITION CONTROL SYSTEM USING SERVOMOTOR “The position control system shown in fig 1.7is @ closed loop system. The syste consists of a servorotor powered by a generator. The load whose position haste be controlled is connected to motor shaft through gear wheels, Potentiometers are used to convert the mechanical motion to electrical signals. The desired load position (0) is set on the input potentiometer and the actual load position (@,) Is fed to feedback potentiometer. The difference between the two angular positions generates an error signal, which is amplified and fed to generator field circuit. The induced emf of the generator drives the motor. The rotation ofthe motor stops when the error signalis zero, e. when the desired load position is reached. This type of control systems are called servomechanisms The servoor servomechanismsare feedback control systems in which the output is méchanical position (or time derivatives of position e.g. velocity and acceleration). Enor signal Amplifier [8 Generator field Lot | Input Feedback | potentiometer potentiometer Fig 1.7: A position control system (servomechanism). 1.3 MATHEMATICAL MODELS OF CONTROL SYSTEMS ‘A control system is a collection of physical objects (components) connected together to serve an objective. The input output relations of various physical components ofa system are governed by differential equations. The mathematical model of a control system constitutes a set of differential equations. The response or output of the system can be studied by solving the differential equations for various input conditions. ‘The mathematical model ofa system is linear if it obeys the principle of superposition and homogenity. This principle implies that if a system model has responses y,(t) and y, (t) to any inputs x, (t) and x, (8) respectively, then the system response to the linear combination of these inputs a,x, (f) +a, x, (1) is given by linear combination of the individual outputs a, y,(t)+a, y,(0), where a, and a, are constants. ‘The principle of superposition can be explained diagrammatically as shown in fig, 1.8. roan ) HG | FG Le | ty ha, EO Sysuem system an() +a) op G Leg) Tel 1 Tel) =a.c,(f) + a,0,0) 3 > ext z a : noe ie 1G Foo | Oo ie] then system Gis linear | er a Fig 1.8 : Principle of linearity and superposition.dle T= Matbeomatical Models of Control Systems 16 A mathematical model will be linear if the differential equations describing the system has constant coefficients (or the coefficients may be functions of independent variables). If the coefficients of the differential equation describing the system are constants then the model is /inear time invariant. If the coefficients of differential equations governing the system are functions of time then the model is linear time varying. ‘The differential equations of a linear time invariant system can be reshaped into different form for the convenience of analysis. One such model for single input and single output system analysis is transfer function of the system. The transfer function of a system is defined as the ratio of Laplace transform of output to the Laplace transform of input with zero initial conditions, Laplace Transform of output | | a | Transfer function = PSS “ESO Oo Laplace Transform of input |e eocauacoamom | wa ay The transfer function can be obtained by taking Laplace transform of the differential equations governing the system with zero initial conditions and rearranging the resulting algebraic equations to get the ratio of output to input. 1.4 MECHANICAL TRANSLATIONAL SYSTEMS The model of mechanical translational systems can be obtained by using three basic elements mass, spring and dash-pot. These three elements represents three essential phenomena which occur in various ways in mechanical systems. The weight of the mechanical system is represented by the-element mass and it is assumed to be concentrated at the center of the body. The elastic deformation of the body can be represented by a spring. The friction existing in rotating mechanical system can be represented by the dash-pot. The dash- pot is a piston moving inside a cylinder filled with viscous fluid. When a force is applied to a translational mechanical system, it is opposed by opposing forces due to mass, friction and elasticity of the system. The force acting on a mechanical body are governed by Newton’s second law of motion. For translational systems it states that the sum of forces acting on a body is zero. (or Newton’s second law states that the sum of applied forces is equal to the sum of opposing forces on a body). LIST OF SYMBOLS USED IN MECHANICAL TRANSLATIONAL SYSTEM x = Displacement, m < HW dx . =.= = Velocity, m/sec dt - dv . ; a = G77 ga = Acceleration, misec* Applied fore, N (Newtons) Opposing force offered by mass of the body, N Opposing force offered by the elasticity of the body (spring), N Opposing force offered by the friction of the body (dash - pot), N fi, M= Mass, kg K = Stiffness of spring, N/m B = Viscous friction co-efficient, N-sec/m Note : Lower case letters are functions of time |17 Gantt sjstemsTaginvering ~) FORCE BALANCE EQUATIONS OF IDEALIZED ELEMENTS, Consider an ideal mass element shown in fig 1.9 which has negligible friction and elasticity. Let a force be applied on it. The mass will offer an opposing force which is proportional to acceleration of the body. Let, f = Applied force bx f= Opposing force due to mass g t—+ Mw 4 @x x Reference Here, fc i, -Moo . ores Im ae OF « Fig 1.9: Ideal mass element By Newton's second Jaw. | M— wn 1.2), Consider an ideal frictional element dashpot shown in fig 1.10 which has negligible mass and elasticity . Let a force be applied on it. The dash-pot will offer an opposing force which is proportional to velocity of the body, Le x Let, f = Applied force f, = Opposing force due to friction g Here, <2 or 4, =B% B Reference at at Fig 1.10: Ideal dashpot with | er By Newton's second law,| f= {, = B x | (13) one end faved to reference. ‘When the dashpot has displacement at both ends as shown in Lex, a fig L.L1, the opposing force is proportional to differential velocity. 4 f+ ———_| or fy wad (%; — x2) B Reference we A) Fig 1.11 : Ideal dashpot with displacement at both ends. Consider an ideal elastic element spring shown in fig 1.12, fox which has negligible mass and friction. Let a force be applied on it. The spring will offer ati opposing force which is proportional to f— ——*—_$ displacement of the body, K K e . Reference Let, f = Applied force Fig 1.12; Ideal spring with one end £, = Opposing force due to elasticity fixed to reference. Herefcx or f=Kx By Newton's second law, | f=f,=Kx |... (1.5) When the spring has displacement at both ends as shown in hex, fox g 1.13 thé opposing force is proportional to differential displacement. > 65H —— f, (Xp Xp) or f, = K(x, —x,) K Fig 1.13 : Ideal spring with displacement at both ends. poe _ [f= f, = K@xy=%,)Chapter T= Mathematical Models of Control Systems 18 | e Guidelines to determine the Transfer Function of Mechanical Translational System 1. In mechanical translational system, the differential equations governing the system are obtained by writing force balance equations at nodes in the system. The nodes are meeting point of elements. Generally the nodes are mass elements in the system. In some cases the nodes may be without mass element. 2. The linear displacement of the masses (nodes) are assumed as x, , X,, X,, ete., and assign a displacement to each mass(node) . The first derivative of the displacement is velocity and the second derivative of the displacement is acceleration. 3. Draw the free body diagrams of the system . The free body diagram is obtained by drawing each mass separately and then marking all the forces acting on that mass (node). Always the opposing force acts in a direction opposite to applied force. The mass has to move in the direction of the applied force. Hence the displacement, velocity and acceleration of the mass Will be in'the direction of the applied force. If there is no applied force then the displacement, velocity and acceleration of the mass will be in a direction opposite to that of opposing force. 4. For each free body diagram, write one differential equation by equating the sum of applied forces to the sum of opposing forces. 5... Take Laplace transform of differential equations to convert them to algebraic equations. Then rearrange the s-domain equations to eliminate the unwanted variables and obtain the ratio between output variable and input variable. This ratio is the transfer function of the system. Note: Laplace transform of x(t) = L£{xt)}= Xs) | Laplace transform of - <{g xo} =5X(9) (with zero initial conditions) i s? X(s) (with zero initial conditions) | 2a 3 Laplace transform of £22 {& x0| de? dt EXAMPLE 1.1 \Write the differential equations governing the mechanical system shown in fig 1. and determine the transfer function. box, Lex K, B Cy | | M pM 9 s * Fig I. SOLUTION es” Inthe given system, applied force '{(t)' is the inputand displacement’ is the output. Let, Laplace transform of f(t) = £{f()} = F(s) Laplace transform of x = L(x} = X(s) Lapiace transform of x, = Lix,) =X(s)1g x Hence the required transfer function is s ‘The system has two nodes and they are mass M, and M,. The differential equations governing the system are given by force batance equations at these nodes. Letthe displacement of mass M, be x, The free body diagram of mass M, is shown in fig 2. The opposing forces acting ‘on mass M, are marked as f,., fy. fy fy and f, fon» fas ax, - dx, . bx, f= M SG BR Lm ins : a ion §=BL(%—x); & =Kits-x) , By Newton's second law, M, f for thr th Hath = tf }J¢——f, 1M, BO 6 BL 6 a Koc, + Ke) = 0 OM Ge Brg TPG PRO AI Fig 2: Free body diagram Ontaking Laplace transform of above equation with zero intial conditions we get, "288M, (nade 1) Mg?X,(s) + B,SX;,(8) + Bs DXi(s) — X(s)] + KX As) +K [X(s)=X(s)} = 0 X(s) [Mss? + (B, + B)s +(K, + K)]- X(s) [Bs +K]=0 Xs) (Ms? + (B, +B)s + (K, + K)] = X(s) (Bs +k] Bs +K . = wee 2X10 ° X00) Goes RaR (1) The free body diagram of mass M, is shown in fig 3. The opposing forces acting on M, are marked as fp, ff, and. ax ft px fho=M, 5%; y= 8, & at ‘at ——_ 1) Q i br8a OR) = KK Xs) fee M, +— By Newton's second law, le 4, foe he + +f =) Kf, Fig 3 : Free body diagram 2, } My Shea, Beod (=x) K(x =x) Ft) of mass M, (node 2) j On aking Laplace transform of above equation with zero initial conditions we get, M,8°X(s) +B,8X(s) +Bs[X(s) — X,(s)] + KPX(s) - X,(s)] = F(s) X{(6) [MS* + (Bz + B)s + K]- X,(s)[Bs +K] = F(s) ‘Subsfituting for X,(s) from equation (1) in equation (2) we get, 24 (8, + ee X(s) [M,s* + (B, +B)s+K]— X(s) Mss (B Bern f°)‘© Chapier I ilatbeniatical Models of Conirol Systems ~) 10 [M,s? +(B, +B)s +K} IMs? +(B,+B)s+(K, +K)]- (Bs +K)? xe) | * Mis? + (B48) 9 =(K+K) | Fe) X(s) i “ Fis) M (B,+B)s+(K, +1) + K)] [Mys* + (Bp + B)s +K]- (Bs + K)? (By+B)s+( RESULT The differential equations governing the systern are, ax, a ua +B) SBS 6-9) ++ Ke, —9) <0 1 Pee ee ee) ‘The transfer function of the systemis, X(8) Ms? +(B, +B) $+(K,+K) Fis) [Ms?+(B, +B) s+(K,+1)] [M,s* +(B, +B) 8+K]- (Bs +K)* EXAMPLE 1.2 YA) Determine the transfer function “2) ofthe ‘system shown in fo 1 Fs) SOLUTION Let, Laplace transform of f(t) = Lif(t)} = F(s) Laplace transform of y, = Lfy,} =Y,(s) Laplace transform of y, = Lfy,} =,(s) ‘The system has two nodes and they are mass M, and M,, The differential equations governing the system are the force balance equationsat these nodes. The free body diagram of mass M, is shovm in fig 2. The opposing forces are marked asf, fy f,, andf,, a g fr MM: ha BM: Hyman ¢ fe Heb ve) By Newton's second law, fay +, +f the =A) a 8 1M + BE Ky Kal, Ya) =f) nl) Fig 2. On taking Laplace transform of equation (1) with zero initial condition we get, Mys?¥\(8) + BS¥4(5) + K,¥4() + Ke1¥4(8)— Yo(S)] = F(S) YASIMs* + Bs +(K; +Kz)]—Yo(s)K2 = Fis) 2) The free body diagram of mass M, is shown in fig 3. The opposing forces acting on M, are f,, and. d fog = Mo % ; Kol¥e-Ys) By Newton's second law, fxo + fig =0 feLit d Oe +Koly2-y) =0 few On taking Laplace transform of abs ati \n taking Laplace transform of above equation we get, ee M,s*¥2(8) + Ka[Yo(s)— Y46)] = 0 M. | . Yo{s) [M,s” +K,]- Ys) Ky = 0 fea ¥48) = Ys) Mae Substituting for Y,(s) from equation (3) in equation (2) we get, 2 vats) ea] [Ms® Bs (K,+K,)]- Yo(s) Kp = FOS) s ve) Mi tae 2a Sa) Fie) 2 ee [Mis? + Bs +(K, +K,)] [Mys? +K,]—K RESULT ‘The differential equations governing the system are, a 1 MGB 3h + Kiyh + Kali ~ Yo) = (0) e 2 MSP + Kye ¥9)= 0 The transfer function of the systemis, YAS) _ Ke Fis) [Ms?+Bs+ (K,+K,)][M,s* +K,]-K? EXAMPLE 1.3 Determine the transfer function, 2) ang X2lS) for the system shown in fig 4 F(s) F(s) 4) fl Poe A bp y +800 M+ 41 M, “BES 4 J K, Le Soo } S4E 5! B, t Fig 1. rs SOLUTION ph) fy Let, Laplace transform of ft) = Ltilt)} = F(s) Mbt, Fig2. Laplace transform of x, = £{x,} = X,(s) fre Laplace transform of x, = £{x,} = Xs)hapter I - Mathematical Models of Control Systems 112 ‘The system fias two nodes and they are mass M, and M,. The differential equations governing the system are the force balance equations at these nodes. The free body diagrarn ofmmass M, is shown in fig 2. The opposing forces are marked a5, fy fayp and f,, dx, fos MS ha By Newton's second law, fri + for + foto + fer = ft) +B, 24g, SO) kx = At a at (On taking Laplace transform of above equation with zero initial conditions we get, Ms°X,(s) + Bs Xi(s) + Bias Ki(s) ~ X(8)] + K.X,¢s) = F(s) X,(6) MS? +(By + By) S+Ky] ~BySXa( F(s) won) The free body diagram of mass M, is shown in fig 3. The opposing forces are marked a F., oy fy. and fp Faas hs ax. Oxy ‘ha = M, : a la = Ma fo=Be d fio = Broa eX); ho = Kx 2 = Bioa (Ka—m) fe = Kok By Newton's second law, fra + foo +for2 + fg =0 dx, Be. oi MoGe 7 Be Gy t Be On taking Laplace transform of equation (2) with zero initial conditions we get, MS°Xa{8) + B,SXa(8) + By28 [Xp(S) — X,(8)] + KyX;(s) = 0 0 We =X) x =0 wel) X,(s) [M,s? + (Bz +Bz2) $+ Kz] ~ Biz 8 X15) 2 X,(s) [M,8* + (By + By) $+ Ka] = Biz $ X,(s) = Bs Xi) (3) *O)~ Ts? (Bp +Bg) 84K] ‘ ‘Substituting for Xs(s) from equation (3) in equation (1) we get, (B,8)° Xi(s) M,s? +(B, +By)S+K, X(6) [IMs? + (8, + Bra) 8+ Ki Mp8? + (Be + Bra) 8 +Kal~(Bre8)?] Fs) M,8? + (By +Byg) $+ Ky Xi(6) IMs? +(B,+ By) 8K; =Fis) - ls) _ M,s? +(B, +B.) 8+K, SF)” Ms? +(8,+8,,) 5+ 1] IM,s? +(B, +B.) 8+K,]-(8, 8)" From equation (3) we get, _ IMs? +(B, +By) sts) X18) a Bys Substituing for X,{s) from equation (4) in equation (1) we get, Xi(s) 2a) MS BeBe) S21 ye? (6,0) 84K ]-B 8 Xl) =F) .153 sag tagineving [M82 +B, + Brg) S+K_] Ms? +(B)+By) s+ Bg8)"] pp) ESS 1 Xol8) _ Bys -— Fle) Mas” + (By + By) 8+ Ka] [Ms* +(B, +B) s*Ki)- (Bs) X2(8) RESULT = Thedifferential equations governing the system are, ox, dx, x, — Xo) SB SB, Se Ky = SE 4B BEBE Kin, ax = x1) at 1M, @x, ax, 2 Magee Boge t Ba + KX = 0 ‘The transfer functions of the system are; XS) _ Ms? +(By +Byz) $+ Kp 1. F(s) IMS? + (B; + By) 8+ Kil [M.S + (B, +B) $+Kal- Bus)" 3, XelS) _ Buys F(S) (Mas? + (By + Biz) § * Ko) IM" + (By + Bz) §* Kil - (Bus) EXAMPLE 1.4 Write the equations of mation in s-domain for the system shown in fig 1. Determine the transfer function of the system. _ Lex) 4 a { f M (t) j 8S 1 >i) 5 77777 * Fig 1. SOLUTION Let, Laplace transform of x(t)= L(x} =X(s) Laplace transform of f(t) = £&tt)} = FOS) Letx, be the displacement atthe meeting point of spring and dashpot. Laplace transform ofx, is X,(s). The system has two nodes and they are mass M and. the meeting point of spring and dashpot “The differential equations governing the systomare the force balance equations atthese nodes. The equations of motion in the s-domain are obtained by taking Laplace transform ofthe differential equations. “The free body diagram of mass Mis shown in fig 2. The opposing forces are marked asf, f,,and fy. ax. dx d fee MGz | f= Brae 3 fe=8e ae) L»x By Newton's second law the force balance equation is, }— ft) fhe =H ae ax dx qd MS +8, 2B Sux) =f) tS aE Bat Beg x)=) On taking Laplace transform of the above equation we get, Fig 2. Ms? X(s)+B, s X(s) +B, $ [X(s) ~ X()] = Fs) [Ms? +(B, +B.) s] X(s) ~ Bz 8X) = Fis) Li The free booy diagram atthe meeting point of spring and dashpotis shown in fig 3. The opposing forces are marked as fand f,. fen BrS ts | 1 By Newton's second law, f, +f, =0 M=0 f bp—te ABS -aeKm= Fig 3. ‘Ontaking Leplace transform of the above equation we get, B, s [X\(s)—X(s)]+K X(s)=0 (By8+K) X{s)~-By 8 X/s) =0 nee) B28 Xi0)= SRE Xie) sl) Substituting for X,(s) from equation (2) in equation (1) we get, 2 =| X(s)=Fis) [Ms? +(6,+B,)s]X(s)-B,s (IMs? =@,+B,)s16, s+K)- (8, Pl x eee Bys+K Xs) _ B,s+K _ FS) Ms*+(6,+B,) 3] (6 s+K) ee ‘The differential equations governing the system are, a ,& 5 ¢ 1 Me eget Ba gem = d 2. Bam =x)+Kx,=0 The equations of motion in s-domain are, 1. [Ms*+(B, +B.) s] X(s)— By S Xs) = F(s) 2. (By $+K) X,(s)-B, s X(s)=0 ‘The transfer function of the systemis, Xs) _ Bysik Fis) (Ms? ~(B,+B,) s}(B)s+1)- (62 s* 1.5 MECHANICAL ROTATIONAL SYSTEMS The model of rotational mechanical systems can be obtained by using three elements, moment of inertia [I] of mass, dash-pot with rotational frictional coefficient [B] and torsional spring with stiffness [K]. The weight of the rotational mechanical system is represented by the moment of inertia of the ass, The moment of inertia of the system or body is considered to be concentrated at the centre of ity of the body. The elastic deformation of the body can be represented by a spring (torsional spring). The friction existing in rotational mechanical system can be represented by the dash-pot. The dash-pot is 2 piston rotating inside a cylinder filled with viscous fluid.LB CGoritrol Systems Engineering When a torque is applied to a rotational mechanical system, it is opposed by opposing torques due to moment of inertia, friction and elasticity of the system. The torques acting on a rotational mechanical body are governed by Newton's second Yaw of motion for rotational systems. It states that the sum of torques acting on a body is zero (or Newton's law states that the sum of applied torques is equal to the sum of opposing torques on a body) UST OF SYMBOLS USED IN MECHANICAL ROTATIONAL SYSTEM © = Angular displacement, rad dO. a = Angular velocity, rad/sec Angular acceleration, rad/sec? = Applied torque, N-m Moment of inertia, Kg-me/rad = Rotational frictional coefficient, N-m/(rad/sec) = Stiffness of the spring, N-m/rad TORQUE BALANCE EQUATIONS OF IDEALISED ELEMENTS Consider an ideal mass element shown in fig 1.14 which has negligible friction and elasticity. The opposing torque due to moment of inertia is proportional to the arigular acceleration. Let, T = Applied torque. T,= Opposing torque due to moment of inertia of the body. T 9 Consider an ideal frictional element dash pot shown in fig 1.15 which has negligible moment of inertia and elasticity, Let a torque be applied on it. The dash pot will offer an opposing torque which is proportional to the angular velocity of the body. Let, T =Applied torque. T, =Opposing torque due to friction. YA | 8 @ . eo T,«<—— T, =B_ B Beg oF B=Bo 79 ® Fig 1.15 : Ideal rotational dash-pot with By Newton's second law, a) os) one end fixed to reference. | When the dash pot has angular displacement at both ends as shown in fig 1.16, the opposing torque is proportional to the differential angular velocity. “ . ls 2 Fig 1.16 : Ideal dash-pot with (1.9) angular displacement at both ends. d d T#£,-9) oF T,=BS-O,-0) fe(Ghabler I =Mathematical Models of Control Systems > 416 Consider an ideal elastic element, torsional spring as shown in fig 1.17, which has negligible moment of inertia and friction. Let a torque be applied on it. The torsional spring will offer an opposing torque which is proportional to angular displacement of the body. Let, T = Applied torque. \ T, = Opposing torque due to elasticity. 76 C& T,<8 or 1,=K@ Fig 1.17 : Ideal spring with one —r (10) end fixed to reference. By Newton's second law, [T % When the spring has angular displacement at both ends as shown in fig 1.18 the opposing torque is proportional to differential angular displacement. \ > -8,) or T,=K(6,-6) T™ 6! KO, a veo L1) Fig 1.18 : Ideal spring with angular 2) displacement at both ends. Guidelines to determine the Transfer Function of Mechanical Rotational System 1. Inmechanical rotational system, the differential equations governing the system are obtained by writing torque balance equations at nodes in the system. The nodes are meeting point of elements. Generally the nodes are mass elements with moment of inertia in the system. In some cases the nodes may be without mass element. 2. The angular displacement of the moment of inertia of the masses (nodes) are assumed as 0,, €,, 8,, etc., and assign a displacement to each mass (node). The first derivative of angular displacement is angular velocity and the second derivative of the angular displacement is angular acceleration. 3. Draw the free body diagrams of the system. The free body diagram is obtained by drawing each moment of inertia of mass separately and then marking all the torques acting on that hody. Always the opposing torques acts in a direction opposite to applied torque. 4. The mass has to rotate in the direction of the applied torque. Hence the angular displacement, velocity and acceleration of the mass will be in the direction of the applied torque. If there is no applied torque then the angular displacement, velocity and acceleration of the mass is in a direction opposite to that of opposing torque. 5. For each free body diagram write one differential equation by equating the sum of applied torques to the sum of opposing torques. 6. Take Laplace transform of differential equation to convert them to algebraic equations. Then rearrange the s-domain equations to eliminate the unwanted variables and obtain the relation between output variable and input variable. This ratio is the transfer function of the system. Note = Laplace transform of 6 = £{0} = %s) Laplace transform of s = 42 = 6(s) (with zero initial conditions) i 2, 2 Laplace transform of % = eh =8°O(s) . (with zero initial conditions) ii itLIT EXAMPLE 1.5 Write the differential equations governing the mechanical rotational systom shown in fig 1. Obtain the transfer function ofthe system. abe Be {Applied Torque) (Output) Fig 1. SOLUTION Inthe given system, applied torque Tis the input and angular displacement is the output, Let, Laplace transform of T = £{T}=T(s) Laplace transform of 0 = {0} =0(s) Laplace transform of 9, = £{0,}= 0,6) Hence the required transfer functionis #3 The system has two nodes and they are masses with moment of inertia J, and J,. The differential equations governing the system are given by torque balance equations at these nodes. Letthe angular displacement of mass vith momentofinertia J, be8, The ree body diagram of J is shown in ig 2. The opposing torques acting on J, are marked as T,, and, Tastee 5 Te=K(@,-9) By Newton's second law, T,.+ 7, =T hh Po oo JANN 14 K(0,-8)=T ve] T 4, 3,28 Ke, -KO=T (1) Fig 2+ Free body diagram of mass with ote moment of inertia J, On taking Laplace transform of equation (1) with zero initial conditions we get, J, 8? 4,(3} +K9,(s)—Ka(s) = T(s) (J, 8? + K) 0,(s)— K 6(s) = T(s) (2) The free body diagram of mass with moment of inertia J, is shown in fig 3. The opposing torques acting on J, aremarked as T, T,andT,. ao ce Tp=teaz i: T=BI i Tk=K0-6; e at 5 T= K0-0) — By Newton's secondlaw, T+, +1, =0 —— 4 “ os 7 a . de oe sage Bay tKe-a)=0 Fig 3 : Free body diagram of mass with moment of inertia J. 2 nie 8B sk KO, = On taking Laplace transform of above equation with zera initial conditions we gat, J,870(s) + Bs O(s) +KO(S)—K8,(s)= 0Chapter t ‘Mathematical Models of Control Systems > 118 (Jz 8? + Bs +K) €(8)—K0,¢s) = 0 (8? +Bs+K) 4 as) = Se (8) ‘Substituting for 6,(s) from equation (3) in equation (2) we get, (i,8? + Bs +) (Js? +K) z ‘8(s) — Ka(s) = T(s) Tus? + 2 _K2 [es K) is? +Bs +K)—K Jos)= L K 28) —K T(s) (Js? +K) zs? + Bs + K) —K? RESULT The differential equations governing the system are, de, 1 Jy a? -Ko=T 2, 2. bE Be Ke Ko, =0 The transfer function of he systems, oS) _ __K F(8) (ds? +K) (Us + Bs +K)—K? EXAMPLE 1.6 Write the differential equations governing the mechanical rotational system shown in fig 1. and determine the transfer function o(s)T(s). SOLUTION Inthe given system, the torque T is the input andthe angular displacement is the output. Let, Laplace transform of T= £{T} =T(s) Laptace transform ofé = £10} =6(s) 19,3 =06) Ss) Hence the required transfer function is ——— Te) . The system has two nodes and they are masses with moment of inertia J, and J, The differential equations goveming system are given by torque balance equations at these nodes. Laplace transform of 8, = Lethe angular displacement of mass with moment of inertia J, be 6,. The free body diagram of J,isshowninfig2. The soposing torques acting on J, are marked as T,,, Ty. and T, mre de, a Th th a 2 Toe BrgGer-8) 3 Te =K(0,-6) aol 1 BSS By Newton's second law, T,+ Ty + =T Fig a : Phe bea dag diagram of mass with moment of inertia J,ao, d #8 go We Beg ‘Ontaking Laplace transform of above equation with zero initial conditions we get, (8,-8)+K(0,-0)=T J870,(s)+8 B;, [0,(s) -Os)] + K0,(s) - KO(s) = Tis) 0,8) US? + SB ie +K] 918) [88,2 +K] = TS) a ‘The free body diagram of mass with moment of inert J is shown in ig3. The opposing torques are marked as Ty, Ts T,andT,. ve Tp Tae Te T d I, ee Tow Bz Gy @- 6) ) VARA we J) TBR; K=KO-0) 3 Fig 3 : Free body diagram of mass with Tet T= 0 moment of inertia J. By Newton's second law, Tj + Toa ao d de de Ge 7 Be OO) + BG + KO 0) =0 a FBS + Bie, +B)+Ke KO =0 3S ate at Ontaking Laplace transform of above equation with zero initial conditions we get, J,870(S) - B,80,(s) + $0(S) [B,2 + B] + K0(s)— Ko,(s) = 0 (8) [s7J, +8(B,, +B) +K]~0,(s) [88,, +K}=0 _ [s4uy+8(B +B) +KI 9 0 Be KI as) (2) ‘Substituting for@,(s) from equation (2) in equation (1) we get, 2 + ie? +s8,,+1q HS +882 +B) M08) gg. (68,,+K) Kk) @(s) = T(s) (Js? + SByz +K) [J287 + 8(Brz + B) + K]- (SByz + KY? - SSB + K) ame foe) TS) 86) (sBy +k) “T(s) (dis? + SBi2 +K) Ns” + 8(B,, + B)+K]- (6B,z +1)? RESULT ‘The ditterential equations governing the system are, / oe, d 1 Sage * Br yy O19) + K(0 1-8) = T ae do, _ do 2 JS 8, SSB, +8)= K-98) <0 arr at at (2 FB) + KO) The transfer function of the systemis, 9) _ (682 +K) (8) (Uys? + 8B + K) [W,8" + §(B,2 +B) + K]—(sBiy +K)?‘ehapier T= Mathematical Models of Control Systems 1.20 1.6 ELECTRICAL SYSTEMS The models of electrical systems can be obtained by using resistor, capacitor and inductor. The current-voltage relation of resistor, inductor and capacitor are given in table-1. For modelling electrical stems, the electrical network or equivalent circuit is formed by using R, L and C and voltage or current urce. The differential equations governing the electrical systems can be formed by writing Kirchoff ’s current law equations by choosing various nodes in the network or Kirchoff’s voltage law equations by josing various closed paths in the network. The transfer function can be obtained by taking Laplace ‘cansform of the differential equations and rearranging them as a ratio of output to input. TABLE-1.1 : [- Relati and € {Element Voltage, across the element Current through the element © oR v)=Ri) (= ~o We x) WL =L4 wet 1H v(t) =L- a iy pivoa vit) i - c _ w= Efi at (t= ch v(t) =XAMPLE 1.7 Obtain the transfer function of the electrical network shown in fig 1. SOLUTION In the given network, inputis e(t) and output is v,(t). Let, Laplace transform of a(t) = £fe()} = Eis) Laplace transform of v,(t)= Ltv,(0) =V,(s) The transfer function of the network is ae Transform the voltage source in series with resistance R, into equivalent current source as shown in figure 2. The network has two nodes. 1e node voltages be v, and v,. The Laplace transform of node voltages and v, are V,(S) and V,(s) respectively. The differential equations governing Senetwork are given by the Kirchof’s current law equations atthese nodes. Atnods-1, by Kirchoff’s current law (referfig3) Wood Mew 8 Mic Mi wee & R dat RR, Ontaking Laplace transform of above equation with zero initial conditions we get, Me v/s R, 4. “se. _ Xefs) +CsV4s)+ vis welt +80, + aL eLal Ciara Sten rggearing tae aT Joo dv Fig 4. ‘At node-2, by Kirchaff's current iaw (refer fig 4) Yo-Vy GMa “241.0, S220 dt Ontaking Laplace transform of above equation with zero initial conditions we get, YOM cs veo[+ “ee Ri127 TABLE -1.3 : Analogous Quantities in Force-Voltage Analogy | Item Mechanical system Electrical system (mesh basis system) Independent variable Force, f Voltage, ¢, V (input) Dependent variable Velocity, v Current, i (ouput) Displacement, x Charge, q Dissipative element Frictional coefficient Resistance, R of dashpot, B Storage element Mass, M Inductance, L Stiffness of spring, K Inverse of capacitance, 1/C Physical law Newton's second law Kirchoff’s voltage law =rF=0 iv=0 Changing the level of Transformer independent variable aN & TABLE. Analogous Elements Force-Curtent Analogy Mechanical system Electrical system Input: Force Input: Current source Output: Velocity Output : Voltage across the element -—-— i¢ Rov Y~ - 109,< 1.28 TABLE-L5 : Analogous Quantities in Force-Current Analogy Item Mechanical system Electrical system (node basis system) Independent variable Force, f Current, i (input) Dependent variable Velocity, v Voltage, v (output) Displacement, x Flux, > Dissipative element Frictional coefficient Conductance G=1/R of dashpot, B Storage element Mass, M Capacitance, C Stiffness of spring, K Inverse of inductance, 1/L Physical law Newton's second law Kirchoff’s current law =f=0 Li=0 Changing the level of Lever Transformer independent variable fo Ny as nN 5. The mechanical driving sources (force) and passive elements connected to the node (mass) in mechanical system should be represented by analogous elements in a closed loop in analogous electrical system. 6. The element connected between two (nodes) masses in mechanical system is represented as a common element between two meshes in electrical analogous system. FORCE-CURRENT ANALOGY ‘The force balance equations of mechanical elements and their analogous electrical elements in force-current analogy are shown in table-1.4. The table-1.5. shows the list of analogous quantities in force-current analogy. The following points serve as guidelines to obtain electrical analogous of mechanical systems based on force-current analogy. 1. In electrical systems elements in parallel will have same voltage, likewise in mechanical systems, the elements having same force are said to be in parallel. The elements having same velocity in mechanical system should have the same analogous voltage in electrical analogous system. Each node (meeting point of elements) in the mechanical system corresponds to a node in electrical system. A mass is considered as a node. 4. The number of nodes in electrical analogous is same as that of the number of nodes (masses) in mechanical system. Hence the number of node voltages and system equations will be same as that of the number of velocities of (nodes) masses in mechanical system. 5. The mechanical driving sources (forces) and passive elements connected to the node (mass) in mechanical system should be represented by analogous elements connected to a node in electrical system. 6. The element connected between two nodes (masses) in mechanical system is represented as a common element between two nodes in electrical analogous system.1.29 EXAMPLE 1.8 Write the differential equations governing the mechanical system shown in fig 1. Draw the foice-voltage and force-current ) [a a % electrical analogous circuits and verify by waiting mesh and nade Boj ee equations. : as i SOLUTION 'B, Fig I. The given mechanical system has two nodes (masses). The differential equations governing the mechanical system are given by force balance equations at these nodes. Let the displacements of masses M, and M, be x, and x, respectively. The comesponding velocities be v, and v,. ‘The free body diagram of M, is shown in fg 2. The opposing forces are marked as * foe for fore and fy wc) t d, dx, br, fai My ot re Mf} >t, d }—# fire BeBe G Om) ff =KiK1—%) te By Newton's second law, fay + fy + fire + fy = f(t) Fig 2 ox, dx, d 2 My Ge Br ge * Biz yO Ka) # Klos Xa) = (9) all) The free body diagram of M, is shown in fig 3.The opposing forces are marked 8 fa, fixe fy, aNd fg x ax, ax, d —pv, Se=Me fa= Ban 1 Toe = Bio Oe) pro fa =a - xX) 5 fe =Kox | ” Me > fre By Newton's second law, fxg + tye + fig + fare + Lys, a fe a Ren KGa m=O on (2) a M27 HF Kin ~ x) (2) Fig 3. ‘On replacing the displacements by velocity in the differential equations (1) and (2) of the mechanical system we get, 2, 7 a * ; Sey andx= vat} MSE Biv Brlvy ve) + Kelis ve) a= sve 3) 2 «Bev, 1 Kelve dt+ Bry —v,)+ Kl vy = (4) FORCE-VOLTAGE ANALOGOUS CIRCUIT The given mechanical system has two nodes (masses). Hence the force-voltage analogous electrical circuit will have twomeshes. The force applied to mass, M, is represented by a voltage source in first mesh. The elements M,,B,,K, and B,,are Connected to first node. Hence they are represented by analogous elementin mesh-1 forming a closed path. The elements Ky, B., M,, K, and B, afe connectedto second node. Hence they are represented by analogous elemientin mesh-2 forming a closed path, The elements K, and B,, are common between node-1 and 2 and so they are represented by analogous element as ‘common elements between two meshes. The force-voltage electrical analogous circultis shown in ig 4 UC, yi, ML, B, 3 R, K> 1c, Yok By Ry L, R, SSW 88 Re R + 2 + TC c. Fig 4 : Force-voltage electrical analogous circuit. ] l+ Fig 6. ‘The mesh basis equations using Kirchoff's voltage law for the circuit shown in fig 4 are given below (Refer fig 5 and 6). di. ; 1. Lage Pt Ralf) +e JC tadat= ef ~6) a 1, tig 4 LG Rilo t Elle at+R,, deel iat =0 (6) Itis observed that the mesh basis equations (5) and (6) are similar to the differential equations (3) and (4) goveming the mechanical system FORCE-CURRENT ANALOGOUS CIRCUIT ‘The given mechanical system has two nodes (masses). Hence the force-current analogous electrical circuit will have ‘wo nodes. “The force applied to mass M, is represented as a currentsource connected to node-1 in analogous electrical circuit. The elements M,, B,, K, and B,, are connected to first node. Hence they are represented by analogous elements connected to node-1 in analogous electrical circuit. The elements K,, B,», M,, K,, and B, are connected to second node. Hence they are represented by analogous elements as elements connected to node-2 in analogous electrical circuit. The elements K, and B,, eré common between node-1 and 2 and s0 they are represented by analogous elements as common element between two nodes in analogous circuit. The force-currentelectrical analogous circuitis shown in fig 7. ‘The electrical analogous elements for the elements of mechanical system are given below. Yi) M, 3 C, B,> UR, K, 71, yoy, M, > C, B,> UR, Kota, v,>¥, B, > UR,L3t AT) vat a ~ w+ Live -wiat=0 svnl8) itis observed thatthe node basis equations (7) and (8) are similar to the differential equations (3) and (4) goveming the mechanicalsystem. EXAMPLE 1.9 I Write the differential equations governing the mechanical system shown in fig 1. Ki A B Drawthe force -voltage and force-current alectrical analogous circuits and verify by writing — x mesh and node equations. * K, ‘f £0 SOLUTION t SOLUTION ve The given mechanical system has three nodes masses. The differential equations mo governing the mechanical system are given by force balance equations at these nodes. Let «8 eB, the displacements of masses M,, M, and M, bex,.x, andx respectively. The comesponding a! ys velocities be v,,v, and v,. a The free body diagram of M, is show in fig 2. The opposing forces are marked as Figl. four fae fy aNd fy jx, x, eX, — rt | M «—t, Lk, Fig 4,> 1.32 fa = Kz 0X2) 5 fia Ko By Newton's second law, fn + fyit feat fra = filt) Bx, dx Mae * BG + Keb ~X2)+ Kx = £() wend) Free body diagram of M,is shown in fig 3. The opposing forces are marked asf. fy fi. fs a, ho Me i fe Ke 0X2) 3 fla = Ky Xs) By Newton's second law, fn2 + fs + fe ths = b(t) a, 4, ge 2 79) * Kala — %1 + Kala — a) = elt) (2) “The free body diagram of M, is shown in fig 4. The opposing forces are markedas ff, and... ax. foo MG i hs Fis) 5 be= Kilts By Newton's second law, fs + fue + fig = 0 a a _ Maat + Ba Gp (M9 ~ Xa) + Kala ~ 2) = 0 _ 7” son) ‘Onrepiacing the displacements by velocity in the differential equations (1), (2) and (8) governing the mechanical syste eget, (,, @x_dv. dx . [te wa and x= Jat Me By, +k, =f (4) Ge By Kyat = Kee, valet =F) vs MyM Bylvy v9) +Ke (ve W)Bt+ Kells Vt =f) 6) dvy , M; <2 +Ba(v, V2) +Kaf(vs—vo)dt = 0 at IOLTAGE ANALOGOUS CIRCUIT ‘The given mechanical system has three nodes (masses), Hence the force-voltage analogous electrical cicuitwill have, Stree meshes. The force applied to mass, M, is represented by a voltage source in first mesh and the force applied to mass, M, ' represented by a voltage source in second mesh. The elements M,, B,, K, and K, are connected to firstnode. Henos they are represented by analogous elementin mesh- { forming a closed path. The elements M,, B,, K, and K, are connected to second node. Hence they are represented by analogous element in mesh-2 forming a closed path. The claments M,, K,and B, are connected to third node. Hence they are represented by analogous element in mesh-3 forming a closed path. The element K, is common between node-1 and 2 and so itis represented by analogous element as common element between mesh 1 and 2. The elements K, and B, are common between node-2 and 3 and so they are represented by analogous elements as common olemenis between mesh-2 and 3. The force-voltage electrical analogous circuit is shown in fig 5. The electrical analogous elements for the elements of mechanical system are given below. fre(0 voi, M,>L, B,->R, K,> 1G, £0 0,0) vi, M,> B,>R, K,>1C, vy iy ML, K,> 11,133 Fig 5 : Force-voltage electrical analogous circuit eo L —_—- ON 1 Ce Te ¢ 1 iL | OY Gea Fig 7. Fig 8 ‘The mesh basis equations using Kirchoff’s voltage law for the circuit shown in fig 5 are given below (Refer fig 6, 7, 8). di, 1 ig ug Rare Git SIM iat = ef) J) -8) dis ng fl bs Gt Ralls ie) += I -iy)at = 0 (9) ttis observed that the mesh equations (7), (8) and (9) are similar tothe differential equations (4), (5) and (6) goveming the mechanical system, _ - FORCE-CURRENT ANALOGOUS CIRCUIT The given mechanical system has three nodes (masses). Hence the foree-current analogous electrical circuit willhave three nodes. ‘The force applied to mass M, is represented as a current source connected to node-1 in analogous electrical circuit. The force applied to mass M, is represented as a current source connected to node-2 in analogous électrical circuit. The elements M,, B,, K, and K, are connected to first node. Hence they are represented by analogous elements as elements connected to nade-1 in analogous electrical circuit The elements M,, B,, K, and K, are connected to second node. Hence they are represented by analogous elements as elemants connected to nade-2 in analogous electrical circuit. The elements M,, B, and K, are connected to third node. Hence they are represented by analogous elements as elements connected to node-3 in analogous electrical circuit. The element K, is common between node-1 and2 and so itis represented by analogous element as common element between node-1 and 2 in analogous circuit, The elements B, and K, are common between node-2 and 3 and so they are represented by analogous elements as common elements between node-2 and 3. The forcs-current electrical analogous circuit isshownin fig 9. ‘The electrical analogous elements for the elements of mechanical system are given below. £010) yy, M,C, B, > UR, Ko 1, fi) VV, M,>C, 8,3 1R, Kom, Wy, M,C, KI,Ghapter 1 - Mathematical Models of ‘Control Systems 1A ‘The node basis equations using Kirchoff's current aw for the circuit shown in fig 9. are given below. (Referfig 10, 11,12). = L 8 wok Fig 9 : Force-current electrical analogous circuit. 1 —](v¥,—v2)dt y, a! ve Ue (eee ie <— five —vp)dt Fig 12: (10) 11) dvs 1 1 Og 7 Hvs~ vale 0 +12) itis observed that node basis equations (10), (11) and (12) are similar to the differential equations (4), (6) and (6) jpveming the mechanical system.1.35 EXAMPLE 1-10 Ka Wie the differential equations governing the mechanical system shown ing, Lz} 3 Mol a fig 1.Draw forcs-voltage and force-current electrical analogous circuits and verify by > 3 K, & writing mesh and node equations. TOO M. SOLUTION O I B, ‘The given mechanical system has three nodos (masses). The differential 7 equations governing the mechanical systom are given by force balance equations at Fig 1 these nodes. Let the displacements of masses M, M,and M, bex,, x, and, respectively. * The corresponding velocities be v,,v,and v,, ‘The free body diagram of M, is shown in fig 2. The opposing forces are marked as ffs fy fy ANA, —>x a, ox, for MVS BGE Ba Km i e— Ff, d qd e-— f, fo =By=(Ks—-%2) ha =Bs—(X-X ; be = Ba grli=%2) # fy = Ba G(s %) u, i By Newton's second law, fay t f+ fa fo the = 0 —t1, le—+, ox 5 &, a d +B, PS Kony + BE) = Ba (1-5) = seul) Mapa + Br Get Kom Ba Geer %2) + Ba Gels 5) =0 ( Fig 2 “The free body diagram of M, is shown in fig 3. The opposing forces are marked aS fp, fyas fy fiay- PX ax, d -—>v, Ge fe Babe) — d — fra = Bas O—%9)! fen =KaalXe —%2) M, — By Newton's second law, fg + fad + fs + fezs =O ax, d gd My SoP + Bap ha Ho) Bas Ga ~ Ha) + Kain — Ma) = 0 (2) Fig 3. ‘The free body diagram of M, is shown in fig 4. The opposing forces are marked aS ffs: za» ANd fe. | ?™ Loy tind . fro = Me Ge : d ea =Bas $e) | fas = Koala —%) By Newton's second law, fs + hs +23 + fs =0 | a&& 2 4 d MAGEE GOS IB GOH) Khe va)30 Fig 4. Onreplacing the displacements by velocity in the differential equations (1), (2) and (3) governing the mechanical syster we get, dx dv dx . (ie ee -vanaxe a) MS Bik fut Bele, ve) +Byl¥e—¥4) =0 M.S + By(va v9) +Baal¥a “¥s)+ Kf —vs}at=0 qt Bae ) + Bag ¥s — Va) + Koo fly — Vayet = 0 Ma:GiapieF I Mathematical Models of Control Systems > SPCE-VOLTAGE ANALOGOUS CIRCUIT The given mechanical systein has three nodes (massés). Hence the force-voltage analogous electrical circuitwill have memeshes. 136 The elements M,, K,, B., B, and B, are connected to first node. Hence they are represented by.analogous elements in =st-1 forming a closed path, The elements M., K,,, B,, and B, are connected to second node. Hence they are represented salogous elements in mesh-2 forming a closed path. The elements M,, K,,, 8, and B, are connected to third node. Hence =) 2'e represented by analogous elements in mesh-3 forming a closed path The elements K,, and B,, are common between node-2 and 3 and so they are represented by analogous element as Simon elements between mésh-2 and 3. The element B, is common between node-1 and 2 and soiitis represented by <22G0us element as comimon elementbetween mest-1 and 2. The element, is common between node-1 and 3and soitis erssented by analogous element between mesh-1 and 3. The force-voltage electrical analogous circuits shown in fig 5. ‘The electrical analogous elements for the elements of mechanical system are given below. vi, M,—L, K, S1C, B, +R, vi, Mol, Ky? 1/C,B, > Ry vy, Mol, BR, B,>Ry fw di, Roly Rolly ie) J Rp Oy TO | | i Fig 5 : Force-voltage electrical analogous circuit Fig 6. Ralls ~ hy) ny Rafle i) + a dz Fat Fig 7. | | | a Te. HE Ile-iadot Realy in) GY Sy * uly R, | Co Ry | wy Fig 8. The mesh basis equations using Kirchoff's voltage law for the Circuit shown in fig 5 are given below. (Refer fig 6, 7 u di 1 . 1 z + Ret ee Lit + Pali —in)+ Pelli) = 0 BB + Ral) + Mla) Palle 1) = 0 2 (7)437 CGaaitrol systems Enginooring di 7 Ladheri A+ itis observed that the mesh basis equations (7), (8) and (9) are similar to the differential equations (4), (5) and governing the mechanical system FORCE-CURRENT ANALOGOUS CIRCUIT The given mechanical system has three nodes (masses). Hence the force-current analogous electrical circuft wilh throe nodes. “The elements M,, K,,B,,.B, and B, are connected tofrstnode. Hence they are represented by analogous element elements vonnected to nede-1 in analogous electrical circuit. The elements M,, K,,,B,, and B, are connected to second.no Hence thay are represented by analogous elements as elements connected to node-2 in analogous electrical cit. elements M, K.,.B,, end B, are connected to third node. Hence they re represented by analogous elements 2s elem connected to node-3 in analogous electrical circu. ‘The elements K,, and B., are common between node-2 and 3 and so they are represented by analogous elemen common elements between nodé-2.and 3in electrical analogous circuit. The element B, is common between node-1 and2. soit is represented by analogous element as common element between node- 1 and 2 in electrical analogous circuit lament, js common between node-1 and and soitis represented by analogous elementas common element between n + and3 in electrical analogous circuit. The force-current electrical analogous circuitis shown in fig 9. “The electrical analogous elements for the elements of mechanical system are given below. vy Y M, > K, > tL, 8, > UR, Vv, > Vy M, > Ky UL B, + 1/R, ye Ny M, > G, B, > UR, B,, > UR, 4‘Ghapler T-Matbeinatical Models of Control Systems > 138 ‘The node basis equations using Kirchoff’s current law for the circuit shown in fig 9 are given below. (Refer fig 10, 11 p12). C Siege ivatt x ~waltgethws)=0 . n(10) iy 1 . CG Me Wt LL Ae volt give v) =0 4) it = Ra GE Oe wit ll wade Bote ms) 20 volt2) itis observed that the node basis equations (10), (11) and (12) are similarto the differential equations (4), (5) and (6) jpverning the mechanical system. EXAMPLE 1.11 ‘Write the differential equations governing the mechanical system shown in fig 1. Draw the erce-voltage and force-current electrical analogous circuits and verify by writing mesh and node uations. SOLUTION “The given mechanical system has two nodes (masses). The differential equations governing pe mechanical system are given by force balance equations atthese nodes. Let the displacement. gmasses M, and M, be x, and x, respectively. The corresponding velocities be v, and v,, The free body diagram of M, is shown in fig 2. The opposing forces are marked asf, f,, and f,. de ~X2) 4:28, ” at ee OE : a= Ki = X2) By Newton's second law, fay + +f Ox, dexy— x0) “The free body diagram of M, is shown in fig 3. The opposing forces are marked asf. fa. + Ky(x,— x2) = 0 (1) ax, dx, d fa MP i he BE 5 = Brg) fe=Kx, fa Kia) By Newton's second law, fro fhe + fe + fy tha = F(t) 2 Xe » Fig 3. On replacing the displacements by velocity in the differential equations (1) and (2) goveming the mechanical system dx, a Ba Get Kana + By ea — x0) + Klee =) = AO meget, =O _ vandx= jvdt dt’ dt MSE Bim veel r—valst= 0 8) My S22 a Bvs + Kel vast + Bp) + Ka Ma - wt= A ®139 : CGintral Systems Engineering” FORCE-VOLIAGE ANALOGOUS CIRCUIT The given mechanical system has two nodes (masses). Hence the force voltage analogous electrical circuit will he twomeshes. The force applied to mass, M, is representedby a voltage source in second mesh. The elements i, K, and B, are connected to first node, Hence they are represented by analogous element in mest forming a closed path. The elements M,, K,, 8, B, and K, are connected to second node. Hence they are represented! analogous element in mesh 2 forming a closed path. ‘Theelements B, and K, are common between node 1 and 2 and so they are represented as common elements betwee mesh 1 and 2. The force-voltage electrical analogous circuitis shown in fig 4. The eloctrical analogous clements for the elements of mechanical system are given below. fe) v, > i Moo Ly, K, > 1c, B, > R, vy, > i Mook Ko Me, BOR nok ‘The mesh basis equations using Kirchoff's voltage law for the circuit shown in fig 4. are given below, (refer fig5 and eft) L Fig 4 : Force-voltage electrical analogous circuit. Fig 5. tis observed that the mesh basis equations (5) and (6) are similar to the differential equations (3) and (4) goveming th mechanical system. FORCE-CURRENT ANALOGOUS CIRCUIT ‘The given mechanical system has two nodes (masses). Hence the force-currentanalogous electrical circuit wil hay two nodes. The force applied to mass M, is represented as a current source connected to node-2 in analogous electrical circu The elements M,. K, and B, are connected to first node. Hence they are represented by analogous elements 2 elements connected to node-1 in analogous electrical circuft. The elements M,, K,, B,, B, and K, are connected to secon ‘hode. Hence they are represented by analogous elements as elements connected to node-1 in analogous electrical circuit‘Mathematical Models of Control Systems £40 The elements K, and B, is common to node-1 and 2 and so they are represented by analogous element as common elements between two nodes in analogous circuit. The force-current electrical analogous circuitis shown in fig 7. ‘The electrical analogous elements for the elements of mechanical system are given below. fi) vy, M,C, B, > 1R, Kom, v,>%, M,C, B,> UR, Ko 1, ‘The node basis equations using Kirchoff's current law for the circuitshown in fig.7, are given below, (Refer fig8 and 9). CML y,-v,)+7f.-v)at=0 soe (7) 4 uy . 1 vp CO, 2+ —y, Pat Ry 2 Itis observed thatthe node basis equations (7) and (8) are similar to the differential equations (3) and (4) governing the mechanical system, Jute» Live ude E five — wpa =i) res @) A jon-veat 4 Fig 8. 1.40 ELECTRICAL ANALOGOUS OF MECHANICAL ROTATIONAL SYSTEMS The three basic elements moment of inertia, rotational dashpot and torsional spring that are used in modelling mechanical rotational systems are analogous to resistance, inductance and capacitance of electrical systems: The input torque in mechanical system is analogous to either voltage source or current source in electrical systems. The output angular velocity (first derivative of angular displacement) in mechanical rotational system is analogous to either current or voltage in an element in electrical system. Since the electrical systems has two types of inputs either voltage source or current source, there are two types of analogies: torque-voltage analogy and torque-current analogy. TORQUE-VOLTAGE ANALOGY The torque balance equations of mechanical rotational elements and their analogous electrical elements. in torque-voltage analogy are shown-in table-1.6. The table-1.7 shows the list of analogous quantities “in torque-voltage analogy.La Laat spsiemsingineering Mechanical rotational system Electrical system Input; Torque Input, : Voltage source Output ; Angular velocity Output : Current through the element B 7 hea * gay VARI . on i Ns ae Tepe I~ ae ye The following points serve as guidelines to obtain electrical analogous of mechanical rotational systems based on torque-voltage analogy. In electrical systems the elements in series will have same current, likewise in mechanical systems, the elements having same angular velocity are said to be in series. The elements having same angular velocity in mechanical system should have analogous same current in electrical analogous system. Each node (mécting point of elements) in the mechanical system corresponds to a closed loop in electrical system. The moment of inertia of mass is considered as a node. The number of meshes in ciectrical analogous is same as that of the number of nodes (moment of inertia of mass ) in mechanical system. Hence the number of mesh currents and system equations will be same as that of the number of angular velocities of nodes (moment of inertia of mass) in mechanical system. The mechanical driving sources (Torque) and passive elements connected to the node (moment of inertia of mass) in mechanical system should be represented by analogous element in 3 closed loop in analogous electrical system. \ The element connected between two nodes (moment of inertia) in mechanical system is represented as a common element between two meshes in electrical analogous system.eee I Hiatbomalica Models of Control Systems > 1a TABLE-L7 : Analogous Quantities in ‘Torque-Voltage Analogy Item Mechanical rotational system Electrical system (mesh basis system) Independent variable Torque, T Voltage, 0, v Gnput) Dependent variable | Angular Velocity, «0 Current, i (output) Angular displacement, @ Charge, q Dissipative element Rotational coefficient Resistance, R. of dashpot, B Storage element Moment of inertia, J Inductance, L Stiffness of spring, K Inverse of capacitance, 1/C Physical law ‘Newton's second law Kirchoff’s voltage law IT=0 iv=0 Changing the level of Gear Transformer |. independent variable Tom ea _N | my a TORQUE-CURRENT ANALOGY The torque balance equations of mechanical elements and their analogous electrical clements in sorque-current analogy are shown in table-1.8. The table-1.9 shows the list of analogous quantities in corque-current analogy. The following points serve as guidelines to obtain electrical analogous of mechanical rotational systems based on Torque-current analogy. 1. In electrical systems the elements in parallel will have same voltage, likewise in mechanical systems, the elements having same torque are said to be in parallel. The elements having same angular velocity in mechanical system should have analogous same voltage in electrical analogous system. Each node (meeting point of elements) in the mechanical system corresponds to a node in electrical system. The moment of inertia of mass is considered as a node. ‘The number of nodes in electrical analogous is same as that of the humber of nodes (moment of inertia of mass) in mechanical system. Hence the number of node voltages and system equations will be same as that of the number of angular velocities of nodes (moment of inertia of mass) in mechanical system. The mechanical driving sources (Torque) and passive elements connected to the node in mechanical system should be represented by analogous element connected to a node in analogous electrical system. The element connected between two nodes (moment of inertia of mass) in mechanical system is represented as a common element between two nodes in electrical analogous system.LB Ciara SystemisTngincering ‘TABLE-L8 : Analogous Elements in Torque-Current Analogy Mechanical rotational system Electrical system Input : Torque Output : Angular velocity Input : Current source Output : Voltage across the element K pT ok T=K0=Kjodt fodt 6 \ a 0 do dt Analogous Quantities in Torque-Current Analogy Ttem Mechanical rotational system Electrical system (node basis system) Independent variable Torque, T Current, i (input) Dependent variable Angular Velocity, o Voltage, v (output) Angular displacement, ® Flux, Dissipative clement Rotational frictional coefficient of dashpot, B Conductance, G = 1/R Storage element Moment of inertia, J Capacitance, C Stiffness of spring, K Inverse of inductance, VL | Physical law Newton's second law ” Kirchoff ’s current law ET=0 di-0 Changing the level of Gear Transformer independent variable 1 |(Githabler FMatbematical Models of Conirol Systems ~) 14 XAMPLE 1.12 g Write the differential equations governing the mechanical otational system & ma EB shown in fig 1. Draw the torque-voltage and torque-current electrical analogous 7 g/L“ a cuits and verify by writing mesh and node equations. i > 7 TTT TITTTOT SOLUTION B, Fig. The given mechanical rotational system has two nodes (moment of inertia ofmasses). The differential equations governing the mechanical rotational system are given by torque balance equations at these nodes. Let the angular displacements of J, and J, be @, and 0, respectively. The corresponding angular velocities be ,ando, “The fee body diagram of Ji shown n fg 2. The opposing torques are marked as Typ Ted Ty dO, 8, Tah ET GE Ty KO.) \ FEY} }— ew Ll kK By Newton's second law, T,,+T,,*T=T T @, Tr Ta Ta Fig 2, 0, 0, —_ eget Bg TO 82) T (1) _ The free body diagram of J, is shown in fig 3. The opposing torques are marked as TT, T,,andT,,. oe, oe Type i ane Ge TL \ Ta=Ke ¢ Tan Ke) Hee By Newton's second law, T+ Tp +T p+ Ty, a, 2 Fig 3. Sir + Be G2 + Ka) + Ki(02 - 0) =9 ‘On replacing the angular displacements by angular velocity inthe differential equations (1) and (2) ‘governing the mechenical rotational system we get, (,. da _da_ 0 e, 22, Bie and 0=Joat| ae dt’ at J J BiB. +KJ(o;-o,)dt=T 3) doy 4 Lege Bata Kalenaat eK [@,-0,)dt=0 (4) TORQUE-VOLTAGE ANALOGOUS CIRCUIT ‘The given mechanical system has two nodes (J, and J.). Hence the torque-voltage analogous electrical circuitwill have two meshes. The torque applied to J, is represented by @ voltage source in first mesh. The elements J,, B, and K, are connected first node. Hence they are represented by analogous element in mesh-1 forming @ closed path. The clements J, B,,K, and K, are connected to second node. Hence they are represented by analogous elements in mesh-2 forming a closed path, ‘The element K, is common between node-1 and 2 and so itis represented by analogous element as common element between two meshes. The torque-voltage electrical analogous circuit is shown in fig 4, The electrical analogous elements for the elements of mechanical rotational system are given below. T set) jo ky BR, kK, = 1c, o, i, beh 9 BOR K, > 1, 2, 1,Figs. Fig 6. ‘The mesh basis equations using Kirchoff's voltage law for the circuit shown in fig 4 are given below (Refer fig 5 and 6). ti, ae tay ug Rae STi (5) di, Tee ; La Git Ral + Jiatt+ fa inat=0 wee(8) itis observed that the mesh basis equations (5) and (6) are similar to the differential equations (3) and (4) goveming the mechanicalsystern. TORQUE-CURRENT ANALOGOUS CIRCUIT ‘The given mechanical system has two nodes (J, and J,). Hence the torque-current analogous electrical citcuitwill have two nodes. The torque applied to J, is represented asa current source connected to node-1 in analogous electrical circuit. The elements J, B, and K, are connected to firstnode. Hence they are represented by analogous elements as elements connected to node-t in analogous electrical circuit. The elements J., B,, K, and K, are connected to second node. Hence they are represented by analogous elements as elements connected to node-2 in analogous electrical circuit The element K, is common between node-1 and 2. So itis represented by analogous element es common element between node-1 and 2. ‘The torque-current electrical analogous circuitis shown in fig 7. vs Lu ve Fig 7: Torque-current elecirical analogous circuit.Ghapiert Mathematical Models of Control Systems. 16 Fig 8. Fig 9. The electrical analogous elements forthe elements of mechanical rotational system are given below. Ti) BUR, 0,>y, 4,36, Kou, BUR, a, >%, 4,56, Kou, ‘The node basis equations using Kirchoff's current aw for the circuit shown in fig 7 are given below (Refer fig 8 and 9). OG ys Lit vadt itp (7) 1 1 eve 7 lvate Tle -v,)Jdt=0 (8) itis observed that the mesh basis equations (5) and (6) are similar to the differential equations (3) and (4) governing the mechanical system, EXAMPLE 1.13 Ky . . a od 7 K, White the differential equations governing the mechanical _\ | | a: HH} B, SL rotational system shown in fig. Draw the torque-voitage and Yi ‘+i. i ye ‘torque-current electrical analogous circuits and verify by writing * By mesh and node equations. Fig !. SOLUTION ‘The given mechanical rotational system has three nodes (moment of inertia of masses). The differential equations, i, Jb, B, > R, K, > 110, oy > i, bok B, > R, K, > 1c, o, > i Jy Ly i L, L L at Le SOS pe ce, eo =o, S ef mae R, a e(t) GY OY Ty | GY | : Rill, -i2) Sa, -j Fig 5 Torque-voltage electrical analogous circuit Fig 6.(Gapter 1 Mathematical Models of Control Systems Ralie—iS R Rv Rilla-i) - ~ GY Fig 7. ‘The mesh basis equations using Kirchott’s voltage law for the circuit shown in fig 5 are given below (Referfig 6, 7 2nd 8). Le Rid +L fll itt = et ol) Puli Pallet + fli =0 ) 1 Ralls h)+ Z finst=0 (9) Itis observed that the mesh basis equations (7), (8) and (9) are similar to the differential equations (4), (5) and (6) 2everning the mechanical system. RQUE-CURRENT ANALOGOUS CIRCUIT ‘The given mechanical system has three nodes (J,,J,and J,). Hence the torque-currentanalogous electrical circuit will save three nodes. The torque applied to J, is represented as a current source connected to node-1 in analogous electrical orcuit ‘The elements K,, J, and B, are connected to first node. Hence they are represented by analogous elements aselements onnected to node-1 in analogous electrical circuit. The elements J, B., B and K, are connected to second node. Hence they ze represented by analogous elements as elements connected to node-2 in analogous electrical circuit The elements J, B,, sod K, are connected to third node. Hence they aré represented by analogous elements as elements connected to node-3 in snalogous electrical circuit. ‘The elements K, and B, are common between node-1 and 2.and so they are represented by analogous element as Gammon elements between node-1 and 2. The element B, is common between node-2 and 3 and so itis represented as common #ement between node-2 and 3in analogous circuit. The torque-current electrical analogous circuits shown in fig 9. Wi-v2) og, aoe LJe—vant Ry Hie) Fe G c Bl Fig 9: Torque-current electrical analogous circuit.1.49 UR, K > 1, &:> Vy 49 8, > UR, K > 1”, o> vy, > G The node basis equations using Kirchoff’s current law for the circuit shown in fig 9 are given below (Refer fig 10, 1 and 12) ay 4 ty . Ora Ved flv) (10) dv, 1 1 1 Co Get Ma Wa vad Je at =0 eel 1) oN os Lvs) “ 0 (12) Itis observed that the node basis equations (10), (11) and (12) are similar to the differential equations (4), (5) and (6 governing the mechanical system. 1.11. BLOCK DIAGRAMS ‘A control system may consist of a number of components. In control engineering to show thi functions. performed by each component, we commonly use a diagram called the block diagram. A bloc diagram of a system is a pictorial representation of the functions performed by each component and o the flow of signals. Such a diagram depicts the intertelationships that exist among the various components The elements of a block diagram are block, branch point and summing point. BLOCK Ina block diagram all system variables are linked to each other through functional blocks. Th functional block ot simply block is a symbol for the mathematical operation on the input signal to th block that produces the output. The transfer functions of the components are usually entered in th corresponding blocks, which are connected by arrows to indicate the direction of the flow of signals Figure 1.25 shows the block diagram of functional block. The arrowhead pointing towards the block indicates the Input, A_[ Transfer | Output, B input, and the arrowhead leading away from the block represents | ee BoAGO the output. Such arrows are referred to as signals. The output signal from the block is given by the product of input signal Fig 1.25 : Functional block. and transfer function in the block.Chapter 1 Mathematical Models of Control Systems 1.50 SUMMING POINT Summing points are used to add two or more signals in the A AB system. Referring to figure 1.26, a circle with a cross is the symbol that adicates a summing operation. i The plus or minus sign at each arrowhead indicates whether the signal s to.be added or subtracted. It is important that the quantities being added or subtracted have the same dimensions and the same units. Fig 1.26 : Summing point. BRANCH POINT Branch point 5 A branch point is a point from which the signal from a block AP OLE Rake zoes concurrently to other blocks or summing points. A CONSTRUCTING BLOCK DIAGRAM FOR CONTROL SYSTEMS Fig 1.27: Branch point: A control system can be represented diagramatically by block diagram. The differential equations joverning the system are used to construct the block diagram. By taking Laplace transform the differential guations are converted to algebraic equations. The equations will have variables and constants. From the sorking knowledge of the system the input and output variables are identified and the block diagram for ich equation can be drawn: Each equation gives one section of biock diagram. The output of one section ill be input for another section: The various sections are interconnected to obtain the overall block Segram of the system. (AMPLE 4 Construct the block diagram of armature controlled demotor. SOLUTION ‘The differential equations governing the armature controlled dcmotorare (refer section 1.7), Vy=ieRe the f+0, ofl) T=Ki, sl) do as By Tad Gt Bo oe3) e =K,o Ad) i on 8 _ oes 8) Ontaking Laplace transform of equation (1) we get, Vols)=1(S) Rotts StF ES) (6) In equation (6), V,(s) and E,(s) are inputs and I,(s) is the output. Hence the V(sFEs) 2=eation (6) is rearranged and the block diagram for this equation is shown in fig 1. Vas) Va(6)~E,(8) = 1(6) [Re +5 La] L106) =e [MoE] “— MO) = Tory Mal)= Fig 1. On taking Laplace transform of equation (2) we get, WO) re? T(s)= KL (6) Fig 2 (7), 1,(s) is the input and T(s) is the output. The block diagram for this equation is shown in fig 2.Conttal En 151 On taking Laplace transform of equation (3) we get. we8) T(e)=Js o(s)+B ais) Inequation (8), T(s) is the input and a(s) is the output. Hence the equation (8) is rearranged and the block diagram o(s) bs this equation is shown in fig (3). Tis) = (Js +B) o(s) 1 ea _ On taking Laplace transform of equation (4) we get, (9) -.0(8)= E(S)=K ols) Inequation (9), o(s) is the input and E,(s) is the output. The block diagram for this equation is shown in fig 4. ‘On taking Laplace transform of equation (5) we get, o(s) #8 4(s) (10) {In equation (10), c9(8) isthe input and 6(s) is the output. Hence equation (10) is rearranged and the block diagram {ort “pias Fig 5. equation is shown in fig 5. 1 &s) = zals) . : “The overall block diagram of armature controlled demmotoris obtained by connecting the various sections shown in fig fig 5. The overall block diagram is showm in fig 6. Vss}-EXs) vss) TMs) 18) | 108) Os) Eds) Fig 6 : Block diagram of armature controlled de motor. EXAMPLE 1.15 Construct the block diagram of field controlled de motor. SOLUTION The differential equations governing the field controlled de motor are (refer section 1.8), ‘On taking Laplace transform of equation (1) we get, V4(s) = Ry W(s) +L 8 (8)452 ln equation (4), V,(s) is the input and |,(s) is the output. Hence the equation (4) is rearranged and the block diagram for =quation is shown in fig 1 v{s) 7 us) is) Jt | V(s)=44s) [P,+8t)] Rosh, ohe)= = vio) Fig L. (On taking Laplace transform of equation (2) we get, 1 Tis) = Ky Is) (8) “2 ge) In equation (5), |,(s) is the input and T(s) is the output. The block diagram for this Fig 2, pssonis shown in fig 2. Ontaking Laplace transform of equation (3) we get, T(s)=Js*0(s) + Bs0() ) of as) In equation (6), T(s) is input and 8(6)/s the output. Hence equation (6) is rearranged Js? + Bs) & Se block diagram for this equation is shown in fig 3. Fig 3 Tis) = (Js? + Bs) @(s) The overall block diagram of field controlled de motor is obtained by connecting the various section shown in fig 1 to The overall block diagram is shown in fig 4. 1_| Ms) Ts) 7 vis) > {s) Ry+sl; aK Js? Bs > 8(s) Fig 4 : Block diagram of field controlled de motor. K DIAGRAM REDUCTION The block diagram can be reduced to find the overall transfer function of the system. The following rules can be used =ieck diagram reduction . The rules are framed such thet any modification made on the diagram does not alter the input-output zor RULES OF BLOCK DIAGRAM ALGEBRA. Rule-1 : Combining the blocks in cascade AG, AGG, AG BS, a ASS > A A_yany 45° Combining Parallel blocks (or combining feed forward paths) AG, Gy, ie AG,+AG,= A(G,+G, > a yao (Ges) Ay Gl AG +G,) Rule-3 = Moving the branch point ahead of the block A ree AG A {] AG. A AG,153. Rule-4 : Moving the branch point before the block AG A AG Aa} ase : Moving the summing point ahead of the block 8 By mi86 \ G2 A AIB a (AXB)IG | : gh—> A AG AG + BG = (A+B)G Sool BS gy AO 4 Rule-6 : Moving the summing point before the block B BIG B HUG] > TT 4.8 = AGHB G AGHB A AG A 6 Rule-7 : Interchanging summing point B B Achy AtB AtB=C = A AC Bs ACABEAIE-C c c Rule-$ ; Splitting summing points 8 8 —] A & AtB-C = A ASB AtB-C ¢ | c Rule-9: Combining summing points B B & ASBC A & AB ee 3 A 4 ¢ ¢ J - Rule-10 : Elimination of (negative) feedback loop (R-CH) (R-CH)G R Ss R G ¢ c> —— -—> cH oe i+GH C=R-CHG 3 C=RG-CHG = C+CHG=RG Proof: Cc GS SCUFHG=RG > FT Gy ~ Rules11 : Elimination of (positive) feedback loop } R ah c(Models istems 154 EXAMPLE 1.16 Reduce the block diagram shown in fig and find C/R. 'G,+—_——, 2 fz] ©) + x O Yo SOLUTION Step 1: Move the branch point after the block. Step 2:Eliminate the feedback path and combining blocks in cascade. GH) G, 1+GH &)-[ G \(e=) GG, +Gy qi ‘The overall transter function of the system, c= ecGoniral Spstems Engineering | Step 1: Moving the branch point before the block 155 EXAMPLE 1.17 Using block diagram reduction technique find closed loop transfer function of the system whose block diagrams shor infig 1 , aa Fi Fig I. SOLUTION RChapter i Mathematical Models of Control Systems > 1.56 Step 5:Eliminating the feedback path and combining blocks in cascade c G 1+GHG2 - GGG, +6.) GiG,G, +6,G, 1 6G, = 14+ GH, = GGG, +G,G,_ 7 GGG) hy > TECH +GGH, TGR, 1+GGH, +G,G,H, + GH GGH, G 1+GGH, Step 7:Fliminating the feedback path —— R GE,6,166, | 1+ G,GH,+G.G.H2 + GH, ___ GG.G,+G,G, C__TGGH=GGH GH _ G,G,G,+6,G, R 4,__ 6S.6s+G6. TG GH, +G,GH, +G,H, +GG,G, +66, TG GH, +G.G,H, + GH; RESULT The overall transfer function is given by, cL GiG,G3+G,Gg R_1+G,G;H, + G.G,H, + GH, +G,G,G, + GG, AMPLE 1.18 Detesmine the overall transfer function at forthe system shown in fig 1157 SOLUTION ‘Step 1: Maving the branch point before the block (S) seis) ‘Siep 3 : Eliminating the feedback path ro Ris) yes) ‘Step 5: Combining the blocks in cascade 158 ‘Step 6 : Eliminating the feedback path RS) GGG, cs) Ty GxG,H, + GzGoHe ‘Step 7: Combining the blocks in cascade 15G,G,Hj+G: . GGG. x 15G,GH,+ 6,64, 7 G, cis) GGG. GLH, CELy ols) GG,G.G, R®) 1+ G,G,H,+ G,GH, +6,G,G,G, SULT The overall transfer function of the system is given by, Ks) _ GiG,G,G, RS) 1+ G,G,H, + G,GH, GGG. =XAMPLE 1.19 Forthe system represented by the block diagram shouim in fig 1. Evaluate the closed loop transfer function when the =cutRis (i) atstation-t (i at station-t { Station-Il 8 Station-| Si SOLUTION @ Consider the input Riis at station-| and so the inputat station-Hlis made zero. Lethe output be Cs. Since there isno inputat statior-Il that summing point can be removed and resulting block diagramis shown in fig 21.59 Step 1 : Shift the take off point of feedback H, beyond Gsand rearrange the branch points i f Gs Fig 2. ‘Step 2: Eliminating the feedback H, and combining blocks in cascade Cc, »iG) Hy ‘Step 3: Eliminating the feedback path R(CGiiapiter T= Mathematical Models of Control Systems > 1.60. Cis) _ GGG. Ris) 1+ GsH, + GH, + G,G,G,H, @ ——_Considerthe inputR atstatior-I, the inputat station-1is made zero. Let output be C,. Since there is no inputin stalion-I that corresponding summing point can be removed and a negative sign can be attached to the feedback path gain H,. The resulting block diagram is shown in fig 3. ‘Step 1:Combining the blacks in cascade, shifting the summing point of H, before G,and rearranging the branch points. Fig 3. ‘Step 2:\nterchanging summing points and combining the blocks in cascade. -GAGGH +H) (+ Gath)LOL Step 5: Eliminating the feedback path GGG, +H) | (14 GHy)Gp = G - G,(1+G,H,) ;-(AGGH+H)) G THGF,+GIGGH TH) ~ 146, +G(GG,H,+H,) 1+GH5 ° 1+G)Hy cG G,(1+ GH.) TR 1sGjHy+GGGH,+H,) RESULT ‘The transfer function of the system with inputat station-lis, GGG. R 14GH)+GH, +GiG,GH, ‘The transfer function of the system with input atstation-tlis, co Gs(14 Gots) R 1+G,+G,(GG,H,+4,) EXAMPLE 1.20 For the system represented by the block diagram shown in the fig 1, determine C,/R, and C./R, * ~) yg} eg} —_S SOLUTION & Case (i) To fina = Inthis case set R, =0 and consider only one output C,. Hence we can remove the summing point which adds R, and need not consider G,, since G, is on the open path. The resulting block diagram is shown in fig 2.Fe [Models of Control Systems 1 ‘Step 1:€liminating the feedback path c > Fig 2. ‘Step 2: Combining the blocks in cascade and splitting the summing point i x i G.GAH, 1G, xe1.63 ‘Step 5: Eliminating the feedback path GG,(1+G,) | G=G,)-GG.CHH, ___ GG(1+ G,) Ry G+6))-G.G.G.HH, GS 1, SGA) GG,(1+G,) GG.GHH, + GG, Ce GiG,G5(14 Ge) R, (1+ GiG,) (17+G,)-G,G,G,44, c, Case 2; To find = a Inthis case set R, = 0 and consider only one output. Hence we can remove the summing point which adds R, and need not consider G., since G, is on the open path. The resulting block diagram is shown in fig 3. ‘Step 1; Eliminate the feedback path,Chapler t= Maibematical Models: ‘of Control Systems) Step 2:Combining blocks in cascade and spliting the summing point | G.GsH, \__, [2.604] 1G, Step 5: Eliminating the feedback path GGiGH R, (+G,) G#G,G,) C, 7 SiG PS ~ (15G,) (+ GG.) Step 6:Combining the blocks in cascade R, | GEGH, a (4G) (14G,G))- GG,G.GH, Ge ___GGsGsGeH, _ Ry (1+G,) (+#G,G,)- G.GGAF,Le RESULT ‘The transfer function of the system when the input and output are. Rr and C: Is given by, c. GiG.G, (1+ Gy) R, 4+G,G,)(1+G,)- GG.GHH, The transfer function of the system when the input and output are Ri and C2is given by, & GG,G,G5 Hp RWG) 0+6,6,)-G.4.G,HH, EXAMPLE 1.21 ‘Obtain the closed loop transfer function C(s/R(s) of the system whose block diagram is shown in fig 1 | Hh SOLUTION ‘Step 7 Splitting the summing point and rearranging the branch points ‘Step 2: Eliminating the feedback path RS) Step 3 Shifting the branch point after the block. R(S)[= Mathematical Models of Control. LOG ‘Step 4 Combining the blocks in cascade and eliminating feedback path Step 6 : Eliminating forward path RS) GGG, ———— — Es 1+ GH, GG, CGH, | {Gu} . fs) GiG.G, “R(s) 1¢GH,+G,GH, - GG] G ULT ' pe GB.G a The transfer function ofthe system's () “7; GH =G,G,H, GGA, Ct =XAMPLE 1.22 ‘The block diagram of a closed loop system is shown in fig 1. Using the block diagram reduction technique determine the sesei loop transfer function C(sV/R(s). R(S) 6 c(s) SOLUTION ‘Step 1: Splitting the summing point. cs)167 ‘Step 2: Eliminating the feedback path. R(S) HAS); (Step 4 : interchanging the summing points and combining the blocks in cascade RS) G(s) cs) 14+ GAs)HAs) GAs) 1+Ga(s}H,(s) a 5 S28) G(s) Hs) l¢———_ * cael) 2 T+G,(s)t(s) Gi(s)H\(s) ‘Step 6 : Combining the blocks in cascade cs) Gls) _ GAs) (G(s) +1 Ris) 1+ GAs) HAs) + G,ls) Gals) H.(S) RESULT ‘The transfer function ofthe systems, cfs) _ Gals) [G4s)+7] Ris) GAS) TLS) Gis) GaliHal cal Models of Control: 18 EXAMPLE 1.23 | _- Using block diagram reduction technique find the transfer function C(s)/R(s) for the system shown in fig 1. = 4 Hil Fig le BO UTION Step 1 : Rearranging the branch points RIS) @ cs RE) go cs) Step 3: Moving the branch point after the block. R(S) + Gy. cs) " 71+ GHA,Lo Step 4: Moving the branch point and combining the blocks in cascade. RIS) ae pe dL cs) 1G.G:I-— TG HA, 1+G,HH, & Step 6 : Eliminating feedback path and interchanging the summing points. Ris) Hy(1+ G2) | GG. EEGs 1 GHH, enenen GGG ~ 17G,HH,— GGG, RS) cs) GGG, TE GHA, -G.G.G,H, Ha(t+ GatsHe) &, eee T.GHH, -6.G,GH, . 66.6.6, ____ GG,G.G, —\( He+GHH)) 1+ G.HH2- G2G,G.H, + GG H.C GHHe) TGHH,-G.G.6H,)( GG,( ahapiler T-Maiiematical Models of Control Systen 170 Step 8 :Eliminating the unity feedback path. RIS) 86,8,6, ' Gia + G.GH(1+ GF) GG,G,G, _O8)__ TGA, -G,6.6,4,+6.6.4,(1+GHHL) “RO) ay GG,G.Gy 1+ GyHH, — GG,G,H, + GiG,Hy(1+ G.HHe) = GGG, TV GHD — G,G,G Hy + GiGH,(1+ GH) = GG,G,G, GGGG; “THHAL(G, + 6,G,G,H,)+ GGx(H, + G:G,)- GGG) ESULT The transfer function of the systemis, ce) OB.8.G, Ris) THHHAG, + GG,G,H) + GGaH, = Gy) - GxG,G,Hy 4:12 BLOCK DIAGRAM REDUCTION USING MATLAB “RANSFER FUNCTION OF A SYSTEM Let, G(s) be the transfer function of a system, When the transfer function is a rational function of ~ then using MATLAB the transfer function can be obtained from the coefficients of the numerator and “sominator polynomials as shown below. Let, the general form of G(s) be as shown below. Mp sit bys) + bs" aps’ tas? +a,88? +. by St by #Oy 48+ ay a= First, the coefficients of the numerator and denominator polynomials are declared as two arrays as own below. [mumcoF > [bo 61 62. > MTs |dencof = [a0 al a2. «ands | Next, the transfer can be obtained using the following commands of MATLAB. 6 = tfC8"); G = ([numcof], [den_cof]) “RANSFER FUNCTION OF CASCADE / PARALLEL / FEEDBACK SYSTEM Consider two systems with transfer functions G,(s) and G(s). Let the two transfer functions be Scional function of "s" as shown below. bys + bys“ + bys" ays tas" +a,8 G)=17 Me agM! ed gh? dysM + dso! +d,s™74. ops +0587 + 0,857 4. GA) = When the two systems are connected as cascade / parallel / feedback system, then the overt transfer function of cascaded system / parallel system / feedback system can be obtained using MATLAl Inorderto obtain the overall transfer function, first the coefficients of the numerator and denominas polynomials of G,(s) and G,(s) are declared as arrays as shown below. [numcofi = [b0 bi bs]; | den_cofl = [a0 at an]; | numcof2 = [d0 di aM]; den_cof2 = [cO ci cn]; When the two systems are connected in cascade as shown below, then the overall transfer functia G(s) of the cascaded system can be obtained using the following commands of MATLAB. —+1G} +E] > —E] Gc = tf('s'); {num_cofc, den_cofc] = series(mum.cofl, den_cofl, numcof2, den_cof2); , Gc = ([numcofc], [den_cofc}) ‘When the two systems are connected in parallel as shown below, then the overall transfer functig G,(3) of parallel system can be obtained using the following commands of MATLAB. GP = tf('s"); [numcofp, den_cofP]=parallel(num_cofi, den_cofl, num.cof2, den_cof2); cp = ([num.cofp], [den_cofp]) j When the two systems are connected in feedback as shown below, then the overall transi function G,(s) of feedback system can be obtained using the following commands of MATLAB. Logs GF = tf('s'); [numcofF, den_cofr] = feedback(num.cofl, den_cofl, numcof2, den_cof2){ GF = ({num_cofF], [den_cofF])<[iiapier 7 itatbematical Models of Control Systems > 172 =ROGRAM 1.1 consider the transfer functions of the two systems given below, G,(s)=8/(s%42s+9) and G(s) =4/(s#6) write a MATLAB program to find the overall transfer function if the two systems are connected as cascade system, parallel system and feedback system. cle clear all Gl=tf(‘s"); G2=tf('s’); Gc=tf('s’);Gp=tf(‘s')iGr=tf(‘s num_cofi=[0 0 8]; den_cofile[1 2 9]; disp(‘system1’); G1=tf({numicofl], [den_cofl]) num_cof2=[0 4]; den_cof2=[1 6]; disp(‘system2'); G2=tf(Lnum_cof2], [den_cof2]) [num_cofc, den_cofC]=series(num_cof1,den_cofl, num_cof2,den_cof2); disp(‘cascade system’); ec-tf([numicofc], [den_cofc]) [num_cofP, den_cofp]=parallel(num_cofl,den_cofl,num_cof2,den_cof2): disp(‘Parallel system’); GP=tf([num_cofp], [den_cofr]) [num_cofF, den_cofF]=feedback(num_cofl,den_cofl, num_cof2,den_cof2) ; disp(‘Feedback system’); GF=tf({[num_cofF], [den_cofF]) DUTPUT systemi Transfer function: & saz +2549 system? Transfer function: 4 ste Cascade system Transfer function: SAP 4 8 S42 + 215 + 54 Parai7e] system Transfer function: 4sA2 +165 + 84 AZ +B SA2 + 215 + 54 Feedback system Transfer function: 8s +48 SAR + 8 5A2 + 21 5 + 861.73 113 SIGNAL FLOW GRAPH The signal flow graph is used to represent the control system graphically and it was developed by S.J. Mason. | A signal flow graph is a diagram that represents a set of simultaneous linear algebraic equations. By taking. Laplace transform, the time domain differential equations governing a control system can be transferred to a set of algebraic equations in s-domain. The signal flow graph of the system can bd constructed using these equations. It should be noted that the signal flow graph approach and the block diagram approach yield the same information. The advantage in signal flow graph method is that, using Mason's gain formula the overall guin of the system can be computed easily. This method is simpler than the tedious block diagrar reduction techniques. ‘The signal flow graph depicts the flow of signals from one point of a system to another and gives the relationships among the signals. A signal flow graph consists of a network in which nodes are connect by directed branches. Each node represents a system variable and each branch connected between twi nodes acts as a signal multiplier. Each branch has a gain or transmittance. When the signal pass throug a branch, it gets multiplied by the gain of the branch. : In a signal flow graph, the signal flows in only one direction. The direction of signal flow ig indicated by an arrow placed on the branch and the gain (multiplication factor) is indicated along th branch | EXPLANATION OF TERMS USED IN SIGNAL FLOW GRAPH Node : Anode is a point representing a variable or signal. Branch : A branch is directed line segment joining two nodes. The arrow on the branch indicates the direction of signal flow and the gain of a branch is the transmittance. Transmittance : The'gain acquired by the signal when it travels from one node to another is called transmittance. The transmittance can be real or complex. Input node (Source) : It is a node that has only outgoing branches. Output node (Sink ) : It is a node that has only incoming branches. Mixed node : Itis a node that has both incoming and outgoing branches. Path : A path is a traversal of connected branches in the direction of the branch arrows. The path should not cross a node more than once. Open path : A open path starts at a node and ends at another node. : Closed path : Closed path starts and ends at same node. Forward path + Itis a path from an input node to an output node that does not cross any nods more than once. : Forward path gam —: {tis the product of the branch transmittances (gains) of a forward path. Individual loop : [tis closed path starting from a node and after passing through a certain of a graph arrives at same node without crossing any node more than once, Loop gain : Itis the product of the branch transmittances (gains) of a loop. Non-touching Loops + Ifthe loops does not have a common node then they are said to be non- touching loops“Licieap iar I thatbomaticalitadels of Control Systems > LA SOPERTIES OF SIGNAL FLOW GRAPH The basic properties of signal flow graph are the following : (i) The algebraic equations which are used to construct signal flow graph must be in the form of cause and effect relationship. (i) Signal flow graph is applicable to linear systems only. (iii) A node in the signal flow graph represents the variable or signal (iv) A node adds the signals of all incoming branches and transmits the sum to all outgoing branches (¥) A mixed node which has both incoming and outgoing signals can be treated as an-output node by adding an outgoing branch of unity transmittance. (vi) A branch indicates functional dependence of one signal on the other. (vii) The signals travel along branches only in the marked direction and when it travels it gets multiplied by the gain or transmittance of the branch. (viii) The signal flow graph of system is not unique. By rearranging the system equations different types of signal flow graphs can be drawn for a given system. (AL FLOW GRAPH ALGEBRA Signal flow graph for a system can be reduced to obtain the transfer function of the system using =e following rules. The guideline in developing the rules for signal flow graph algebra is that the signal at is given by sum of all incoming signals. Rule J + incoming signal toa node through a branch is given by the product ofa signal at previous node and the gain of the branch. Example: x a, of en x x * a Xa, M7OX FAX Rule 2 : Cascaded branches can be combined to give a single branch whose transmittance is equal to the product of individual branch transmittance. Example: ab o—b—0 ro = o> x x % % % Rule 3: Parallel branches may be represented by single branch whose transmittance is the sum ~ of individual branch transmittances. Example: a+b ‘ a>: — emrantLB Rule 4 : A mixed node can be eliminated by multiplying the transmittance of outgoing brand (from the mixed node) to the transmittance of all incoming branches to the mixed node Example ; a x ac i x Se > : x x ,) (C* x xs be Rule § : A loop may be eliminated by writing equations at the input and output node and rearranging the equations to find the ratio of output to input, This ratio gives the gain of resultan: branch. Example ab ibe b x % To XRT 2 RK id i b i 6 o—$— be, ! Proof: x % % Xp = OX, $OKy 7 y= bey Put x; = ax; +x in the equation for x5. £X3=b (ax +Oq) => xy=abxytbexs => x;—bex,=abx, = x,(1—be)=abx, . X%}__ab “xy T-be ‘SIGNAL FLOW GRAPH REDUCTION The signal flow graph of a system can be reduced either by using the rules of a signal flow grapt algebra or by using Mason's gain formula. For signal flow graph reduction using the rules of signal flow graph, write equations at every no and then rearrange these equations to get the ratio of output and input (transfer function). The signal flow graph reduction by above method will be time consuming and tedious. $.J.! Masol has developed a simple procedure to determine the transfer function of the system represented as a sign: flow graph. He has developed a formula called by his name Mason's gain formula which can be directs used to find the transfer function of the system. MASON'S GAIN FORMULA : | The Mason's gain formula is used to determine the transfer function of the system from the signal flow graph of the system. Let, . R(s) = Infput to the system C(s) = Output of the system Now, Transfer function of the system, T(s) = s a ‘Mason's gain formula states the overall gain of the system [transfer function] as follows, snl Overall gain, T=— 2 PyAxGeib alibomahal Masel of Control Systems > 1.76 where, T = T(s)= Transfer function of the system Forward path gain of Ke forward path Number of forward paths in the signal flow graph 1 — (Sum of individual loop gains) . (= of gain products of all possible ) combinations of two non - touching loops, A = ‘Sum of gain products of all possible combinations of three non - touching loops, Ax = A for that part of the graph which is not touching K* forward path CONSTRUCTING SIGNAL FLOW GRAPH FOR CONTROL SYSTEMS A control system can be represented diagrammatically by signal flow graph. The differential equations governing the system are used to construct the signal flow graph. The following procedure can be used to struct the signal flow graph of a system. 1, Take Laplace transform of the differential equations governing the system in order to convert them to algebraic equations in s-domain. 2. The constants and variables of the s-domain equations are identified. 3. From the working knowledge of the system, the variables are identified as input, output and intermediate variables. 4. For cach variable a node is assigned in signal flow graph and constants are assigned as the gain or transmittance of the branches connecting the nodes. 5. For each equation a signal flow graph is drawn and then they are interconnected to give overall signal flow graph of the system. PROCEDURE FOR CONVERTING BLOCK DIAGRAM TO SIGNAL FLOW GRAPH The signal flow graph and block diagram of a system provides the same information but there is no Sendard procedure for reducing the block diagram to find the transfer function of the system. Also the ‘ock diagram reduction technique will be tedious and it is difficult to choose the rule to be applied for mplification. Hence it will be easier if the block diagram is converted to signal flow graph and Mason's sain formula is applied to find the transfer function. The following procedure can be used to convert Jock diagram to signal flow graph. 1. Assume nodes at input, output, at every summing point, at every branch point and in between cascaded blocks. 2. Draw the nodes separately as small circles and number the circles in the order 1, 2, 3, 4, ..... etc. 3. From the block diagram find the gain between each node in the main forward path and connect all the corresponding circles by straight line and mark the gain between the nodes. 4, Draw the feed forward paths between various nodes and mark the gain of feed forward path along with sign 5. Draw the feedback paths between various nodes and mark the gain of feedback paths along with sign.177 Ragan Systm ginooring EXAMPLE 1.24 Constructa signal flow graph for armature controlled de motor. SOLUTION ‘The differential equations goveming the armature controlled de motor are (refer section 1.7). LR thet ey: T=Kias T=J 2sbo; e=Ko; o=d8/dt On taking Laplace transform of above equations we get, V{S)=1(9)R, +L, 81,6) +E(s) wed Tis) = KI(s) venl2) Ts) =Jso(s)*Bolsy ee (3) AO a 4 as) =50(s) 5) The input and outputvariables of armature controlled de motor are armature voltage V,(s) and angular displacemen 0(s) respectively, The variables Is), T(s). (8) arid o(6) are intermodiate variables. ‘The equations (1) to (5) are rearranged & individual signal flow graph are shown in fig 1 to fig 5. V,(s)— Ex(s)=h(s) [Ry +8 La] & 1 yo AS) = eT (Malo) te] VO. 7 38 Rar LO TS)=KY) WT) X Fig 1 Fig E,() T(s) =0(s) [Js +B] ; = 3 . 1 +0 , -0(8)= Tg 8) Te os) Fig? ‘o(s) EJs)=K,0(6) . ho ™ Fig 4 = 80 (8) . (s) (8) Vs (3) ; 2846)=0(6) oso Fig 5 ‘The overall signal flow graph of armature controlled dc motoris obtained by interconnecting the individual signal flow graphs shown in fig 1 to fig 5. The overall signal flow graph is shown in fig 6. 1 Ve) 4 Ts) JssB (8) is es) 0) Fig 6: Signal flow graph of armature controlled de motor. 178 =XAMPLE 1.25 Find the overall transfer function of the system whose signal flow graph is shown in fig 1. =H A(s). G Gg G, G, 1 o> > > > > ed 1 4 4 5 8 a, > G Fig SOLUTION . % Forward Path Gains ‘There are two forward paths. -. K=2 Let forward path gains be P: and P, Ris) 1 G, G, G, G, G, 1 Cs) os > ——— > _ 1 2 3 4 5 6 7 38 Fig 2 : Forward path-I. Ris) 1 G, G, 1 C(s) Fig 3 : Forward path-2. Gain of forward path-1, P,= G, G,G,G,G, Gain of forward path-2, Ps GGG, Individual Loop Gain ‘There are three individual loops. Let individual loop gains be Py, Pri and Pa. Fig 4 : Loop-t. Fig 5 : Loop-2. Fig 6: Loop-3. Loop gain of individual loop-1, Pir Loop gain of individual loop-2, Pr Loop gain of individual loop-3, Par = -Gs Ha Gain Products of Two Non-touching Loops ‘There are two combinations of two non-touching loops. Let the gain products of two non touching loops be Przand P22, 3 G 46 Gs 7 -H é Ge 7 7 NI Gs Hy Hs 4 5 Hs Fig 7: First combination of 2 non-touching loops. | Fig 8: Second combination of 2 non-touching loops. Gain product of first combination = PyPar = (CGpH,) (Get) = of two non touching loops | Pra= Pua = (Gath) Got) = GaGetiy Gain product of second combination atte non tousing ope } Pa =PePar=(-GrGeh) (Gils) = GaGiGiHHe179 IV. Calculation of A and A, AVS 1-(Pri+Pa+Pm) + (Pre+Pn) = 1 ¢G2H1- GeGsHe- Gots) + (GGsHiHs+ GG iGsH:H) = 14 Galli + GzGsHe + Gat + GeGsHsHy + GiGsGsHoHs A: = 1, Since there is no partof graph which is not touching with first forward path. «The part of the graph which is non touching with second forward path is shown in fig 9. 3 4 Ag = 4-Pi=1-(-GeHh) = 1+ G2Hr Ns? V._ Transfer Function, T i Fig 9 By Mason's gain formula the transfer function, Tis given by, & T= iz Pebe =t (PA; +Ppz) (Number of forward paths is 2 and so K= 2) a G,G,G,G,G5 +G,G,G, (1+G,H,) 1+ GH, +G,G;H, + G.H, + GG.HH, + GG.GqH,H, . GG_G,G.G, +G,G.G, +G,G,G,G,H, 1+ GR, + G,G,H, + GH, + G,G,HA, ~ G,G,GH, G,G.G; [G,G, +G, 1G, + Get] 1 GH, = G,GoH, + GH, + G,G,H,H, + GG. -H. EXAMPLE 1.26 Find the overall gain of the system whose signal flow graph is R(s) 4 shown in fig 1. oe SOLUTION Letus number the nodes as shown in fig 2. 1. Forward Path Gains There are six forward paths. -. Lethe forward path gains be P1, Ps, Ps, Ps, Ps and Po. ve °8 7 8 Fig 3: Forward path-1, Fig 4 : Forward path-2.(Chaplet 1 Mathematical Models of Control Systems) 180 7 - Fig 7 : Forward path-5 Fig 8; Forverd paths Gain of forward path-1, P: = Gz Gs Ge Gain of forward path-2, G3GsGr Gain of forward path-3, Ps = G: Ge Gr Gain of forward path-4, Ps = Go Ge Ge Gain of forward path-5, Ps=-G:GsGrGeH: Gain of forward path-6, Ps =-G:GeGsGeHz Individual Loop Gain There are three individual loops. Hy Letindividual loop gains be Pn, Pai and Ps: +H, 7 8 Gc, Fig 9: Loop-1 Fig 10: Loop-2 Loop gain of individual loop-1, P11=-Gultt 7 8 Loop gain of individual loop-2, P= -GsHt2 Loop gain of individual loop-3, Pa = GiGshHe i. Gain Products of Two Non-touching Loops G; Hy +H, Fig 1 : Loop-3 There is only one combination of two non-touching 4 caps. Let gain product of two non-touching loops be Px H, Fig 12 : Combination of 2 non-touching loops PPer= CGiH) -GsH.) = GiGsHH Gain product of first 1- (Gal) = 1+ GiH: ‘There is no part of the graph which is non-touching with forward paths 3, 4, 5 and 6 +H, V. Transfer Function; T Fig 4 By Mason's gain formula the transfer function, Tis given by, T= + (z P, 4) (Number of forward paths is six and so K=6) z 1 ~ Em ar+ Peds Pads +Pae + Phy Pode) G,G.Go(1+ GH,)+ GxG,G_(1+ GH.) + GGG, +G,G,G, GsG,GH, - G,G,G,G5H, : 17 GH, + GH, -G.G.AH, +G,G,AR, EXAMPLE 1.27 Find the overall gain C(s)/R(s) for the signal fiow graph shown in fig 1. SOLUTION \__ Forward Path Gains ‘There are two forward paths. .. K=2. Letthe forward path gains be P, and P.. R@) <4, G, a GOs) o> 9 9 oo 4 2 3 4 5 Fig 2 : Forward path-1 RG) G, G cs) ° 1 2 3 4 5 Fig 3 : Forward path-2 @ Gain offorward path-1, P,=G,G,G,G, Gain of forward path-2, P,=G,G,G,* eee toop-1 Fig 6: loop-3 Gs _ . 4 Fig 7: loop-4 Fig 8 : loop-5 Loop gain of individual loop-1, P, Loop gain of individual loop-2, P,,=-H,G, Loop gain ofindividual loop-3, P,,= Loop gain of individual loop-4, P,, Loop gain of individual loop-5, P,, E -Gain Products of Two Non-touching Loops There are two combinations of wo non-touching loops. Let the gain products of two non-touching loops be, and P,,. NJ? 0s 4 Fig 9 : First combination of Fig 10 : Second combination of two non-touching loops two non-touching loops Gain product of first combination Pag = PoP of two non touching loops Gain product of second combination] ,, _ of two non touching loops a | Calculation of A and A, (-GH,} (Gs) = G,G3H, "51 = (-G,GzH,) (Bs) = -G,G.GeHy, A212 (PyitPaePoy+PrrtPa (Pa +P) =1-(-G,G,H,-H,G, -G,G,G,H, +@, -G,G,H,) + CGH,G5 - G2G,G,Hs) Since there is no part of graph which is not touching forward path-1, A1= 1 ‘The part of graph which is not touching forward path-2 is shown in fig 11. 2 A= 1-Gs | Ssansfer Function, T y Mason's gain formula the transfer function, Tis given by, = z SPA, (Number of forward path is 2 and so K = 2) Figit[Padi + Prd] =< [G,6,6,6, «1+6,G,6,01-6,)] GG,G,G, +G,G,G,-G,G,G.G, 1+G,G,H, +H,G, + G,G,G,H, — G; + G,G,H; — G,H,G, — G,G,G,H, EXAMPLE. 1.28 Find the overall gain C(s)/R(s) for the signal flow graph shown in fig 1 Figt SOLUTION ‘ Forward Path Gains There is only one forward path. .. K= 1. Let the forward path gain be P, RS) 4 G G G, G. 1 CG) > > > > > oe 1 2 3 4 5 6 7 Fig 1: Forward path-1 Gain of forward path-1, P,=G,G,G,G, Il. Individual Loop Gain ‘There are three individual loops, Let the loop gains be P,,, P,P, & Ss + NU" Les “+H, 3 4 Fig 3: loop-1 Fig 4: loop-2 Fig $ : loop-3 Loop gain of individual loop-1, P,,=-G,G,H, Loop gain of individual loop-2, P,, Loop gain of individual loop-3, P,, Ill. Gain Products of Two Non-touching Loops There are no possible combinations of two non-touching loops, three non-touching loops, etc. IV. Calculation of A and A, B=1-(P +P) +P) =1-(-G,G,H,-G,G.H, -GG,G,G,) =1+G,G,H, + G,G,H,+G,G,G,G, Since no part of the graph is non-touching with forward path-1, A,= 1.| Lifer F-Mitbonateal Models of Control Storis > 1.84 Transfer Function, T By Mason's gain tormula the transfer function, Tis given by, 1 | Ts Ret z Fade = Pri (Number of forward path isan 60 = 9) = GiG,G,G, 1+ G,G,H, + G,G,H, + GG,G,G, EXAMPLE 1.29 BeAMPLE 1.29 The signal flow graph for a feedback control system is shown in fig 1. Determine the closed loop transfer function IRIS). LUTION : Forward Peth Gains ‘There are two forward paths. =. K=2. Let forward path gains be P, and P,. RS), G G, Gq g 1 2 3 4 5 6 | Fig 2: Forward path-1 | S R&) g, G, Gg, Cs) 1 2 3 a 5 6 Fig 3 : Forward path-2 Gain of forward path-1, P:= GiG2 GeGe Gs Gain of forward path-2, P2= G:G2GsGs individual Loop Gain ‘Thore are four individual loops. Let individual loop gains be Pr, Par, Ps and Pas. G a i. 5 SS 5 A, H, H, H, Hy Fig 4: loop-1 Fig 5: loop-2. Fig 6; loop-3 Fig 7: loop-4 Loop gain of individual loop-1, P,, = GH, Loop gain of individual foop-2, P,,= GH, Loop gain of individual loop-3, P,, = GH, 6. Loop gain of individual loop-4, P,, = G,H,H, 22 -& 2 oo B. Gain Products of Two Non-touching Loops N77 A, Thereis only one combination of two non-touching loops. Let the gain Fig 8: First combination of seriscts of two non-touching loops be P,,. two non touching loopsLe Gain product of first combination ] of two non- touching loops J Pre = (GH) (G.Ha) =G,G/HH, \V._ Calculation of A and A, A= 1— (Phy + Part Pat Pas) Phe, = AM(GiH, + GaHp + GH + GgHaH,) + GoG.HH5 = 1- GH, — GH ~ GH — GeH pH, + GG, Since there is no part of graph which is non-touching with forward path-1and2, 4,=A,=1 V. Transfer Function, T By Mason's gain formula the transfer function, Tis given by, T -t Dd Pde = tea, +P,A,) (Number of forward paths is two and so K = 2) < . GG,G.G,G;+66,6,6, 1 GH, GH, — GAs — GH H, + GG. EXAMPLE 1.30 Convert the given block diagram to signal flow graph and determine C(s)R(s). Cs) SOLUTION ‘The nodes are assigned at input, output, at every summing point & branch point as shown in fig 2. 1. Forward Path Gains There are two forward paths. . K=2 Let the forward path gains be P, and,Fig 5 ; Forward path-2 Gain of forward path-1, Gain of forward path-2, G, FG, 1 l_ Individual Loop Gain Theres only one individual loop. Let the individual loop gain be P,,. NS? Loop gain of individual loop), P,,=-G,H. Gain Products of Two Non-touching Loops There are no combinations of non-touching Loops. W. Calculation of A and A, Pl 214+GH Since there are no part of the graph which is non-touching with forward path-1 and 2, A= A=1 * Transfer Function, T By Mason's gain formula the transfer function, Tis given by, _i zi 3 - GG -G, Tao Di Pade = SPP] = “Ga i EXAMPLE 1:31 Convert the block diagram to signal flow graph and determine the transfer function using Mason's gain formula. a SOLUTION The nodes are assigned at input, ouput, at every summing point & branch point as shown in fig 2.E87 The signal flow graph for the above black diagram is shown in fig 3, RS) 4 1 G, 1 G, Gg 1 1 CG) ° > > 1 & 9 1 10 Fig3 |. Forward Path Gains There aretwo forward paths... Ke2 Let the gain of the forward paths be P, andP,. Rs) 4 1 G, 1 Gq G, 1 1. & -—>—_ 0 —_ >——0- P——o- > >——o > o— > > 0 1 2 3 4 5 6 7 8 9 g, 9 Gain of forward path-1, P,=G,G,G, Fig 5 : Forward path-2 Gain of forward path-2.. P,=G,G, idual Loop Gain There are five individual loops. Let the individual loop gain be P,,, P,.. Psy P,,8nd P,,. Loop gain of individual loop-1, P,, Loop gain of individual loop-2, P, "at Loop gain of individual loop-3, P,; Loop gain of individual loop-4, P, Loop gain of individual loop-5, P.,=-G,H, Fig 6 : loop-I. 1 & Fig 8: loop-3- Fig 9 : loop-74.188 Gain Products of Two Non-touching Loops There are no possible combinations of two non-touching loops, three non-touching loops, et... Calculation of A and A, A=1-[P, AP, tPytPatPed = 1+G,GG,+G,GH,+G,G.H,+G,G,+G,H, Since no part of graph is non touching with forward paths-and 2, A,=A,=1. Transfer Function, T By Mason's gain formula the transfer function, T is given by, 1 1 T=—> RA, =—[PA,+ PA. TERM =HiPasePael = Bs +GiGy T1G,G,G, + GGH,+ G,G,H, + GG, GH, (MPLE 1.32 Convert the block diagram to signal flow graph and determine the transfer function using Mason's gain formula. BO_UTION ‘The nodes are assigned at input, output, at every summing point & branch pointas shown in fig 2. Earward Path Gains ‘There are two forward path, ~. Ke2 Let the forward path gains be P, and P,,(C Ganirol Sjstems ingineering RS) 1 1 G, G@ G, 1 1 Cs) — — > ~ do +9» 0 1 2 3 4 5 6 7 8 Fig 4 : Forward path-1. RG) 1 1 Cs) ° ° ° ° 3 4 5 6 Fig 4 : Forward path-1 Gain offorward path-1, P.=G,G,G, Gain offorward path-2, P,=G, jual Loop Gain ‘There are three individual loops with gains P,,,P,, and P,, G, G =H, tot 1 $ 5 9 Hy Fig 6 : loop-1. Fig 7 : oop-2. Fig 8 loop-3. Gain of individual foop-1, P,,= G,G,H, Gain of individual ioop-2, P,,= —G;G,H, Gain of individual loop-3, P,,= —G,H, ll Gain Products of Two Non-touching Loops ‘There are no possible combinations oftwo-non touching loops, three non-touching loops, ett... IV. Calculation of A and A, IV._Galculation of A and 4, A= 1-[P,,#P,+P,,] = 1-G,G,H,+G.G,H,+@H, Since no partof graph touches forward path-1, A, = 1 The partof graph non touching forward path-2 is shown in fig 9. 2A, = 1-[G,GH, -G,G,H,-GH,] = 1-G,GH,+G,G,H,+GH, V. Transfer Function, T By Mason's gain formula the transfer function, Tis given by, T -t Y RA =t [P,41+PoA9] (Number of forward paths is 2 and so K = 2) 1 = 5 [GGrGa + Gi(1-GiG,H, +G.G.H, + GH,)] [G.G.G, + G,-G,G,G,H, + G,G,6,H, + G,G,H,] _ G,G,Gp + G,-G6,G,G,H, + G,G,G,H, +G,G,H, 1 GGH,+ G,GyH, + GA,190 =XAMPLE 1.33 Drawa signal flow graph and evaluate the closed loop transfer function of a system whose block diagram is shown d 7 tig 1. SOLUTION ‘The nodes are assigned atinput, output, at every summing point & branch pointas shown infig 2. Fig 2 ‘The signal flow-graph for the block diagram of fig 2, is shown in fig 3. a = Forward Path Gains 8 ‘There are four forviard paths, -. K=4 Let the forward path gains be P,P, P,and P,, 9 yo y ¥ > > »~—0 1 2 3 4 5 6 7 Fig 4: Forward path-1. 1 1 * >—0 6 7 Fig 5: Forward path-2. 9 do yo Fig 7 : Forward path-4.Lo Gain of forward path-t, P= GG, Gain offorward path-3, P.= G,G,G,H, Gain offorward path-4, P,=-G,G,G,H, IL. - Individual Loop Gain There are two individual loops, let individual loop gains be P, and P,, -G, GG i 5 3 6 a 2 A, 8 Fig 7: loop-1 Loop gain of individual loop-1, P,,=—G,G,H,H, Loop gain of individual loop-2, P,,=G,G,H,H, I. Gain Products of Two Non-touching Loops There are no possible combinations of two non-touching loops, three non-touching loops, otc... IV. Calculation of A and 4, A=1- [sum of individual loop gain]=1- (P,,+P,,) =1-[G,G.H,H,+G.G,H.HJ = 1+6,G,H,H,- .G.HH, Since no part of graph is non touching with the forward paths, A,= A,= A,= Ay V. Transfer Function, T By Mason's rin formula the transfer function, Tis given by, tai tz Ped = PPAF APs umber of forward paths is 4 and so K~ 4) : oa GG, +G,G.G,H, -~G.G,G,H, 1+ G,G.HH> — GGHH, _ GAGs-Gy)+ G.G.HAG-G,) _ Gy(Gy-G2\1+G.He) 1+GHH,(G3 — Ga) 1+ GHH,(G3— Ge) 14_ SHORT QUESTIONS AND ANSWERS Qi1 What ts system? When a number of elements or components are connected in a sequence to perform a specific function, the group thus formed is called a system. Q1.2 What is control system? Asystom consists of a number of components connected together to perform a specific function. In a system when the output quantity is controlled by varying the input quantity, then the systemis called control system. The output quantity is called controlled variable or response and input quantity is called command signal or excitation. Q1.3 What are the two major type of control systems? The two major type of control systems are open loop and closed loop systems. Q14 Define open loop system. The control system in which the output quantity has no effect upon the input quantity are called open loop control system. This means that the output is not fedback to the input for. correction,Ghar Meaibornaical » sels of Control Systems) 192 Define closed loop system. The control systems in which the output has an effect upon the input quantity In order to maintain the desired output value are called closed loop control systems. What is feedback? What type of feedback is employed in control system? The feedback is @ control action in which the output is sampled and a proportional signal is given to input for automatic correction of any changes in desired output. Nogative feedback is employed in control system. What are the components of feedback control system? The components of feedback control system are plant, feedback path elements, error detector and controller. error detector Open loop ystem (Plant) Controller Reference input Output Feedback @-2 Why negative feedback is invariably preferred in a closed loop system? The negative feedback results in better stability in steady state and rejects any disturbance signals, It also has low sensitivity to parameter variations. Hence negative feedback is preferred in closed loop systems. 29 What are the characteristics of negative feedback? The characteristics of negative feedback are as follows : () accuracy in tracking steady state value. (li) rejection of disturbance signals. (iil) “low sensitivity to parameter variations. (iv) reduction in gain at the expense of better stability. What is the effect of positive feedback on stability? - The positive feedback increases the error signal and drives the output to instability. But sometimes the positive feedback is used in minor loops in control systems to amplify certain internal signals or parameters. Ii. Distinguish between apen loop and closed loop system. 1B! ypen loop Open loop Closed loop 1.Inaccurate & unreliable. 1. Accurate & reliabie. 2.Simple and economical |2. Complex and costly. 3.Changes in output due to external |3. Changes in output due to extemal disturbances disturbances are not corrected automatically} are corrected automatically. 4.They are generally stable. |4. Great efforts are needed to design a stable system. € 12 What is servomechanism? The servomechanism is a feedback control system in which the output is mechanical position (or time derivatives of position e.g. velocity and acceleration). State the principle of homogenity (or) State the principle of superposition. ‘The principle of superposition and homogenity states that if the system has responses c,(t) and c,(t) for the inputs r,(t) and r,(() respectively then the system response to the linear combination of these input a,r,(t) + a, (2) is given by linear combination of the individual outputs a,c,(t) + a,¢,(t), where a, and a, are constants.1% Qn14 on1s Q1.16 Q117 Q1.18 OL Q1.20 Define linear system. Asystem is said to be linear, if it obeys the principle of superposition and homogenity, which states that the response of a system to a weighed sum of signals is equal to the corresponding weighes sum of the responses of the system to each of the individual input signals. The concept of linear system is diagrammatically shown below. a(t) 10 16 Leow 1) fa Pe System System ayn) + axtdt) }—» o,(t) r(t) oO > c(t) rt) >) a zal edt) = a,c) + ae(t) system ” Fig QI.14 : Principle of linearity and superpo: What is thme Invariant system? Asystemis said to be time invariant if its input-output characteristics do not change with time. A line: time invariant system can be represented by constant coefficient differential equations. (In line: time varying systems the coefficients of the differential equation governing the system are function time). | Define transfer function. The transfer function of a system is defined as the ratio of Laplace transform of output to Le transform of input with zero initial conditions. (It is also defined as the Laplace transform of t impulse response of system with zero initial conditions). What are the basic elements used for modelling mechanical translational system? ‘The model of mechanical translational system can be obtained by using three basic elements m: spring and dashpot. Write the force balance equation of ideal mass element, Let a force f be applied to an ideal mass M. The mass will offer an opposing force, f,, which proportional to acceleration Lex f—>M Hrite the force balance equation of ideal dashpot. Leta force f be applied to an ideal dashpot, with viscous frictional coefficient B. The dashpot will of an opposing force, f, which is proportional to velocity. | box bx | pe t= Se,-x,) reference | 8 B Write the force balance equation of ideal spring. Let a force f be applied to an ideal spring with spring constant K. The spring will offer an opposi force, {, which is proportional to displacement. g Lex ex g f— TO K Les COS K reference“Thai TMathemaiical Models of Control Systems > 194 p24 What are the basic elements used for modelling mechanical rotational system? The model of mechanical rotational system can be obtained using three basic elements mass with moment of inertia, J, dash-pot with rotational frictional coefficient, B and torsional spring with stiffness, K. Write the torque balance equation af an ideal rotational mass clement, Let a torque T be applied to an ideal mass with moment of inertia, J J. The mass will offer an opposing torque T, which is proportionalto. —T¥ wg, angular acceleration. tet ay oe i at? Write the torque balance equation of an ideal rotational dash-pot. Leta torque T be applied toa rotational dash-pot with frictional coefficient B. The dashpot will offer an opposing torque which is proportional to angular velocity | \ | \ o ~« ! g Par T=1,-B2% T 6 4 at B B T=7,-B 5-63) Write the torque balance equation of ideal rotational spring. Let torque T be applied to an ideal rotational spring with spring constant K. The spring will offer an opposing torque T, which is proportional to angular displacement. Name the two types of electrical analogous for mechanical system. The two types of analogies for the mechanical system are force-voltage and force-current analogy. Write the analogous electrical elements in force-voltage analogy for the elements of mechanical translational system, Force, fF > Voltage,e | Frictional coefficient,B > Resistance, R Velocity, v ~ Current.i | Stiffness, K —> Inverse of capacitance, 1/C Displacement, x Charge,q _| Newton's second law, 2f = 0-> Kirchoff’s voltage law, Ev = 0 Mass, M + Inductance, L Write the analogous electrical elements in force-current analogy for the elements of mechanical translational system. Force, f > Current, i Frictional coefficient, B > Conductance, G=1/R Velocity, v > Voltage, v Stiffness, K > Inverse of Inductance, 1/L Displacement, x + Flux, Newton's second law, f= 0 > Kirchoff’s current law, 2 Mass, M > Capacitance, C Write the analogous electrical elements in torque-voltage analogy for the elements of mechanical rotational system. Torque, T Voltage, e |Stitiness of spring, K —> Inverse of capacitance, 1/C Angularvelocity,@ —>Current,i [Frictional cosfficient,B_ > Resistance, R Moment of inertia, J —rInductance, L Newton's second law, ET = 0 > Kirchoff’s voltage law, Sv = 0. ‘Angular displacement, 0 > Chargo, q|19 Gartrol Systems tagincoring 21.29 Write the analogous electrical elements in torque-carrent analogy for the elements of mechanic rotatlonal system. Torque, T ~ Current, i | Frictional cosfficient,B > Conductance, G= 1/R Angular velocity, « — Voltage, v | Stiffness of spring, K > Inverse of inductance, 1/1: Angular displacement, © -> Flux,@ | Newton's second law, ST = 0 —» Kirchoff ’s current law, Si = Moment of inertia, J > Capacitance, C 21.30 What is block diagram? What are the basie components of block diagram? Ablock diagram of a system is a pictorial representation of the functions performed by each componen of the system and shows the flow of signals. The basic elements of block diagram are block, brane: point and summing point. Q1.31 What is the basis for framing the rules of block dlagram reduction technique? ‘The rules for block diagram reduction technique are framed such that any modification made on the diagram does not alter the input output relation. Q1.32 Write the rule for eliminating negative feedback loop. Proof | R (R-CH)__ (R-CH)G C=(R-CH)G | “ , C=RG-CHG CH C+CHG=RG Cl + HG)=RG c_G R 1+GH Q1.33 Write the rute for moving the summing point ahead of a block. 8 —) 8G A (A+B)G = AG AG +.BG = (A+B)G Q1.34 What ts a signal flow graph? A signal flow graph is @ diagram that represents a set of simultaneous linear algebraic equations. By taking Laplace transform, the time domain differential equations governing a control system can b {transferred to a set of algebraic equations in s-domain. The signal flow graph of the system can be constructed using these equations. Q1.35 What is transmittance? The transmittance is the gain acquired by the signal when it travels from one node to another node in signal flow graph 91.36 What is sink and source? Source is the input node in the signal flow graph and it has only outgoing branches. Sink is a outpul Rode in the signal flow graph and it has only incoming branches. Q1.37 Define non-touching loop. The loops are said to be non-touching if they do not have common nodes. 21.38 What are the basic properties of signal flow graph? The basic properties of signal flow graph are, (@ Signal flow graph is applicable to linear systems.‘—Ghapler Mathematical Models of Control Systems 1.96 (i) It consists of nodes and branches. A node is a point representing a variable or signal. A branch indicates functional dependence of one signal on the other. (ii) _ A node adds the signals of all incoming branches and transmits this sum to all outgoing branches. (v) Signals travel along branches only in the marked direction and when it travels it gets multiplied by the gain or transmittance of the branch. (v) The algebraic equations must be in the form of cause and effect relationship. 9.39 Write the Mason's gain formula. Mason's gain formula states that the overall gain of the system [transfer function] as follows, Overall gain, T=1 Peay Ak T-=T(s) = Transfer function of the system K = Number of forward paths in the signal flow graph P= Forward path gain of K" forward path L “ of wt + [ sum of gain products of all possible | loop gains combinations of two non- touching loops ‘sum of gain products of all possible . - [service of three non- touching tops] - ‘A for that part of the graph which is not touching K” forward path 31.40 For the given signal flow graph, identify the number of forward path and number of individual loop. Number of forward paths = 2 Number of individual loops = 4 4.15 EXERCISES For the mechanical system shown in fig E1.1derive the transfer function. Also draw the force-voltage and force-current analogous circuits. x For the mechanical system shown in fig E1.2 draw the force- voltage and force-current analogous circuits. = Tx Lowy, bow Fig EL. 70) » LS M is —| LE -W— 2, SBF By 1 Ba Fig E12 Write the differential equations governing the mechanical system shown in fig £1.3(a) & (b). Also draw the force-voltage and force-current analogous circuit.LIT Fig E1:3(a) ™ Fig E1.3(b) 1.4 Consider the mechanical translational system shown in fig E1.4, Draw(a) force-voltage and (b) force. current analogous circuits. bev bows a cy pw oI B, Ls. whan paar Fig ELS Fig E14 E1.5 Wile the differential equations governing the rotational mechanical system shown in fig E1.5. Also draw the torque-voltage and torque-current analogous circuits. E16 In an electrical circuit the elements resistance, capacitance and inductance are connected in parallel across the voltage source E as shown In fig E1. Draw(a) Translation mechnical analogous system (b) Rotational mechanical analogous system. Fig E16 E1.7 Consider the block diagram shown in fig E.1.7(a), (b) (c) & (d). Using the block diagram reduction technique, find C/R. Fig E1.7(a) ffi} J A «——________| Fig El.7(6)1.98 Fig EL) Fig EL.(@) E 3 Convert the block diagram shown in fig E1.8 to signal flow graph and find the transfer function of the system. Fig E18 Consider the system shown in fig E1.9(a), (b), (c) & (d). obtain the transfer function using Mason's gain formula. 1 -—~o cs) An Aa As Fig E1.9(b) Xe199 Fig E190) Fig E1.9(d) Fig E110 Xs Xe E1.10 Consider the signal flow graph shown in fig E.1.10 obtain [4nd % 7 £1.11. Find the transfer functions of the networks shown in fig E1.11(a); (b), (c) & (@). R, R L yo ert Cc a) ga rn) ed) Fig El.11(a) R, l, L M55 [ CO % Ly e® °F ) RS tam — fem Rr, tet - | * Fig ELU(c) Fig E1.11(d) R, R, £1.12 Find the transfer function of the circuit shown in fig £1.12. 20} © uy) . ia) gu ett] ls “BIS Fig E112~~ The transfer function is ©) 1 F(s) (Ms? Bs 1K) 8S 1 | f oR Tt c | et ™ | Ait gL I | fe) MOL KOTG Qo MSO KOT yi «BOR vou BOR Force-voltage analogous circuit Force-current analogous cireult Bz OO RL, b a ee 4 Ti, re | _| £8, MOL, BOR BR ive, Mel KOU, Kom, vi BOR KU, Ka, vi, BOR, KC ur Force-voltage analogous circuit Force-current analogous circuit i) MS By, +8, ve) eK, [vet +Kflvr—vabdt= f(D d 2 MS +B,¥2 + B(vp — vy) + Ko] vodt + Kf (Ve — vd = F201) LR Rk uy “8 T W wr) +o $ et) 4M FF ™ T° bev R, C. R: Loo [i] We) Mol, BR, WoW MOC, BTR, voi, Mol Kc, VY, MC, Ko tL, v,7i, BR, K,>1C, v,>%, BUR, Kot, Force-voltage analogous circuit Force-current analogous circuit Bisib) MS ok, [vist Ka ov —vaddt + Ka fv vedet =f) Ka J v2 —vidt + Bly, — vs) =O; dy, Mae +B(V,-v,) HK Jy -v,)dt=0 Cc, v I ey ot 7 as 3 OY Oe) TL AS, Sb Ait) SC, FL. =o, te ; yy felt) ML, KC, fi) M>C, KM, vi Mol KC, yoy, MG, Koil, vp, BoR K-14, v,3¥, BOUR K,> 1, vy-> ly Vyas Force-voltage analogous circuit Force-current analogous circuit41.101 EL4 R, Ly NV 88° u R, a™ * R a" a(t) a) it) (! o FR ie ) iL 855" i) 00). Mol, B,>R, fi) M,C, vi, Mol, BOR, Woy, M,C, vp, ML, B,>R, vv, M,C. Y>h BSR Ko>1c MM BUR Force voltage-analogous circult Force-current analogous circuit do, es 4,22 PBK [(o- o,)dt = T(t); 2G + Baa + Kf (02 —0,)dt4 Kf (oy 3)dt 19% ao, fley-0-0 ve wv Te ee] 7 {| elt 1) Rg cC C, ‘Tey oye Tre] Ty i Toa Jol, B,+R, KC, TOI JC, By > 1k, K, > TL, ei, JO BR Koc, o,5y, Lo, BUR, Kan oi Jol BR, «31 o>¥, L9G BOR, 6. 5 Vy Torque-voltage analogous circuit Torque-current analogous circuit E16 Vs V2 Vs Bb B a) | \o] Oo Ly on M yj Lage K x Zé K Z J v Tt) es Bo ef) iv, iv, RoB WoT) io, io, RoB icv, CoM UC+K ve, LJ WK Analogous mechanical translational system — Analogous mechanical rotational system. GG, +G¥ ELT wee GH +G,+6GH,GGR, & cL GiG1+ G3) (Gi + G5) - Ro 1+(14+G,) (G, +G5) Hy +(1+ G,) (G, +G5) GH, @ S« GG, +GiG, Ro 1+GG,H,+G,G,+G,G, +G.H,+GH, w &- Sa, R 1+G,GjH +GiG,H,+G,G,G,1.102 is +G,+G,G,G,H,+G,GH, +6,G,G4 T+ GiGaH, + GA, + G/GHy Os) _ 6G.G, RS). 1 GG, + G,GH, +GG,G, C_ AAAs + AA Aun tApAmAuhas + ApAy Aes Ro T=Avghar Aaa + Aaglgs) + ApAareaes + AraAcraeAgs FAA agAseAgs) ~ (AraAzAaeAacAseAss) (a) (b) © _ GGG,G, (1+ GsGg + GgGz) + G,SsQgG7(1+ GG, + G,G5) © B°+GG,+G,G,+G,G, +G,G, +G,G,G,G, - G,G,G,G, +G,G,G,G, + G,G,G,G, ( Sa _ 16EGG1][1-(CrG2+ 6.6/5, +6) + G-GxGu] % 11G,G.Gp+ GpG,+G,G,G,G,G, + G,G,G,6,,+E,Giz 46,66; + Gx] +G,G,GeGp5 + GpB iG + G/GxGrg * =GG,G,6,+G6,6,4,+66.G,6,G, ; 2 G,G,6,G, @ FolS)__1+8 RC Es) 1+8(R)+R.)C Eu(s) _ 1 Ee SLC+sRC+1 w@ BE. ___ s&s E\(s) (S',C+SR,C+1 (S1,C+sR,C+1-1 @ Eal®) _ SRRO+Re E(s) sRR,C+(R,+R,) os) _ s*L,C : E(s) [sR,C+i][s*L,+L,)C+sR, C+1]-1CHAPTER 2 TIME RESPONSE ANALYSIS 2.1 TIME RESPONSE ‘The time response of the system is the output of the closed loop system as a function of time. Itis denoted by o(f). The time response can be obtained by solving the differential equation governing the system. Alternatively, the response c(f) can be obtained from the transfer function of the system and the input to the system. Co) _ I functi — = The closed loop transfer function, = Ta wey M(s) woe 20 The Output or Response in s-domain, C(s) is given by the product of the transfer function and the input, R(s). On taking inverse Laplace transform of this product the time domain response, ¢(t) can be obiained, Response in s-domain, C(s) = R(s) M(s) snn(2.2) Response in time domain, oft) = £'{C(s)} = £{Ris)x M(s)} ven2.3) where, M(s)= —@®)__ 1+G(s)H(s) The time response of a control system consists of two parts : the transient and the steady state response. The transient response is the response of the system when the input changes from one state to another. The steady state response is the response as time. t approaches infinity. RG) C(s) Ri {s) {>_> RO rep Input es} Response Meh (8) \(or Output). => M-—2®)_ His} 1+G@H@) Fig 2.1 : Closed loop system 2.2 TEST SIGNALS The knowledge of input signal is required to predict the response of a system. Ii most of the systems the input signals are not known ahead of time and also it is difficult to express the input signals mathematically by simple equations. The characteristics of actual input signals are a sudden shock, 2 sudden change, a constant velocity and a constant acceleration. Hence test signals which resembles these characteristics are used as input signals to predict the performance of the system. The commonly used test input signals are impulse, step, ramp, acceleration and sinusoidal signals. The standard test signals are, 1. a) Step signal 2. a) Ramp signal 3. a) Parabolic signal b) Unit step signal b) Unit ramp signal b) Unit parabolic signal 4, Impulse signal 5. Sinusoidal signal.Time Re sts, 2.2 Since the test signals are simple functions for time, they can be easily generated in laboratories. The mathematical and experimental analysis of control systems using these signals can be carried out easily. The use of the test signals can be justified because of a correlation existing between the response characteristics of a system to a test input signal and capability of the system to cope with actual input signals. STEP SIGNAL ‘The step signal is a signal whose value changes from zero to A at t= 0 and remains constant at A for t> 0. The step signal resembles an actual steady input to a system. A special case of step signal is unit step in which A is unity. r(t)a The mathematical representation of the step signal is, ri 320 tO (2.4) - 3 RAMP SIGNAL Fig 2.2: Step signal. The ramp signal is a signal whose value increases linearly with time *() from an initial value of zero at t = 0. The ramp signal resembles a constant velocity input to the system. A special case of ramp signal is unit ramp signal in which the value of A is unity. The mathematical representation of the ramp signal is, r()=At ; t20 : =0 5 tS0 00 (2.5) Or 2 ft Fig 2.3 : Ramp signal. PARABOLIC SIGNAL In parabolic signal, the instantaneous value varies as square of the.) time from an initial value of zero at t = 0. The sketch of the signal with respect to time resembles a parabola, The parabolic signal resembles a 4.54 constant acceleration input to the system. A special case of parabolic signal is unit parabolic signal in which A is unity. 2A The mathematical representation of the parabolic signal is, O.5Ab = 554 aes = 3 20 Fig 2.4 : Parabolic signal. =0 5 ¢<0 © 0 (2,6) [Note : Integral of step signal is ramp signal. Integral of ramp signal is parabolic signal, IMPULSE SIGNAL ‘A signal of very large magnitude which is (DO = B01 available for very short duration is called impulse *(t) = 5(2) A signal, Ideal impulse signal is a signal with infinite a magnitude and zero duration but with an area of A. => ‘The unit impulse signal is a special case, in which A is unity, The impulse signal is denoted by 3(t) and , ats > mathematically it is expressed as, Fig 2.5 : Impulse signal. 8) =; t=0 and focu-a myte0 (2.7)2.3 Gare Ssiemstagincerng) Since a perfect impulse cannot be achieved in practice itis usually approximated by a pulse of small width but with area, A. Mathematically an impulse signal is the derivative of a step signal. Laplace transform of the impulse function is unity. TABLE 2-1; Standard Test Signals Name of the signal Time domain equation Laplace transform of of signal, r(t) the signal, R(s) A Step A | a . 1 Unit step I i s i / A Ramp | At = | | 1 Unit ramp | t 2 2 A Parabolic a . 2 1 Unit parabolic 5 2 Impulse a 1 2,3 IMPULSE RESPONSE The response of the system, with input as impulse signal is called weighing function or impulse response of the system. It is also given by the inverse Laplace transform of the system transfer function, and denoted by m(t). Impulse response, m(t) = £' {R(s) M(s)} = £7 {M(s)} (2.8) sy’ : where, MO~ SHO R(s) = 1, for impulse Since impulse response (or weighing function) is obtained from the transfer function of the system, it shows the characteristics of the system. Also the response for any input can be obtained by convolution of input with impulse response. 24 ORDER OF ASYSTEM ‘The input and output relationship of a control system can be expressed by n® order differential equation shown in equation (2.9). a qt a a Fart) + aye) + a aeeP) +. d a Fn GPE) + ay p(t) = bog ad qm a 4 +b; ger) tbe U+ see Dn AO) + by a(t) wun(2.9) where, p(t) = Output / Response ; q(t)=Input / Excitation. The order of the system is given by the order of the differential equation governing the system. If the system is governed by 1! order differential equation, then the system is called mt order system.2.4 Altematively, the order can be determined from the transfer function of the system. The transfer funetion of the system can be obtained by taking Laplace transform of the differential equation governing the system and rearranging them as a ratio of two polynomials in s, as shown in equation (2.10). eb stab, oe? Transfer function, Te) = 2 bos = bist bys. (9) ast as (2.10) where, P(s) = Numerator polynomial Q(s) = Denominator polynomial The order of the system is given by the maximum power of s in the denominator polynomial, Q(s). Here, Q(s) =a, s*+a, 1 +a, 9" + taste, Now, n is the order of the system When n= 0, the system is zero order system When n= 1, the system is first order system. When n = 2, the system is second order system and so on. Note : The order can be specified for both open loop system and closed loop system. ‘The numerator and denominator polynomial of equation (2.10) can be expressed in the factorized form as shown in equation (2.11). 1(9 = PO _ StaNet er (2.1) Qs) (s+ p)(s+pa) where, 2, 2, z,, are zeros of the system. Ps Ps Now, the value of n gives the number of poles in the transfer function. Hence the order is also given by the number of poles of the transfer function. p, are poles of the system, Note : The zeros and poles are critical value, of s, at which the function I(s) attains extreme values 0 or 2. When s takes the value of a zero, the function T(s) will be zero. When s takes the value oj |a pole, the function T(s) will be infinite 2.5 REVIEW OF PARTIAL FRACTION EXPANSION The time response of the system is obtained by taking the inverse Laplace transform of the product of input signal and transfer function of the system. Taking inverse Laplace transform requires the knowledge of partial fraction expansion. In control systems three different types of transfer function are encountered. They are, Case? : Functions with separate poles. Case2 : Functions with multiple poles. Case 3: Functions with complex conjugate poles. ‘The partial fraction of all the three cases are explained with an example. Case 1: When the transfer function has distinct poles K bel, T9)= OS Pi) (8+ Po)2.5 on Ss ns By partial fraction expansion, T(s) can be expressed as, i -—_K—_-A._B_,_©¢ s(s+Pi)(S+P2) S$. S+Py S+P2 ‘The residues A, B and C are given by, A=) x9,_, B=T@ x +p) C= Ts) * G+ Pa) ban — Example 2 Let, Ts) <= oT) GID) By partial fraction expansion, T(s) can be expressed as, tee BS Serra 8 stl sid A is obtained by multiplying T(s) by s and letting s 2 2 2 tol =1 A= 1) 4,247 S641) (+2) vl, GHG6DL__, 12 B is obtained by multiplying T(s) by ned letting s =—1. 2 2 TK +1 —— y= 2.» B=Tis)= +3} “geo xD). Ge, aI) | C is obtained by multiplying T(s) by (s +2) and letting s = -2. 2 2 2 1 +2) — x (542 fy | C= Tis) x +2) wpern S27 TeeD),, , BaD | 2 12 40 | “19=Saye@ry sgt 342 Case 2: When the transfer function has multiple poles rete has multiple poles bet T= aw (*P.)” By partial fraction expansion, T(s) can be-expressed as, kK A, Be sG+p)G+P2) 8 S+P, tPs) +P) The residues A, B, C and D are given by. A=T(s) x3 B=T(s) x (s+p))| T(s)= =p . D= {fr x(S+P2 yl, =-0, | Example Let, T(s)= s(s+1) +2)? By partial fraction expansion, T(s) can be expressed as, K A,B c D —f_ =A +t s(s+1) (8+2)" (stl) (s+2)? (s+2) Ts)=€ 2 Tine. ysis) 2.6 A is obtained by multiplying T(s) by s and letting s = 0, A=T(s) x4 — 4 -—_2_| = sre s(stiy(st2y eo (s#1)(s#2)"). 4 B is obtained by multiplying T(s) by (s +1) and letting s = -1. 2 B=T +) = st = (6+ Da * Zar ear 1 \ C is obtained by multiplying T(s) by (s +2)? and letting s = C=T)x (+2)"|_ = x(8+2) *S(6+l) =~ =I 3 2241) 2 8(8+1) (8+2)" = D is obtained by differentiating the product T(s) (s +2)? with respect to s and then letting s = —2. 220d) _ =22(-2) +1) aus , 241) 4 ay) -4{2 d= j,[Tex06+2y 1 > & (=a 3 2 2 1 15 a Ste SS s(sti)(st2 0s) stl (st2)? s+2 S(s+l (2 Ts) Case 3 : When the transfer function has complex conjugate poles Let, T(s) = — Ke stp) tose) By partial fraction expansion, T(s) can be expressed as, K LA Bs+C bstc) stp, #+bs+e Tis)= svon(2.12) (s+P,) The residue A is given by, A=T(s) x (s+ PJ], The residues B and C are solved by cross multiplying the equation (2.12) and then equating the coefficient of like power of s. Finally express T(s) as shown below, ty) = Aa Bste [arya sinyay? stp, s+bs+e Let us express, s? + bs, in the form of (x+y). This will require addition and subtraction of an extra term (b/2/. Bs+C A Bs+C ——+ ——> FF SP paanber(bh ec (B) *P (5b) fc-2) 22 a 2) \like power of s, tA, Bate Fest > ay GiD@istl) s+2 Sesel =F 4208e(3) «-(3] 1=A(? +541) +(Bs+O) (8+2) ry 2 * \2 4841) +Bs?+2Bs+Cs+2C -(s+3) (3) =(5+0.5) +0.75 § Example 1 Let, T() = 5 —— ~ (s+2)(s'+s+1) By partial fraction expansion, 1 A Bs+C Ts) = +--+, G+2)@+stl) s+2 Sts A is obtained by multiplying T(s) by (s +2) and letting s 1 were o7) ARTE) xG+2),_ 1 3 “ya ee To solve B and C, cross multiply the following equation and substitute the value of A. Then equate the 2 1-5 43414 Be +2Bs+C5+20 On equating the coefficient of #* terms, 0= is B; 1 On equating the coefficient of s terms, seopee : 1 3 3 dos ji S 4st] 38 3(+s+l) 3(s+s+]) 1 s wt 1 ~ 36405 40.75 3 (8405) +0.75 2.6 RESPONSE OF FIRST ORDER SYSTEM FOR UNIT STEP INPUT The closed loop order system with unity feedback is shown in fig 2.6. R(s) / 1 +Q) a Fig 2.6 : Closed loop jor first order system. The closed loop transfer function of first order system, C= _1_ RG) 1+Ts cy ro ,[ 1 | og 1+Ts If the input is unit step then, 1(t)-1 and R)= = 1 11 «1 The response in s~domain, O(8) =R@F5= G5 ty ais)ar ET Resp gos > 28 By partial fraction expansion, _T A, BL fy) T. T. Ais obtained by multiplying C(s) by s and letting s= 0. 1 1 A=C()xd_, = sx} = -t =I T —L {s+2) sit T. 0 T 5-0 B is obtained by multiplying C(s) by (s+1/T) and letting s = -1/T. 4 T s fo BsC(s)xjsto VT. 1 C(9)= > - The response in time domain is given by, 1 -eT (2.13) c(t) = £'{C(s)} = £+ tL Sse | + vy) ‘The equation (2.13) is the response of the closed loop first order system for unit step input. For step input of step value, A, the equation (2.13) is multiplied by A. 7 «For closed loop first order system, Unit step response =1-e * ‘ | Step response = ali “| J When, t = 0, c(t) = When, 1 = 11, c(t) = When, t = 27, c(t) = When, t = 3T, c(t) = When, i = 4T, o(t) = When, t= ST, c(t) = When, t = ~, c(t) eel Here T is called Time constant of the system, In a time of ST, the system is assumed to have attained steady state. The input and output signal of the first order system is shown in fig 2.7.2.9 r(t) + o(t) 1 1 0.95 0.865} ~ 0.632) ---y oy ¥ 150 t i rr ree t Fig 27a: Unit step input. Fig 2.76 : Response for Unit step input. Fig 2.7 : Response of first order system to Unit step input. 2.7 SECOND ORDER SYSTEM ‘The closed loop second order system is shown in fig 2.8 a cl S) RQ) oR ©) COS + Oe Fig 2.8 : Closed loop for second order system. Ris) The standard form of closed loop transfer function of second order system is given by, &) o% Rs) +2f0,8+0° where, ©, = Undamped natural frequency, rad/sec. = Damping ratio. The damping ratio is defined as the ratio of the actual damping to the critical damping. The response c(t) of second order system depends on the value of damping ratio, Depending on the value of ¢, the system can be classified into the following four cases, Case: Undamped system, c=0 Case2 : Under damped system, O1, 5,5 =-6o, todo 1 {peers ween ae 219) When0<<1, 5,.s)=~o,to,y@?-1 --Go, t0,4(-D 1-07) =-€0,20,0-1 1-0 - " Fig 2.9.0 + Input. Fig 2.9.6 + Response. Fig 2.9 : Response of undamped second order system for unit step input. Using equation (2.24), the response of undamped second order system for unit step input is sketched in fig 2.9, and observed that the response is completely oscillatory. | Note : Every practical system has some amount of damping. Hence undamped system does not exist in practice. The equation (2.24) is the response of undamped closed loop second order system for unit step input. For step input of step value A, the equation (2.24) should be multiplied by A. | ~. For elosed loop undamped second order system, Unit step response = 1 cos at Step response =A(1 — cos ot) 2.7.2 __ RESPONSE OF UNDERDAMPED SECOND ORDER SYSTEM FOR UNIT STEP INPUT The standard form of closed loop transfer function of second order system is, os) 8 Rs) s*42Co,s+@, For underdamped system, 0 < ¢ t Fig 2.10.0 : Input Fig 2.10. : Response. Fig 2.10 : Response of under damped second order system for unit step input. 2.7.3 RESPONSE OF CRITICALLY DAMPED SECOND ORDER SYSTEM FOR UNIT STEP INPUT The standard form of closed loop transfer function of second order system is, C{s) a2 Rs) s?+2C@,s+@2 For critical damping ¢ = 1. - <9 _ oO a . = a (2.29) R@) s?+20,s+@2 (sto)? When input is unit step, r(t) = 1 and R(s) = 1/s. +. The response in s-domain, 2 o? eo? Cs) =R =1 Se =n (2.30) = ore S+O,) 8 (sta,) s(sta,)* G By partial sation expansion, we can write, oF A B c + 9) 3 (sto,) sto, A=sxC(s)),_ 0” Gro, B=(+0,)' «CO. = =-0, ™ lwa d 2 d{or _-03| _ c= <[s+0,) xo] a3} . FL =-1A OOS ero tere, 370, The response in time domain, (t= C@}= ell - 1 (s+o,) ao ot) = 1-0, teat erent e()=1-e™"(1+a,t) The equation (2.31) is the response of critically damped closed loop second order system for tinit step input. For step input of step value, A, the equation (2.31) should be multiplied by A. ~. For closed loop critically damped second order system, | Unit step response = 1-e"*"(1+0,t) | | Step response __= AfI-e**(1+0,t)] | Lp ee | Using equation (2.31), the response of critically damped second order system is sketched as shown an fig 2.11 and observed that the response has no oscillations. r(t) a 1 0 0 Fig 2.1.0 : Input. Fig 2.11.5 Response. Fig 2.11 : Response of critically damped second order system for unit step inpue. oy ov 2.7.4 RESPONSE OF OVER DAMPED SECOND ORDER SYSTEM FOR UNIT STEP INPUT The standard form of closed loop transfer function of second order system is, cy) 0} Ris) 2Co8+o; For overdamped system ¢ > 1. The roots of the denominator of transfer function are real and distinct. Let the roots of the denominator be s,,s,, Se )=~ Go, £0,461 [60,201] 232) Let j--sands--§ 2s) =Co,-o,y we(2.33) sate, to, nf234) The closed loop transfer function can be written in terms of s, and s, as shown below. Cos) oF CH (2.35) RG) 42%osror 645) (+5)2.15 For unit step input r(t) = 1 and R(s) = 1/s. 2 2C(s)= (2 = eee (s+3,) (S+5,) 8 ($+) (S+3)) By partial fraction expansion we can write, a A,B c (s)= ——*2 = 44 = + —_ 3645) GF) SHS, S+S 2 > A=sxC(S)kco= SX So On sGrHGtS), 85 = On - —=—= [sent [500+ Onye?=1 Co, -o a ) OD B=(s+9,)x C()|, s+) Tes, Fs) iB | “tn, soak 1 * [oak Ja NET & ot CHC(s)x(s+5)), = el sap _ @% o} = om 1 [-to,-0, fe -1+G0,-0Ne-1] [aoe] aga & f 7 lo 1 o 1 1 | o(t)= DN =-~— bee — + i 8 2-1 1% aay" 2-1 & (sts)| oft) = 1-— ae Let, tgs 2yC-1 % 2ye?-1 & em See ond 2.36 1 ( s & } where, s;=G0,-0,70?-1 FOOT The equation (2.36) is the response of overdamped closed loop system for unit step input. For ster input of valug, A, the equation (2.36) is multiplied by A. :, For closed loop over damped second order system, | fen i Unit step response = 1-—@2— 1 | £ where, 5; =Co,—o,¥C?-1 We=1 85 82 = 6, +@nyO =1 Step response = A 1-(i) et) > ~ t Fig 2.12.b : Response. 0 t Fig 2.12.4: Input Fig 2.12 : Response of over damped second order system for unit step input. Using equation (2.36), the response of overdamped second order system is sketched as shown in fig 2.12 and observed that the response has no oscillations but it takes longer time for the response to reach the final steady value. 28 TIME DOMAIN SPECIFICATIONS The desired performance, characteristics of control systems are specified in terms of time domain specifications. Systems with energy storage elements cannot respond instantaneously and will exhibit transient responses, whenever they are subjected to inputs or disturbances. The desired performance characteristics of a system of any order may be specified in terms of the transient response to a unit step input signal. The response of a second order system for unit-step input with various values of damping ratio is shown in fig 2,13. Os ©, 1 Ww ov Fig 2.13.0: Ingrut. Fig 2.13.b : Response. Fig 2.13 : Unit step response of second order system. ‘The transient response of a system toa unit step input depends on the initial conditions. Therefore to compare the time response of various systems it is necessary to start with standard initial conditions. The most practical standard is to start with the system at rest and so output and all time derivatives before t= 0 will be zero. The transient response of a practical control system often exhibits damped oscillation before reaching steady state. A typical damped oscillatory response of a system is shown in fig 2.14. The transient response characteristics of a control system to a unit step input is specified in terms of the following time domain specifications. 1. Delay time, t, 2. Rise time, t, 3, Peak time, t, 4, Maximum overshoot, M, 5, Settling time, t,7, Allowable error 2% or 5% out t & t Fig 2.14 : Damped oscillatory response of second order system for unit step input. The time domain specifications are defined as follows. 1, DELAY TIME (¢,) + It is the time taken for response to reach 50% of the final value, for the very first time. 2. RISE TIME (t) 2 Tt is the time taken for response to raise from 0 to 100% for the very 3. PEAK TIME a 4. PEAK OVERSHOOT ™) first time. For underdamped system, the rise time is calculated from 0 to 100%. But for overdamped system it is the time taken by the response to raise from 10% to. 90%. For critically damped system, i is the time taken for response to raise from 5% to 95%, Itis the time taken for the response to reach the peak value the ver first time. (or) It is the time taken for the response to reach the peak overshoot, M,. It is defined as the ratio of the maximum peak value to the final value where the maximum peak value is measured from final value. Let, ¢(s) = Final value of c(t). o(tp) = Maximum value of c(t). Now, Peak overshoot, M, elt) =A) (O37 eo) % Peak overshoot, %M, = p=) 199 (238 o(~) S.SETTLINGTIME(t) = Itis defined as the time taken by the response to reach and stay withir a specified error. It is usually expressed as % of final value. The usua tolerable error is 2 % or 5% of the final value. EXPRESSIONS FOR TIME DOMAIN SPECIFICATIONS Rise time (t) ‘The unit step response of second order system for underdamped case is given by, Att=t, oft) = c(t.) =1 (Refer fig 2.14).= sin(@,t, +8)=1 | Note : On constructing right | 1-¢? —— ws | angle triangle with Cand ji-E, sin(o gt, +8) =0 | we get i | Since ~e**s'r 40, the term, sin (w,t, +0) =0 | When, 6 =0,7,27,37... , sin 6=0 Sgt, +020 ot, =n-0 | +. Rise Time, t, = i 242.39) { a axl ag Here, @= tan! 2 og =0,f1-C (refer note) | | ~. Rise time, 1, = S insec | (2.40) 1-¢ | — 2 ld be measured in radian: Peak time (t,) To find the expression for peak time, t,, differentiate o(f) with respect to t and equate to 0. . ies FOlee, =o The unit step response of under damped second order system is given by, FS silos 8) yi-g Differentiating c(t) with respect to t. 5 = Jootot-0)0 ine a coat a) ee Co,) sin(o jt + only Put, oy =0,y1- cl an" ice os =o =e into +0) I= C wioss8] of (Go,) sin(o ,t+6)— Wat cos(ca yt +6) -¢ RE €r'[cos8 sin(wst-+0)—sin® cos(o,t+8)] (refer note) me en" [sin(@,t + 6) cos — costed gt + 9)sin8]2.19 Conirol Systems Engineering) e*** sin(@t) “ThE e*[sin((@,t+8)-@)] = i att=t,, say= 0 | Note ; On constructing right angle 1 ig! ig | cw, | wiangle with Sand fi-P, we get es"? sin(wgty) =0 | ® & . re We sind= fi-@ =F Since, e“°*"? +0, the term, sin(w tp) =0 080 =6 When @ =0, x, 22, 3x, sing =0 + @yty=m Thane Ot | -. Peak time, ,=— | | 4 | The damped frequency of oscillation, 4 =0,1-C? on (2.42) +. Peak time, tp Peak overshoot (M,) hae] %Peak overshoot, %M, = se x100 on 2.43) | oe) . where, o(t,) = Peak response at t= t, . 0(00) = Final steady state value. The unit step response of second order system is given by, erent vi-c et) = sin(o jt +8) Att=, oft) = (a) = MS sin(ogt+®) = 1-0 = 1 t oni Att=ty, oft)=c(t,)=1-5— sinfoost, +8) Vee oar 1 Og=0, yl-C =1- sin} @g—-+0 7 hoe toy sin (x +)=—sin9| oa aN gine +8) [Note : On constructing right angle ale with Cand 4fi-¢?, we get | vale sind= 1-2aera Rasps Ana 2.20 Aig) gg _ tte YP AI (20) 1 Percentage Peak Overshoot, %M, 100 aan (2.45) Percentage Peak Overshoot, %Mp Settling time (t. The response of second order system has two components. They are, € 1. Decaying exponential component, 2. Sinusoidal component, sin(og + é). In this the decaying exponential term dampens (or) reduces the oscillations produced by sinusoidal component. Hence the settling time is decided by the exponential component. The settling time can be found out by equating exponential component to percentage tolerance errors. For 2 % tolerance error band, at t For least values of € , eS"s's = 0,02 . On taking natural logarithm we get, ~C6,t,=10(0.02) => -to,t,=-4 => For the second order system, the time constant, T= Mp Settling time, .=—L-=4r (for 2% error) en(246) fo For 5% error, os = 0.05 On taking natural logarithm we get, —Goyt, = In(0.05) => -lo,t,-3 = fos | + Settling time, t, = = =3T (for S%erron)| (2.47) | ©, In general for a specified percentage error, Settling time can be evaluted using equation (2.48). : % 7 «Setting time, 4, = 202208) _ n(% error) Sn T2.21 EXAMPLE 2.1 baa the response of unity feedback systsm whose open oop transferfunction's G(s) = 4 5 on ‘and when the input isunitstep. SOLUTION Ris) ce Cis) The closed loop system is shown in fig 1. The closed loop transfer function, XS) - SiS) 1 Re) a 1+ is) Fig 1: Closed loop system. Cs)__s@+5) 55) 4 ~ 44 “Ris) 1,4 S(S+5)+4~ s(6+5)+4 «8745544 (544) (S#1) 3(8+5) (88) Teresponseins-domain, Cs) RS) aay Since the input is unit step, R(s)=—; — .Cls)= ince the input is unit step, FS) =~ 8) - SHED By partial fraction expansion, we can write, os)-——4+__- 4,8, o “s(set(era) s sti 844 4 | 44 Ax Ols) * 5 “SDEA,., bea 4 4 3(5+4)|,__, 1+ 4) B=C(s) x (s#1),_ 4 C=COs) x (+4), =a 3 The time domain response c(t) is obtained by taking inverse Laplace som of). (144 41 a 3s 3 oi 3 5+ + [fe*- Response in time domain, e(t)= £-'{C(s}}= £ RESULT Response of unity feedback system, c(t) = EXAMPLE 2.2 A positional control system with velocity feedback is shown in fig 1.. Whatis the response of the system for unit step input. SOLUTION The closed loop vansfer function, Cls)___ Gis) t lotse1 F(s)” 1+G(s) H(s) -, ves 100 Fig 1 : Positional control system. Given that, G(s)=——— and H(s)=O.1s +1 s(s+2) 100 100 Cs) _ s(st2) = s(s+2) = 100 _ 100 Ris) 100 © $(8+2)+1000.1S+1) s?+2s+10s+100 s*+12s+100 +( 2, Jorsen 3(8+2)Chapt ETime Response Analysis 22 Here (s? +128 + 100) is characteristic polynomial. The roots ofthe characteristic polynomial are, ae fies i00 =400 _ -12+j16 2 The roots are complex conjugate. The system is underdamped and so the response of the system will have damped oscillations. Su Sp 6+j8 The response in s-domain, C{s)=R(s) ———20 __ $74 1284100 Since input is unit step, R(s) -1 1 100 100 Cl) = = 557261100 ~ s(e*+ 12s +100) By partial fraction expansion we can write, 100 LA, Bs+C 3 (8*+128+100) 8 s*+12s+100 The residue A is obtained by multiplying C(s) by sand lettings =0. = 100 _ “100 - Cs)= The residue B and C are evaluated by cross muttiplying the following equation and equating the coofficients of like power of. 100 A, Bs+C Sistetaes 10) 6 a iaseT00 8 (8° +128+100) § s°+125+100 100 = A(s* +12s+100)+(Bs+C)s 100 = As* + 12As + 100A 4 Bs? +Cs On equating the coefficients of s*weget, O=A+B 2 BSA On equating coefficients of s we get, 0=12A+C 2 CS-12A=-12 2 14 s+12 _1__s+6+6 754252100 S74 128236464 9S (S46) +8 s+6 6 1 s+6 6 8 ” Breese (s+6+8 os (S46) +8" BB (S+6) +8" The time domain response is obtained by taking inverse Laplace transform of C(s).. Time response, c(t) = C'{C(s)} = of ited ws} s (s+6)?+8 8 (s+6)* ee =1-e'cos8t +4 @* sin8t= 1- «(3 singt cosa] The result can be converted to another standard form by constructing right angle triangle with Gand 1—¢? . The ‘damping ratio ¢ is evaluated by comparing the closed loop transfer function of the system with standard form of second order 100 0, =102.23 Co Constructing right angled triangle with Gand 1-C? weet, sind -08 ; cos = 08; tnd = 33 2.0 = tar t08 53-530 x 06 73 rad = 0.925 rad. 0° 2. Time response, o(t) = 8110 10 +[Esine +cosst| =1- on Bl Renee + 2 cost] 1.25 e [sin (8t+0)]=1- 125 e* sin(8t+ 0.925) RESULT ‘The response in time domain, t= we sin8t + cosa] or eft) =1-1.25 6 sin(8t+0,925) EXAMPLE 2.3 ‘The response of a servomechanism is, cit) = 1 + 0.2 e® - 1.2. 6% when subject to a unit step input. Obtain an expression for closed loop transfer function. Determine the undamped natural frequency and damping ratio. SOLUTION Given that, c(t) =1+0.2e-1.2e Ontaking Laplace transform of c(t) we get, +10)+ + Cs) 2-02 wen? ETD (s+ 60) (s i a = 8” +70s+600+02s?+2s-12s?-728 © 600 _i 600 5 +60) (8+ 10) S+60) (+10) s (6 +60) (s+10) Since inputis unitstep, R(s) = 1/s. * Os) = Fis) era ery * 7706785 Qs)____ 500 The ck func tem, Wo 70s 1600 fe closed loop transfer function of te system, > = aoe “The damping ratio and natural frequency of oscillation can be estimated by comparing the system transfer function with standard form of second order transter function. Cs) 2 600 * Ris)” sa ate,sr0% 87 +708+600 ‘On comparing we get, 3 = 600 <1. @, = V600 = 24.49 rad / sec RESULT cs) 600 The closed function of , SS) e closed loop transfer function of the system, 1° = 3a. aa Natural frequeney of oscillation, »,=24.49 rad/sec 43 Damping ratio, galys 2.24 EXAMPLE 2.4 ‘The unity feedback systems characterized by an open loop transfer function G(s) =K/S (+10). Determine the gain K, so thatthe system will have a damping ratio of 0.5 for this value of K. Determine peak overshoot and time at peak avershoot for aunitstep input SOLUTION Rs) wea ‘The unity feedback system is shown in fig 1. Os) __Gis) __ The closed loop transfer function Fray 7. e(s) Fig 1 : Unity feedback system. Given that, G(s)= Kis (s +10) —_K_ 3(s+10) K K “44% 6(S+10)+K 57+ 105+K {s+ 10) ‘The value of K can be evaluated by comparing the system transfer function with standard form of second order transfer function. . &)_ os __k_. “ R(s) 8? +2Co,s+w7 $7+108+K ‘On comparing we get, of =K Xe, =10 K=100 0, = 1K Put =0.5 and o, = VK o, = 10rad/sec 2 2x05xVK=10 vK=10 The value of gain, K=100. li? 499 -o5eie0s? . 199 = Percentage peak overshoot, %M, 1463x100 = 16.3% z k x Peak time, 1, = -=—2__ = = 0.363 see Pog waltat® 10vi-05? RESULT The value of gain, K => 100 Percentage peak overshoot, 4M, = 163% Peaktime, t, = 0.363500 EXAMPLE 2:5 ‘The open loop transfer function of a unity feedback systemis given by G(s) = Kis (ST + 1), where K and T are positive constant. By what factor should the amplifier gain K be reduced, so that the peak overshoot of unitstep response of the system is reduced from 75% to 25%. SOLUTION The unity feedback system is shown in fig 1. Ss). _ Os) Ris) 1+Gis) Re) cts) (S}-—p The ciosed loop transfer function, Fig I: Unity feedback system2.25 Given that, G(s)= Kis (sT +1) © Gls)__K/sisT #1) K Ris) T+K/s(sT +1) 8 (ST #1)4K 3 TT Expression for ¢ and », can be obtained by comparing the transfer function with the standard form of second order transferfunction, - Os) of K/T ” RS) On comparing we get, og =K/T 20,8 VKTT ‘The peak overshoot, M, is reduced by increasing the damping ratio . The damping ratio is increased by reducing the gain K. When M,=075, Let¢=¢, and K=K, When M, = 0.25, Let¢=Cp and K=K, anf Peak overshoot, M, = €)5*¥" Taking natural logarithm on both sides, in M, On squaring we get, (in M,)? On crossing multiplication we get, On equating, equation (1) & (2) we get, 1 (aM)? 1) aM). ee? uF Een ne AT Eero 2 1_ 4T (nM)? inna)? cn Mya? 4. (nM)? —G2(in M)? = GPa Cn MF (in M,)? 2 en? 4 2, 2 = (in M,)? = C?n? +¢7(In My ae (in M,)? =M4= 8 x100=e¥F 9 x 100 = 0163 x 100= 16.3% Time constant, T=1-= 1 = 05 sec Coq O54 For 5% error, Settling time, t, = 3T = 3x 0.5= 1.5 sec 1T=4x0.5=2sec For 2% error, Settling time, RESULT ‘The time domain response, oft) = 1— | Fania cost i2 | (or) oft) pr Prats 1047]Ghaprer 2 Time. ysis) 2.28 Rise time, t, ).6046 sec Peak time, i= 0.907 sec % Maximum overshoot, %M,= 16.3% Settling time, |S sec, for 5% error sec, for 2% error EXAMPLE 2.7 Aunity feedback control system is characterized by the following open loop transfer function G(s) = (0.48 +1)s(s +0.6). Determine its transient response for unit step input and sketch the response, Evaluate the maximum overshoot and the corresponding peak time. SOLUTION The closed loop transfer function, StS) -__Ss)_ Ris) + Gis) H(s) Given that, G(s) = (0.45 +1)/s(s +0.6) For unity feedback system, H(s) = 1. O4s+1 . &s) G(s) s(s+0.6) O4s8+1 “Ris) 1+Qs) 7, 048 5 (5+0.6)+0.4s+1 3(s+0.6) = 04841 OAS +1 08s +0.4s+1 s?+s+1 The s-domain response, C(s)= Fis) « 248*4 seset For step input, R(s) = 1/s. 1 O4s+i O4s+4 s s?+st1 s(s*+s+ . G(s) By partial fraction expansion C(s) can be expressed as, O4s+1 A Bs+C As)= = (ese) 8 séesed ‘The residue A is solved by multiplying C(s) by s and letting s=0, O4s+1 s?4se1 The residues B and are solved by cross multiplying the following equation and equating the coefficients of ike powers of. A= CIS) « |g =1 O4s+1 A, Bs+C s(st+se) 8 st4se7 On cross multiplication we get, 0.4541 =A(s?+s+1)+(Bs+C)s 0.4541 =As?+As +A+Bs?+Cs On equating coefficients of s?we get, O= A+B On equating coefficients ofswe get, 04=A+C -.C=04-A ce=4 -s-06 1 s+0.6 1 $+0.6 S S481 8 $7+5+025+075 5 (5 +2x05s+05)+0.75 s+0.5 40.75 s+0.5+01 _ (640.5 +0.75 J0.75 (s+0.5)?+0.75 © (+05) +0.75 -1 52.29 “The time domain response is obtained by taking inverse Laplace transform of Cis). The response in time domain, =o ceili 8105 0.75 at)= {ets} -f (1057 +075 078 wiotan 01 — =1-e% cos ¥0.75 t- sin 0.75 t Jars = 1-0 *[01185 sin(/0.75 t) rc08( 10.75 | The transient response is the part of the output which vanishes as t tends to infinity. Here ast tends to infinity the ‘exponential component e** tends to zero. Hence the transient response is given by the damped sinusoidal component. 0 sor 155 sin(V0.75 t) + cos(/0.75 | The value of and o, can be estimated by comparing the characteristic equation othe system with standard form of second order characteristic equation. ety ‘The transient response of c(t) = 84 2C0,5+0, 1.163 ‘On comparing we get, 1 ¢ opal +: @q = Trad/ sec 0 178.628 sec % “08 _gjg3 Fig: 1: Response of under damped system. Tee % Maximum overshoot, %M, = M, x 100 = 0.163 x 100 = 16.3% Maximum overshoot, M, = @ ze 5. = 3628 sec a = - © @, 1-0? 1 Vi-05 ‘The response of the system is underdamped and itis shown infig 1. ESULT [01185 sin(JO.75 t)+cos(V0.75 » Transient response of the system, c(t) ‘% Maximum peak overshoot, %M, = 163% Peaktime, t, = 3.628sec EXAMPLE 2.8 unity feedback control system has an amplifier with gain K, = 10 and gain ratio, G(s) = 1/s(s +2) in the feed forward path. A derivative feedback, H(s)=sK, is introduced as a minor loop around G(s), Determine the derivative feedback constant, K, so that the system damping factoris 0.6. SOLUTION ‘The given system can be represented by the block diagram shown in fig 1. RG), ce) \— ok, Eo Here, K,= 10; G(s)= 5 and H(s)=sK, ‘The closed loop transfer function of the system can be obtained {AG} by block diagram reduction techniques. |— ? trans! OR, Fis) se (24K,) 8410 ‘The given systemis.a second order system. The value of K, can be determined by comparing the system transfer function with standard form of second order transfer function given below. Standard form of } Cs) _ o? 2) Gs) ____ ee Second order transfer function} R(s) s?+20o,s+03 ‘On comparing equation (1) & (2) we get, o2=10 2. @, = ¥10 = 3.162 rad/ sec ro, 2 x0.6 x 3.162- 2=1.7944 RESULT ‘The value of constant, K, = 1.7944 EXAMPLE 2.9 ‘Aunity feedback control system has an open loop transfer function, G(s) = 10/s(s+2). Find the rise time, percontage overshoot, peak time and settling time for a step input of 12 units. SOLUTION | ‘Note : The formulae for rise time, percentage overshoot and peak time remains same for unit step and step input.2.31 Caoniratsystoms Engincoring * The unity feedback system is shown in ig 1. —_ . C6) The closed loop transfer function, gs 86 RO), ee ® ‘The closed loop transfer function, Given that, G(s) = 10/s (s+2) Fig 1: Unity feedback system 10 _ Os) s+2) wo 10 tt Ris) 4,__10 8(S+2)+10 s*+28+10 s(s+2) The values of damping ratio and natural frequency of oscillation @, are obtained by comparing the system transfe? function with standard form of second order transfer function, ‘Standard form of cs) «2 Second order transfer function| Fis) ‘On comparing equation (1) & (2) we get, 2 =10 2 10 = 3.162 rad/sec wees =0316 20, 3162 ee ~ 2 Geran tS — pap MICOS 1040 rad c 0316 @y=0,y1-0? = 3.162¥1- 0.316? =3 rad/sec Rise time, 1, -2—9 _ = 1249 _ 963 sec @g 3 a 23165 Percentage overshoot, %M, =e <100=e%"23 «400 = 0.3812 x 100=36.12% 12 Peak overshoot = 29:12 42 units = 4.2144 units Peaktime, ty 1047 sec 7 1 1 t, TH ee Time constant, T= = == S365 S162 *. For5% error, Settling time, t,= 3T = 3sec ‘tsec For2% error, Settling time, 1, =4T =4sec RESULT Rise time, t, = 0.63sec Percentage overshoot, NM, 3.12% Peakovershoot 4.2144 units, (fora inputof 12 units) Peaktime,t, 1.047 sec Setting time, t, 3secfor 5% error 4 sec for 2% error2,32 EXAMPLE 2.10 2, Aciosed loop servo is represented by the differential equation + at =6he Where cis the displacement of the output shaft, ris the displacement of the input shaft and e=r—c. Determine undamped natural frequency, damping ratio and percentage maximum overshoot for unit step input. SOLUTION ‘The mathematical equations governing the system are, @e_ de Grr eg ote esto (2) Pute=r-c in equation (1), we gic ‘ oS. Gage - 3 ” area 64{r -c) (3) Let L{c}=C(s) and {7} = R(s) On taking Laplace transform of equation (3) we get, 8? C(s) + 88 C(s) = 64 [R(s)-C(s)] 2 S*C(s) +85 C(s) +64 C(s)=64 Ris) (s?+ 85 + 64) C(s) = 64 Ris) _ a) 64 “ R{s) 87+ 85+64 eel) The ratio C(s)/R(s) is the closed loop transfer function of the system. On comparing the system transfer function with standard form of second order transfer function, we can estimate the values of Sando, Secondordertransferfuncion[ Rs) es 2estar ‘On comparing equation (1) & (2) we get, Standard form of : oe 8 8) a5" Percentage peak overshoot, %M, =e RESULT Undamped natural frequency of oscillation, = 8 rad/sec Damping ratio, 5 = 05 Percentage peak overshoot, %M, = 163%2.33 29 TYPE NUMBER OF CONTROL SYSTEMS ‘The type number is specified for loop transfer function G(s) H(s). The number of poles of the loop transfer function lying at the origin decides the type number of the system. In general, ifN is the number of poles at the origin then the type number is N. The loop transfer function can be expressed as a ratio of two polynomials in s. POS) _ 4. (+2) (S4Zp) (SHZs) eneenen G(s) H(s) =K——-=K. As) SS (+1) (S+P2) (8+ Ps) are zeros of transfer function are poles of transfer function wal (2.49) where, 2,» 2» Zy Py» Pa» Py K = Constant N=Number of poles at the origin The value of N in the denominator polynomial of loop transfer function shown in cquation (2.49) decides the type number of the system. If N = 0, then the system is type — 0 system. IfN = 1, then the system is type - 1 system If N = 2, then the system is type — 2 system IfN =3, then the system is type — 3 system and so on. 2.10 STEADY STATE ERROR The steady state error is the value of error signal e(t), when t tends to infinity. The steady state error is a measure of system accuracy. These errors arise from the nature of inputs, type of system and from non linearity of system components. The steady state performance of a stable control system is generally judged by its steady state error to step, ramp and parabolic inputs. Consider a closed loop system shown in fig 2.15. Let, R(s)_ = Input signal RO) rE) os) E(s) = Error signal C(s) H(s) = Feedback signal C(s)H(s) C(s) = Output signal or response Fig 2.15. The error signal, E(s)=R(s)-C(s)H(S) ae (2.50) The output signal, C(s)= ES) GS) a (2.51) On substituting for C(s) from equation (2.51) in equation (2.50) we get, E(s) = R(s) ~ [E() GS)] HG) E(s) + E(s) G(s) H(s) = RS) E(s) [1 + G(s) H()] = RG) =) __ sn (252) POT G6 HO Os Let, e(t) = error signal in time domain.2.48 RG) \ ve (2.53) ete CUE} = ne(t)= CBS) } = £ 16) He) Let, ¢,, = steady state error. ‘The steady state error is defined as the value of e(t) when t tends to infinity. 2@,= Lt ef) “ sewl2.58) The final value theorem of Laplace transform states that, If, F()=L{f@} then, Lt f() = Les Fs) wned2.55) Using final value theorem, ‘The steady state error, ¢,,= Lt e(t)= Lt sE(s)~ Lt —R)_ nns(2.56) be +0 340 14G(s) HS) 2.11 STATIC ERROR CONSTANTS: When a control system is excited with standard input signal, the steady state error may be zero, constant or infinity. The value of steady state error depends on the type number and the input signal. ‘Type-0 system will have a constant steady stafe error when the input is step signal. Type-1 system will have a constant steady state error when the input is ramp signal or velocity signal. Type-2 system will have a constant steady state error when the input is parabolic signal or acceleration signal. For the three cases mentioned above the steady state error is associated with one of the constants defined as follows, Positional error constant, K,= Lt G(s) H(s) sn (2.57) Velocity error constant, = K,= Lt sGS)HS) nel (2.58) Acceleration error constant, K, = Lt, G(s) H(s) 2.59) * The K,, K, and K, are in general called static error constants. 2.12 STEADY STATE ERRORWHEN THE INPUT IS UNIT STEP SIGNAL sR) Si =L teady state error, ex = Lt ey When the input is unit step, R(s)=1/s 1 s 1 1 1 =u ——S et — 1 - 30 THGG)HO 99 16GG)H() 1+ Lt GG) HG) 1+K, where, Ky = Lt, G(s) H(s) 18 enu(2.60) The constant K, is called positional error constant. Type-0 system 2.290 Py Pa-Pa- K, = LEG() His)= Lt x Sta Sia) stehe _ constant 40 ($+ py) (S+P2) (s+ Ps ney = = constant i+ K, Hence in type-0 systems when the input is unit step there will be a constant steady state error.2.35 Type-1 system (#2) (S+2,) (St 25a ar G(s) H()= Lt Ke BOO HO BBS GoD) GPs) O#Ps)onn In systems with type number 1 and above, for unit step input the value of K, is infinity and so the steady state error is zero. 2.13 STEADY STATE ERROR WHENTHE INPUT IS UNIT RAMP SIGNAL i — ©. 9 1+6@) HO) Steady state error, ¢,, = When the input is unit ramp, R(s) => 1 st eget — arp —1 __ -__t_ii eo (2.61) 0 14GG)HS) 0 stsG@)HG) Lt sG()H@) Ky 5 where, K, = Lt s G(s) His) ‘The constant K, is called velocity error constant. Type-0 system (s+4) (8+ Z) (s* (s+ Pi) (+ P2) S+Ps: K, = Lt sG(s) H(@)= Lt sk UK, =1/0=0 Hence in type-0 systems when the input is unit ramp, the steady state error is infinity. Type-1 system Ly Pi-P2-P; {8 +24) (8429) (S+ Zg)orene s(S*P,) (S*Pz) (S+Ps Ky= Lt sG(s) H(s)= Lt sk =K constant .¢,,=1/K, = constant Hence in type-] systems when the input is unit ramp there will be a constant steady siate error. Type-2 system (s+) (S+Z) (S+Z). & (S+p,) (8+ P2) (5+ Ps K,= LUsG(s) His)= Lt sk 2eyel/K,=1/0=0 In systems with type number 2 and above, for unit ramp input, the value of K, is infinity so the steady state error is zero.er Tima Response analats 2.36 2.14 STEADY STATE ERROR WHENTHE INPUT IS UNIT PARABOLIC SIGNAL tr —SR6)_ 590 14+-G(s) HG) Steady state error, e,, = When the input is unit parabola, R(s) = 2. $ ss 1 . 1 _l Lt —~S——_= lt =, = —_,______= seve(2.62} So G9 H@ 2 Pre GOHO Lt 7G(s) His) ee) where, K, = Lt G(s) H(s) The constant K, is called acceleration error constant. Type-0 system K,= Lb 266) His)= Le PK SEEDS +E) (SF 2)ore (s+ Py) (S*P2) (8+ Ps K, 0” Hence in type-O systems for unit parabolic input, the steady stale error is infinity. Type-1 system . 2g (8%) (+2) (6+ 23) Ky LesG6) HS) = 9K Gp) +P.) 61 Ps =0 Hence in type-1 systems for unit parabolic input, the steady state error is infinity. Type-2 system K,= L1G) HO)= Lt 2 (EF) (642) (84 tsar $°(8*p,) (+ Pp) (8+ Py)- constant He == constant R, Hence in type-2 system when the input is unit parabolic signal there will be a constant steady State error. Type-3 system Ke soe ni)= 1 ex Steere) 62 fF ON SS (5p) a) (SF Palo In systems with type number 3 and above for unit parabolic input the value of K, is infinity and so the steady state error is zero.2.37 ‘TABLE-2.2: Static Error Constant for ‘TABLE-2.3 : Steady State Exror for ‘Various Type Number of Systems ‘Various Types of Inputs [Error “Type number of system Input ¢ number of system |Constant [~ 0 Zz Signal 0 T 2 3 I unit sep |x| 9 | © | ° Unit Ramp © 0 a K, 0 ayy | Unit Parabolit ~ a K 2.15 GENERALIZED ERROR COEFFICIENT The drawback in static error coefficients is that it does not show the variation of error with time and input should be a standard input. The generalized error coefficients gives the steady state efror as a function of time. Also using the generalized error coefficients, the steady state error can be found for any type of input. The error signal in s-domain, E(s) can be expressed as a product of two s-domain functions. RR) 1 1+G() HS) 1+G(s) HE) where, F(s)= E(s)- R(s) = F(s) R(S) (2.63) 1+6@ HG) Let, e(t) = £+{E(s)} (error signal in time domain) “f= LF} r(t) = £°4R(s)} (input signal in time domain) The convolution theorem of Laplace transform states that the Laplace transform of the convolution of two time domain signals is equal to the product of their individual Laplace transform. ive, L{Rt) * (H}= Fs) RG) where « is the symbol for.convolution operation 2 LF) RO)}PH MY + 1) con(2.64) From equation (2.63) & (2.64) we can write, e(t) = ft) +) Mathematically the convolution of f(t) and r(t) is defined as, £0 + =f fC nt-T)aT ; where Tis a dummy variable 4 o e(t)= J £(D r(t-T) aT Itis assumed that the input signal starts only at t= 0 and does not exist before t= 0. Also we are interested in finding error signal at any time t after t= 0 (Le., for t> 0). Hence in the above equation the limit of integral can be changed as 0 to t. t ne= J f(T) t-T) aT 32.38 Using Taylor’s series expansion the signal r(t=T) can be expressed as, r= T= 10) TH) +e FORD PO. aT al where, i(t)=1* derivative of r(t) i(t)=2™ derivative of r(t) ‘T()=n™ derivative of x(t) On substituting the Taylor's series expansion of r(t — T), the error e(t) can be written as, cp Fo] aT nt f Ta Pa a=] f(T) [rota -F 1O-Z For. ‘ £ 1 oo ett) “J AT) eit ar] fT) THE) af fF FO ar t 3 t a -f oT rornef sen coh reo ar pea ! al Since r(t), i(t), i(t),.....r(t) are constants when the integration is done with respect to T, the error signal can be written as, ' ' ary t e@=ney f cr) at-iy f TAD a1 OF TECH) dt ° 3 0 | 2 FOF aa ai PAD dtt.. HD) ard T’ AT) dt. Let, C, =f £(1) dT c =f Tar) aT ) Cos = 3 9 : Qe- f TACT) aT ° t ' = ‘J TACT) aT C= orl TE(T) aT e(t) = r(t) Cy +8) C, +9 & rte) Se. 1G) See bese =Co (t+ Cy (8) “2 WO+ S 7 tt Sa F(t) onn(2.68) ‘The equation (2.65) is the general equation for error signal, e(t). ‘The coefficients C,, Cy. C, ...u.C, are called the generalized error coefficients or dynamic error coefficients. The steady state error e, is obtained by taking limit t —> 2 on e(t). +. Steady state error, ¢,, = And fee Cy + F(t) Cy +) Se ww S Ht +r = = Con(t)+C, we a+ + F(t) Fu vt (teen (2.66) _2.39 2.16 EVALUATION OF GENERALIZED ERROR COEFFICIENTS The generalized error coefficient is given by, 2 f 0 . =! c,=Cn J T°#(T) dT; where F(s) iran We know that £{f(T)} = F(s), hence by the definition of Laplace transform, F(s) “J £(T) “TdT On taking Lt on both sides of equation (2.67) we get, ‘ Lt F(S)= Lt, J £1) ed -j £(T) A erar~f f(T) dT=C, ° > + [Cy= LER@) On athrinng equation (2.68) with respect to s we get, £ 2 Fo) “ai £1) eT at = i rs 2 eat “| £1) CD ar = -J TAC) Tar 0 On taking Lt on both sides of equation (2.69) we get, 20 ut 2r@)= Le -[ tem ear soo ds 30 2 = f TET) Lt ear=-[ TET) aT=C, @ a d ByasF| On differentiating equation (2.68) on both sides with respect to s we get, afa _4l fF st x [Sern|- a TAT) € «| # F(s)= | TET) a (aT --| TET) (-T) ed ds ds 4 ¢ aoe. [rameter eoe(2.67) ove 2.68) wee 2.69) ved (2.70) swe(2.71)Applying the limit s > 0 on both sides of the equation (2.71) we get, @ ee Le (eH e* Harton yr amet ar ‘ : -jrm Le e“ar= [TAT AT=C, 0 F(s) seee2.72) “4 Similarly it can be shown that, Cc, = Lt 2 | } n(2.73) 30 ds* 2.17 CORRELATION BETWEEN STATIC AND DYNAMIC ERROR COEFFICIENTS The values of dynamic error coefficients can be used to calculate static error coefficients. The following expressions shows the relationship between them. “Tsk, we A2.74) Gq “e s(2.75) Get sno(2.76) Proof — _ CBO" Berane CORT 2.18 ALTERNATE METHOD FOR GENERALIZED ERROR COEFFICIENTS coats . RG) - E(s)-—RO)_ The error signal in s-domain, B(s)- = AO. I (2.77) “RG 1+G(s) HG) The equation (2.77) can be expressed as a power series of s as shown in equation (2.78). BO tg oes Ge. Sey, svon(278) R(s) 1+G(s) H(s) 2 3! 7 E(s)=Cy R)+CsR()+ LE RO)+Sh & es R(s}t...... ssen(2,79), On taking inverse Laplace sensor of equation (2.79) we get, e(t)= Corte Cnt) e ry> Gs Sr 2.80) The equation (2.80) is same as that’ of equation (2.65) in section 2.14.This method will, be useful to find the generalized error coefficients without using differentiation, but using laplace transform.2.4 EXAMPLE 2.11 Fora unity feedback control system the open loop transfer function, G(s) = a) the position, velocity and acceleration error constants, b) the steady state error when the inputis R(s), where R(S) = SOLUTION a) To find static error constants 3 5 Fora unity feedback system , H(s)=1 Postion error constant, K, = Lt, G(s)H(s)= Lt G(s) “hee Velocity error constant, K, = Lt,s G(s)H(s) = Lts Gis) = Lt, yaw Acceleration error constant, K,= Lt, S?G(s}H{s) = Lt s?G(s) 10(s+2) _ 102 zt s% 210x2, be ery 4 b) To find steady state error Method-I ‘Steady state error for non-standard input is obtained using generalized error series, given below. = 20 The error signal, e(t) = r()Cy HNC, HV yee, 2 Given that, rg)=3 3 Input signal in time domain, r(t)= £'{R(s)} = ©" 32,4 " s 38 = iv e = 9m = 3-24 2i@)- Sn(t)=-2+ 12-24 nit) = Silt) = 24 32t= 242 ity Sty aos ry=2 Sa Si =0 ae ‘The derivatives of (t)iszero after second derivative. Hence we have to evaluate only three constants C,,, C, and, ‘Tho gonoralized error constants are given by, : -ut : Co= Ltrs); C= LEAs); “Ge ia Sr F(s)= 1 on 1 _ s'(s+ . - s'+s? “T£G)Hs) 1+G(s) 4, 108+2) “s'(5++10(5+2) s°+5?+10s+20 (6+) ) sis? S587 +106+20|"° y= Lt F(s)= als2.42 C= Ut tris) = uf peed ssods| 55 +s? +10s+20 auf +8? +105 + 20)(35? + 2s) +8?)(3s? + 25 + 10) (ss? + 108+20)" alt 3s° + 2s! +384 + 28° + 308° + 20s? + 60s” + 40s - 35° - 2s* — 108° - 35¢ - 25° - 105? (s° +5? +105 +20)" C= Ly Srt8)= wale F(s) | > We 20s? +705? + 40s | (e+8?+ 108+ 20)" (s° +s? +108 +20) (60s? +1403 + 40) “poo cre +408) 2x (5? +s? +108 +20) (39? +28 + 10) _ 207x401 = (0? +s?.+108 +20)" 20% "10 L Error signal, e(t)=1(0)Cy +), +i(t) a oe a 60 Steady state error, @,, = Lt oft) = Method - Il The error signal in s-domain, E(s) __s 23s) __ . P(s+1)+10(s+2) “s%(s+1) s(s+) s%(s+1) _2 (s+) tl 2(s +1) s|5%(s+1)+10(s+2)| s?|s%(s+1)+10(s+2) | 3s°|s%(s+1)+10(s +2) The steady state error e,, can be obtained from final value theorem. Steady state error, @,, = Lt e(t)= Lt s E(s) e.-t sl3| 2(s+1) 2 | s+), 1 | (s+1) O55 8) 5] eeri)+10(6+2)|- s?| S(e++148+2)| 35°| 554+ 10(6+2) elt as{s+) sist) (s+1) =0-0+-1 soo |S%(er)+10(s+2) S(6+1)+10(s+2) 3s%(s+1)+30 (s+2) 60 12.43 Cetra emi Engincering > Method - Il Error signal in s- domain, E(s) “eGHS oo a “R(s) 1+ G(S)H(s) / 2 3 | sis. | Gwen that, Gis)= 20842). pyeynt 20 40 | s(s+1) 20+10s+s?+s4 s?+5? | - Els) 1 s(si1} 2 TRG) 5, 10(8+2) “$7541 +1062) Oo2 s(s+1) so+s? s?+s° 7s? +572108+ 20 20+10s+s+5° 20° 407” 2 E(s)= Rel. [20° 40 4 | = aeRO) + Esher. ‘On taking inverse Laplace transform of the above equation we get, 1 i —__ Wa sys Geto [Pre nono | by denominator polynomial, Given that, R(s) = =A) = £YR(S)} = cf it) Gry =-2+ ae = e a -Sry= Sin =t 0 3) 60 Stoady state error, ex. = Lt eft)= Lt b= ¥ ess oe ae 60 60 RESULT @ — Positionerror constant, Velocity error constant, Acceleration error constant, 3.201 (&) When, Ris)==-S+am EXAMPLE 2.12 For servomechanisms with open loop transfer function given below explain what type of input signal give rise toa constant steady state error and calculate their values. 20(8 +2) a) 89-3! ©) Gs)= 10 s(e> S72)Cigar Tame Raposo ina) 2.4 SOLUTION 20(s+2) (s+ 1(s+3) Letus assume unity feedback system, -. H(s)=1 ‘The open loop system has a pole atorigin. Hence itis a type-1 system. In systems with type number-1, the velocity (camp) inputwill give a constant steady state error. a) Gs) 1 The steady state error with unit velocity input, @,, = Velocity error constant, K, = .Lt, $ Gis) H(s)= LE, 8 G(s) 5 —2US+2)_20x2_ 40 S90” s(s+iXs+3) 1x3 3 075 Steady state error, ess = 10 (s+ 2\6+3) Letus assume unity feedback system, -. H(s)=1. The open loop system has no pole at origin. Hence itis 8 type-O system. In systerns with type number-0, the step input will give a constant steady state error. b) G{s)= The steady state error with unit step input, e,5 10 10 Position error constant, Ky = Lt Gl@)H(s)= Lt, G(s) = Neyer Bd Steady state error, e.= “to (s+ 1(s+2) Letus assume unity feedback system, «.H(s)=1. The open loop system has two poles at origin. Hence itis a typs-2 system. In systems with type number-2, the acceleration (parabolic) input will give a constant steady stale error. ‘The steady state error with unit acceleration input, e,, z 1, GOH)= 1 s*6l) = bys ©) Gis)= 1 Acceleration error constant, K, o Steady state error, e., RESULT 1. Insystem (@) with unit velocity input, Steady state error = 0.075, 375 2 2. In system (b) with unit step input, Steady state error= 3.in system (0) with unit acceleration input, Stoady state error =2.45 EXAMPLE 2.13 ‘The open loop transfer function of a servo system with unity feedback is G(s) = 10/s(0.1s+1). Evaluate the static error constants of the system. Obtain the steady state error of the system, when subjected to an input given by the polynominal, = ap rate Be, SOLUTION To find static error constant For unily feedback system, H(s) = 1. + Loop transfer function, G(s) Hs) = G(s) ‘The static error constants are K,, K,andK,. 10 Posi LK, = Lt Gis)= Lt ition error constant, K, = i(s)= Zoe + a to. ‘Velocity error constant, K, = Bt 8G(8)= ins HOts*1) 10 Acceleration error constant, K, = Lt, $*@(s)= Lt,s* werd To find steady state error Method - | ‘Steady state error for non-standard inputs obtained using generalized error series, given below. The error signal, e(t) =r(9C, +H(HC, + WZ. +r) S Given that, r(t)= a, +ajt+ oe ‘ y= Saty= Sfasate Ze) 2.08 i= weal (0)-Ste.+2,)=2, 2 f= Sn ${S0)-Se.)20 Derivatives of r(t) is zero after 2nd derivative. Hence, let us evaluate three constants C,,C, & C, ‘The generalized error constants are given my Cy = Lt Fis) ; oS FS); Cam Liar Fes) Fs)= 1 1 1 -_Oist+s ~ HS) T+G(s) 4g, 10 sts +410 Oe? +s+10 ‘s(01s +1) 0.18? +5 Ut Fs) = Lp ©) = Leow +s+ 10 1, df 01s? +s ] ont afer Hy 4 A aul = lt =04 =| (o's? +s+10)" low =o) (02s +1)-(018*~s) (0.25+1) 28410 10 ap 28410 (or%+s+t0) 10hy Time R Analysts 246 2 df 7 ] é a a} 2s+t0 | G=u =u t[ tee) 2p, 2) +10 | by arPO= Uae) sett eu 4 [ (01s? +9410)" «2—(28+10) x2 (01s? +s +10) (0.25 +1) se ds (01s? +5+10)" | 10? x 2-10% 2x 10x41 2 C20 Error signal, (9 FNC FNC, FEE = Ft}C,+0+0 =(a, +a,t) 04 +» Steady state error, @,, = Lt e(t)= Lt [(e:+e2t) 01] Method - Il anal in ¢ dome! -__ Fis) The error signal in ¢-domain, (6) =a Given that, r(t)= a + ats SE ar; C08)" spagay His)=1 On taking Laplace eaéstor of r() we get RS), 220491, 2 21 ay RO Se De ee se Bo, a » F()= BO) - 3 ets 1+G(S)H(s) s(01s+)}+10 s (ist) =2! 8 (0.1s+7) af s(0.1s+1) _s(.ts+t) ] | a) s(O1s+) ] 8 |8Qis+)+10/" s*/s spurr| we) (O.1s+7)+10} ‘The steady state errore,, can be obtained from final value theorem. Steady state error, 2, = Lt e(t)= Lt s Els) wk BS s018+1)_] , af_s(0+1) | a,[ (01841) | 0) S(03s+ 7} +10] © S*| s(018+7)+10| “s*| 01+ 9+10 | = 1p | 208(028 +1) Sines e Liatearr* - Method - Il! RE). Es) 1 & - , Eis) = 8); HS). rr signal in 6- domain, Ele)= 72S Ae) OS 10 _ Gwon that, Gls)= So yepq andiils)=1 . Eis) _ 1 __s(01s+1)_,_0.1s?+s 8 +0.18? -s_s . Ris), 10 s(0ts+)+10 Ois?+8+10 10+s+07s* 10 7000 Os =F a SOs d Dividing numerator polynomi: > Es) = 3p) Fae FO | by denominator polynomial2.47 On taking inverse Laplace transform, 1 1 t) =F -—__ F(t 850° To00" 10+s+Q.1s?| s+0.1s? , ae | e 3 Given that, 1() =a, +at+ Ser | Shiotio9 s d Ss Fe GD = ay teat 100 + 5 | | = qd. 7) = Gi =0, | : r= Fava La | “.Eror signal in tme domain, et) =Lit)= <1 (a, +2) Steady state error, Ut e(t= Lt ae +a,t) =a RESULT (@) Postion errorconstant, Kee ©) Velocity errorconstant, kK,=10 (©) Acesieration error constant, K=0 @. When input, r()=a, eat + oe Steady state error, @,, =< EXAMPLE 2.14 ‘i . i &s)__ Ks+b Consider a unity feedback system with a closed loop transfer function =~ —"**°__. Determine open loop Ris) s?+as+b transfer function Gis). Show that steady state error with unttramp Inputs glven by 2— d SOLUTION Forunity foodback system, H(s)=1 The closed loop transfer function, M(s) = C8) Ris) Gls) _ “tag MO) On cross muitiplication of the above equation we get, Gls) = M(S)I1+G(8)]= Mis) + M(s) G(s) ++ G(s)=M(s) GIS) = Mis) => Gis)[1-M(s]=Mis) => M(s) = ee <. Open loop transfer function, _ Ks+b (s? +as-+b)-(Ks +b) Ks+b Ks+b S+b “s?saseb-Ks-b s’+(a-Ks s+(aK)]hap TT Response tnalysis > 2.4% Velocity error constant, Ky = Lt s Gis)H{s)= Lts Gis) = Lt, s—“S=P Teen na a-K 1_a-Kk With velocity input, Steady state error, e, RESULT Ks+b s[s+(a-K)] Open loop transfer function, Ge) With veloctiy input, Steady state error, @, EXAMPLE 2.15 Aunty feedback system has the forward transfer function G(s) = eat = When the input rt) = 146, 8 (68+ H(I+8) Setermine the minimum value of K, so that the steady erroris less than 0.1. SOLUTION Given that, input r(t)=1+6t ‘On taking laplace transform of f(t) we get R(s). & ss 2 Rls) = L{r(t)} = £{1+61} = ‘The errorsignal in s-domain E(s) is given by, 1,6 1,6 2 F(s)2_—P@)_.__ ss * - Para Here His) W+GS)H) 4, _K(2s+1) 8 (6s+1) (148) +K(25+1) s(Ss+i(1+s) 3(58+ (I8)° - s (Ss +1) (1s) |-3f s(Ss+1)(1+8)° Tearaeske| 3 | s(Gs+1) (les) K(2s+ S| The steady state error e,, can be obtained from final value theorem. = Lt olt)= Lt s E(s) us {[__sesmve? Jef stsen0-9? _ s(s+1(1+s) +Ky2s+1) |) soe" |8| sGs+1(1+s)'+Ki2s+1)) 5° =u s(5s-+7{1+s)” ‘ &Ss+1(1+s)" = 928.8 © 290] g(6s+ Mls) 4K(2s11) 868+ (Irs) +K (2517) KK Given that, @,, < 04, o os~ & or 7 RESULT For steady state error, e,, < 0.1, the value of K, should be greater than 60.2.49 ystems Engineering 2.19 COMPONENTS OF AUTOMATIC CONTROL SYSTEM The basic components of an automatic control system are Error detector, Amplifier and Controller, Actuator (Power actuator), Plant and Sensor or Feedback system. The block diagram of an automatic control system is shown in fig 2.16. Automatic controller Reference: enor detector §=-——__—y! yc ; input [Ampiiier ana | ico} ———— Da LOuteut Contalior +[ Actuator |-+} Plan a (input ‘B Error signal Ld Ct signal) | LS Sensor or | Feedback signal Feedback system [7 Fig 2.16: Block diagram of auomatle control system. ‘The plant is the open loop system whose output is automatically controlled by closed loop system. The combined unit of error detector, amplifier and controller is called automatic controller, because without this unit the system becomes open loop system. In automatic control systems the reference input will be an input signal proportional to desired output. The feedback signal is a signal proportional to current output of the system. The error detector compares the reference input and feedback signal and if there is a difference it produces an error signal. Arvamplifier can be used to amplify the error signal and the controller modifies the error signal for. better control action. The actuator amplifies the controller output and converts to the required form of energy that is acceptable for the plant. Depending on the input to the plant, the output will change. ‘This process continues as long as there is a difference between reference input and feedback signal. Ifthe difference is zero, then there is no error signal and the output settles at the desired value. Generally, the error signal will be a weak signal and so it has to be amplified and then modified for better control action. In most of the system the controller itself amplifies the error signal and integrates or differentiates to produce a control signal (i.e., modified error signal). The different types of controllers are P, PI, PD and PID controllers. 2.20 CONTROLLERS A controller is a device introduced in the system to modify the error signal and to produce a control signal. The manner in which the controller produces the contro! signal is called the controf action. The controller modifies the transient response of the system. The electronic controllers using operational amplifiers are presented in this section. ‘The following six basic control actions are very common among industrial analog controllers. - Two-position or ON-OFF control action. .. Proportional control action. Integral control action. Proportional- plus- integral control action. Proportional-plus-derivative control action. Proportional-plus-integral-plus-derivative control action. SeayGhapler:2 “Time Response Analysis)” 2.50 Depending on the control actions provided the controllers can be classified as follows. ‘Two position or ON-OFF controllers. Proportional controllers. Integral controllers. Proportional-plus-integral controllers. Proportional-plus-derivative controllers. . Proportional-phus-integral-plus-derivative controllers. ON-OFF (OR) TWO POSITION CONTROLLER The ON-OFF or two position controller has only two fixed positions. They are either on or off. The on-off control system is very simple in construction and hence less expensive. For this reason, it is very widely used in both industrial and domestic control systems. AYR YY The ON-OFF control action may be provided by a relay. There are different types of relay. The most popular one is electromagnetic relay. It is a device which has NO (Normally Open) and NC (Normally Closed) contacts, whose opening and closing are controlled by the relay coil. When the relay coil is excited, the relay operates and the contacts change their positions (i.e., NO—> NC and NC — NO) Let the output signal from the controller be u(t) and the actuating error signal be e(1). In this controller, u(t) remains at either a maximum or minimum value. ut) =m; fore) <0 = u,; fore(t) > 0 Ri U, 8) 5), ‘ r 2 UG) Le Feedback signal Fig 2.17 : Block diagram of on-off controller. E@=Lfe)}s U()= Lf} PROPORTIONAL CONTROLLER ( P - CONTROLLER ) The proportional controller is a device that produces a control signal, u(t) proportional to the input error signal, e(t). In P-controller, u() « e(t) WO = Ke nr (2.81) where, K, = Proportional gain or constant On taking Laplace transform of equation (2.81) we get, UG) = K,E(s) (2.82) Transfer function of P - controller, ao The equation (2.82) gives the output of the P-controller for the input E(s) and equation (2.83) is the transfer function of the P-controller. The block diagram of the P-controller is shown in fig 2.18. R(s) Es) Ue) z kK Feedback signal Fig 2.18 : Block diagram of proportional conrroller.2.51 From the equation (2.82), we can conclude that ‘the proportional controller amplifies the error signal by an amount K,. Also the introduction of the controller on the system increases the loop gain by an amount K,, The increase in loop gain improves the steady state tracking accuracy, disturbance signal rejection and the relative stability and also makes the system less. sensitive to parameter variations. But increasing the gain to very large values may lead to instability of the system. The drawback in P-controller is that it leads to a constant steady state error. EXAMPLE OF ELECTRONIC P-CONTROLLER The proportional controller can be realized by an amplifier with adjustable gain. Either the non- inverting operational amplifier or the inverting operational amplifier followed by sign changer will work as a proportional controller. The op-amp proportional controller is shown iit fig 2.19 and 2.20. R VW R ww—Lf> ut) sc nn os Tnverting amplifier L + Sign changer - ¥ Fig 2.19 : Op-amp P-controller using Fig 2.20 : Op-amp P-controller using non-inverting amplifier. inverting amplifier By deriving the transfer function of the controllers shown in fig 2.11 and 2.12 and comparing with the transfer function of P-controller defined by equation (2.83), it can be shown that they work as P-controllers, ANALYSIS OF P-CONTROLLER SHOWN IN FIG 2.19 In fig 2.19, the input o(t) is applied to positive input. By symmetry of op-amp the voltage of negative input is also e(t). Also we assume an ideal op-amp so that input current is zero. Based on the above assumptions the equivalent circuit of the controller is shown in fig 2.21 By voltage division rule, R RFR, One + u®) ; M= a esa) + woe On taking Laplace transform of equation (2.84) we get, ot) S& | 2 4 (2.85) ¥ Fig 2.21 : Equivalent circuit R +R U(s)=—1— Fs) Ry - UB) _RitRe oRé of P-controller EG) R 2.86) shown in fig 2.19. The equation (2.86) is the transfer function of op-amp P-controller. On comparing equation (2.86) with equation (2.83) we get, R,+R2 1 Therefore by adjusting the values of R, and R, the value of gain, K, can be varied. Proportional gain, Ky = 2.87)Ghar Tie Repost analysis > 2.52 ANALYSIS OF P-CONTROLLER SHOWN IN FIG 2.20 The assumption made in op-amp circuit analysis are, 1. The voltages at both inputs are equal 2. The input current is zero. Based on the above assumptions, the equivalent circuit of op-amp amplifier and sign changer are shown in fig 2.22 and 2.23. ap AR BR AAA + + aR TART ules) | 2) ug) uo) ul) Fig 2.22 : Equivalent circuit of amplifier. Fig 2.23 ; Equivalent circuit of sign changer. From fig 2.22, e(t)=i,R, ; -2 oof 2.88) 1 WR (2.89) Substitute for i, from equation (2.88) in equation (2.89). sen(290) From fig 2.23, u(t) = w(2.91) u®Q=iR svo(2.92) Substitute for i, from equation (2.91) in equation (2.92). £ o (= R=-u(t) o--(2.93) On equating the equations (2.90) and (2.93) we get, Op | oR “1-2 Rp; ult) z e(t) nn(2,94) On taking Laplace transform of equation (2.94) we get, R, UGs)= R ER (2.95) Us) _ Ro Es) R The equation (2.96) is the transfer function of op-amp P-controller. On comparing equation (2.96) with equation (2.83) we get, Rp Proportional gain, K, = w(2.97) Pe gain, Kya ae Therefore by adjusting the values of R, and R, the value of gain K,, can be varied.2.53 INTEGRAL CONTROLLER (T-CONTROLLER) ‘The integral controller is a device that produces a control signal u(t) which is proportional to integral of the input error signal, e(t). In I-controller, u(t) a feo dt; (t)=K, f e(t) dt en2.98) where, K, = Integral gain or constant. On taking Laplace transform of equation (2.98) with zero initial conditions we get, Us) =K, *o wnl99) ‘ Us) _ Ky *. Transfer fu f I-controlier, [= — ransfer function of I - controller, Ee) s ssee(2.100) The equation (2.99) gives the output of the I-controller for the input E(s) and equation (2.101) is the transfer function of the I-controller. The block diagram of I-controller is shown in fig 2.24. K, | Us) at > RE) E(s) Feedback signal Fig 2.24 : Block diagram of an integral controller. The integral controller removes or reduces the steady error without the need for manual reset. Hence the I-controller is sometimes called automatic reset. The drawback in integral controller is that it may lead to oscillatory response of increasing or decreasing amplitude which is undesirable and the system may become unstable. EXAMPLE OF ELECTRONIC I-CONTROLLER The integral controller can be realized by an integrator using op-amp followed by a sign changer as shown in fig 2.25. R By deriving the transfer function of the Ly— > u(t) controller shown in fig 2.25 and comparing . ‘ + uo with the transfer function of [-controller Integrator f sigtchan defined by equation(2,101), it can be shown ag Sign changer that it work as I-controller. Fig 2.25 : I-controller using op-amp. ANALYSIS OF I-CONTROLLER SHOWN IN FIG 2.25 ” ot at) RB, . The assumptions made in op-amp circuit analysis are, 1. The voltages of both inputs are equal. 2. The input current is zero. Based on the above assumptions the equivalent circuit of op-amp integrator and sign changer are shown in fig 2,26 and 2.27. From fig 2.26, e(t)=i,R, ; vee2101) u(t) = svoe(2.102)Ci Ghapler2 Time Response Analysis 2.54 Substitute for i, from equation (2.101) in equation (2.102). Ly ot) Lips suQ=- tf Sa =-— we ql R, RG 1c (2.103) Fig 2.26 : Equivalent circuit of integrator. Fig 2.27 : Equivalent circuit of sign changer _ nu From fig 2.27, u(t)=-i,R, > RY, BW) RG) EG) UB) >| K,G+Ts) Feedback signal Beedback signal Fig 2.32 : Block diagram of PD- controller. The derivative control acts on rate of change of error and not on the actual error signal. The derivative control action is effective only during transient periods and so it does not produce’ corrective measures for any constant error. Hence the derivative controller is never used alone, but it is employed in association with proportional and integral controllers. The derivative controller does not affect the steady- State error directly but anticipates the error, initiates an early corrective action and tends to increase the stability of the system. While derivative control action has an advantage of being anticipatory it has the disadvantage that it amplifies noise signals and may cause a saturation effect in the actuator, The derivative control action is adjusted by varying the derivative time. The change in the value of K, affects both the proportional and derivative parts of control action. The derivative control is also called rate control, EXAMPLE OF ELECTRONIC PD-CONTROLLER ——S—Ssveaeaes ee CG R The PD-controller can be realized by an ys wo op-amp differentiator with gain followed by a sign | ° ‘changer as shown in fig 2.33. © > rR | | | ct uO By deriving the transfer function of the controller shown in fig 2.33 and comparing with + uo the transfer function of PD-controller defined by Differentiator Sign chanpét equation (2.124) it can be proved that the circuit with gain - ene shown in fig 2.33 will work as PD-controller. Fig 2.33 : PD controller using op-amp. ANALYSIS OF PD-CONTROLLER SHOWN IN FIG 2.33 The assumptions made in op-amp circuit analysis are, 1. The voltages at both inputs are equal. 2. The input current is zero. Based on the above assumptions the equivalent circuit of op-amp differentiator and sign changer are shown in fig 2.34 and 2.35. From fig 2.34,» 3,-2.4 ¢, &@ 9 ig eo Ay Ry +O ‘dt swun(2.125) sv 2126) On equating the equations (2.125) and (2.126) we get, id R, d Ee Recs, 2 20 =-(Ban nc, 4e0)] wn(2127)—b at o “ea | i, R, i, RB booR eR Law Ww 7 aR 7iR, ERT TiR- ane Au woe no ar) + + Fig 2.34 : Equivalent circuit of differentiator Fig 2.35 : Equivalent circuit of sign changer. From fig 2.35, u(t)=-iR 3 8 p= (2.128) u()=iR snu(2.129) Substitute for i, from equation (2.128) in equation (2.129). 2 u@=— “ R=~u(t) svnne(2.130) On equating the equations (2.127) and (2.130) we get, R. a -u(t) = Ro +R,C, se} Ray 4d ve u(t) = Re + RCO) we(2.131) On taking Laplace transform of equation (2.131) with zero initial conditions we get, us) =e E(8) + R2C, s Els) (2.132) 1 UG) “EO ‘The equation (2.133) is the transfer function of op-amp PD-controller. On comparing equation (2.133) with equation (2.124) we get, ze RCs) (2.133) Proportional gain, Ky = 7% 1 Derivative time, Ty =R,C, By varying the values of R, and R,, the value of gain K, and T, can be adjusted. PROPORTIONAL PLUS INTEGRAL PLUS DERIVATIVE CONTROLLER (PID-CONTROLLER) ‘The P1D-controller produces an output signal consisting of three terms : one proportional to error signal, another one proportional to integral of error signal and the third one proportional to derivative of error signal.2.59 In PID-controller, u(t) [tty fet a+ fa] L at k, d nm( 21341 fu) = Ky e+ fein KT Sed ‘ where, K, = Proportional gain T, = Integral time T, = Derivative time On taking Laplace transform of equation (2.134) with zero inital conditions we get, K, UG) = K, E46) +2 2 FO, Ts Bb) «(20135) 2. Transfer finction of PID -controtter, US » (1-13) Feedback signal The proportional controller stabilizes the gain but produces a steady state error. The integral controller reduces or eliminates the steady state error. The derivative controller reduces the rate of change of error. EXAMPLE OF ELECTRONIC PID-CONTROLLER The PED-controller can be realized by op-amp amplifier with integral and derivative action followed by sign changer as shown in et) G RG R R fig 2.37. a P| By deriving the transfer function of the - ‘ + uo controller shown in fig (2.37) and comparing with | Amplifier + | Sign changer the transfer function of PID-controller defined + by equation (2.136) it can be proved that the Fig 2.37: PID- controller using op-amp. circuit shown in fig 2.37. work as PID-controtler. ANALYSTS OF PID-CONTROLLER SHOWN IN FIG The assumptions made in op-amp circuit analysis are. 1. The voltages of both inputs are equal. 2, The input current is zero.2.60 Based on the above assumptions the equivalent circuit of op-amp amplifier and sign changer are shown in fig 2.38 and 2.39. cs 4 G be rR eR ie, iy ia +aRO ik Q oe 2 uC) WO ut OF fo o we ¥ ¥ Fig 2.38 : Equivalent circuit of amplifier. Fig 2.39 : Equivalent circuit of sign changer. . _ et) de(t) From fig 2.38; Rt (2.137) On taking Laplace transform of equation (2.137) with zero initial conditions we get, l= & E(s)+C, sE(s) 1,6)= (Eres) Es) 138) From fig 2.38, iR.+ aie a (2.139) " On taking Laplace transform of equation (2.138) with zero initial conditions we get. 1@)R) +> 20. -U,) “1® (k, ae (8) Substitute for 1,(s) from equation (2.138) in equation (2.140). wf{ eos BG) R.+—-|=-U6) ae 2+ lu or {2+ +e + ca + Recs) F(8) = U,(s) 141) From fig 2.39, u()=-i,R j= -2 seve (2.142) u()=i,R svne(2143) Substitute for i, from equation (2.142) in equation (2.143). u()= ~HOr- Rt (2.144) On taking Laplace “ransform of equation (2.144) we get, WSU (2.145)2.61 From equations (2.142) and (2.146) we get, RG, 1 UGSs)= (&- ct RCs + Race) E(s) ~ Us) _f Rx, + RC, L 1 WO(BERE sn RC, RCs aR (RGR G 1 ("Ee Res Rea) eae (2.146) ‘The equation (2.146) is the transfer function of op-amp PID-controller. On comparing equation (2.146) with equation (2.136) we get, Proportional gain, Ki = = 1 Derivative time, Tj =R,C,; Integral time, Tj = R,C, Also, RC. aC By varying the values of R, and R, the values of K,, T, and T, are adjusted. 2.21 RESPONSE WITH P, PI, PD AND PID CONTROLLERS In feedback contro! systems a controller may be introduced to modify the error signal and to achieve better control action. The introduction of controllers will modify the transient response and the steady state error of the system. The effects due to introduction of P, PI, PD and PID controllers are discussed in this section. EFFECT OF PROPORTIONAL CONTROLLER (P-CONTROLLER) The proportional controller produces an output signal which is proportional to error signal. The transfer function of proportional controller is given below. (Refer equation 2.83). UG) Transfer function of P-controller, a> = Kp The term K, in the transfer function of proportional controller is called the gain of the controller. Hence the proportional controller amplifies the error signal and increases the loop gain of the system. The following aspects of system behaviour are improved by increasing loop gain. 3 Steady state tracking accuracy. % Disturbance signal rejection. + Relative stability. In addition to increase in loop gain it decreases the sensitivity of the system to parameter variations. The drawback in proportionai control action is that it produces a constant steady state error. EFFECT OF PI-CONTROLLER The proportional plus integral controller (PI-controtler) produces an output signal consisting of two terms : one proportional to error signal and the other proportional to the integral of error signal. 1 Ts+1) ‘Transfer function of Pl-controller, G.(s) = K, (+4)-s(2 | Refer equation 2.113) where, K, is proportional gain and, T, is integral time.© Ghipler d= Time Response Analysis _) 2.62 ‘The block diagram of unity feedback system with Pl-cortroller is shown in fig 2.40. 3) | c® iO) RG) 3) — Kol Ace Ts) Fig 2.40 : Block diagram of feedback system with PI-controller. Let the open loop transfer function G(s) be a second order system with transfer function, as. shown in equation (2.148). Open loop transfer function, G(s) ~ en 2147) (+2%0,) Now, loop transfer function = G,(s) G(s) His) = G,(s) Gs) [HO x {itts) o _ Koh +s) a Ts )* s6+20,) 9hs+2Zo,) 2148) Now the closed loop transfer function is given by, K,o,(1+Ts)_ Cis) __ G(SG(s)__ FT (8+2KO,) = Ko,(+Ts) R@). 146,966) K,o\(1+Ts) FE(6+20 @,)+K,02 (+s) STF, (8+ 2%a,) - (1+Ts) +K,03 T, s+K,o KT) O04 T9. 4 2 25, Np K, #+260,9' +K,03 s+Poz K,=-2 T mee = Keg +ts) swn(2.149) s° +260," +K,o2s+ Ko? From the closed loop transfer function (equation (3.149) it is observed that the Pl-controller introduces a zero in the system and increases the order by one. The increase in the order of the system results in a less stable system than the original one because higher order systems are less stable than lower order systems. From the loop transfer function (equation (3.148)) it is observed that the PI-controller increase the type number by one. The increase in type number results in reducing the steady state error. For example if the steady state error of the original system is constant, then the integral controller will reduce the error io zero. EFFECT OF PD-CONTROLLER The proportional plus derivative controlier produces an output signal consisting of two terms : one proportional to error signal and the other proportional to the derivative of error signal. The transfer function of PD - controller, G(s) = K, (1+T, s) (Refer equation 2.124) where K,, is Proportional gain, T, is Derivative time.2.8 The block diagram of unity feedback system with PD-controller is shown in fig 2.41. ce G6.) =K,(1+Ts) Fig 2.41 : Block diagram of feedback system with PD-controller. Let the open loop transfer function G(s) be a second order system with transfer function as shown in equation (2.150). Open loop transfer function, G(s)= sne(3.150) 5 Now, loop transfer function = G,(s) G(s) H® = G.(s) Gs) [HGE)=1 o K,o; (1+Tys) = K,(1+Ty()) x Re el {2.151} . 8(+2f0,) 8+260,) Now the closed loop transfer function is given by, K,@2(1+Tys) CG) __ GANGS) ___ 8G F2E0,) R(s) 1+G(s)G,(s) Ss oo (1+Ts) 8(8+250,) = K,o5 (1+ Ts) s(s+24w,)+K,; (1+ Tys) _ K,o3 +Ts) # +260,8+K,o, +K,orTs = Kos +18) s'+(2o, +K,021,) s+K,o3 Ky=Kply o3 Ky +Kes) + QC, +Kyo2) s+K,o2 From the closed loop transfer function (equation (2.152)) it is observed that the PD-controlle: introduces a zero in the system and increases the damping ratio. The addition of the zero may increase the peak overshoot and reduce the rise time. But the effect of increased damping ultimately reduces the peak overshoot. From the ioop transfer function (equation (2.151)) it is observed that the PD-controller does not modify the type number of the system. Hence PD-controller will not act modify steady state etror. EFFECT OF PID-CONTROLLER A suitable combination of the three basic modes : proportional, integral and derivative (PID) can improve all aspects of the system performance. The proportional controller sizbilizes the gain but produces a steady state error, The integral controller reduces or eliminates the steady state error. The derivative controller reduces the rate of change of error. The combined effect of all the three cannot be judged from the parameters K,, K, and Ky2.64 2.22 TIME RESPONSE ANALYSIS USING MATLAB In general, the closed loop transfer function of a system is denoted as M(s). Let, M(s) be a rational function of "s", as shown below. dys +b 5 ays +a,s% + a,s' bys, Wy, M(s For time response analysis, the coefficients of the numerator and denominator polynomials are declared as two arrays as shown below. num_cof = [60 bl b2 .......- BMI; dencof = [a0 al a2 vies an]; UNIT STEP RESPONSE To compute step response ‘The unit step response can be computed and displayed using following commands. syms § complex; R = 1/5; M = (bO*SAM+b1*SA(M-1)+. .. +bM)/(a0*SANGA1*SA(N-L)+...42N) 5 S = RM; disp(‘unit step response of the system is,*); step_res = ilaplace(s) To plot step response Method 1 : The unit step response can be plotted using the following command. step(num_cof, den_cof); Method 2: The unit step response of the system can be plotted using the following commands. [te = tstart : tstep : tend ; |e = step(numcof, den_cof,t); [plot(t,c, ‘k')} | where, ¢ is an array where the values of response are stored. | The unit step response can be computed "n" times by varying some parameter of the system (coefficient / damping ratio / natural frequency of oscillation) using the following commands. t = tstart ; tstep : tend ; fori=ili:n cCl:k, 1) = stepCnumcof, den_cof,t); end plot(t,c,'k'); where, c is an array where the values of response are stored. k is the number of samples of response to be computed,2.6 kaniral sFsiems Kgineering >) Method 3 : ‘The unit step response of the system can be plotted using the following commands. s = thts"); M = (bO*SAMSbL*SA(M~1)+,. .4bM)/(a0*SANtal#SACN-1)+...4aN)5 t= t.start : tstep : tend ; sr = step(M,t); plot(t,sr,"k"); IMPULSE RESPONSE To compute impuise response The impulse response can be computed and displayed using following commands. syms 5 complex; M. = (bO*sAm+b1*sA(M-1)+...+bM)/(a0*sANtal*sA(N-1)+...4aN); disp(‘unit step response of the system is,’); imp_res = ilaplace(s) To plot impulse response Method 1 : ‘The impulse response can be plotted using the following command. ‘impuise(num_cof, den_cof); Method 2: The impulse response of the system can be plotted using the following commands. | t = t.start : tstep : tend ; m = imputse(numcof, den_cof,t); plot(tsm,"k")5 I where, m is an array where the values of impulse response are stored, Method 3 : The impulse response of the system can be plotted using the following commands. 5 = TFC's); | M = (bO*SAMSbI*SA(M-1)+,..+bM)/(Q0*SANSAL4SA(N-D) 4.4. 40ND t= tstart : tstep : tend ; imp = impulse(m,t); plotct, imp, 'k'); RESPONSE FOR ARBITRARY INPUT The response of a system for an arbitrary input, x(t) can be plotted using the following commands. t tstart : t.step : tlend ; c = Lsim(numcof, dencof, r, t); plot(t,c,'k'); where, c is an array where the values of response are stored.2.66 >ROGRAM 2.1 Consider the standard closed loop transfer function of the second order system given below. M(s)=0,2/(st#26o,s+0,2) write @ MATLAB program to find the unit step response for various values of damping ratio,¢. Take, natural frequency of oscillation, @,=1 rad/sec. Kunit step response for various values of damping ratio, zeta. Xthe natural frequency of oscillation, wn=1. ele t=0:0,2:125 Mspecify a time vector cezeros (61,6); %initialize response array as zero zeta=[0 0.2 0.4 0.6 0.8 1]; %store zeta as an array for n=l %for loop to compute c(t) 6 times numcof=[0 0 1); den_cof=[1 2*zeta(n) 1]; c(1:61,n)=step(num_cof,den_cof,t); end plot(t,c,"k'); grid xlabel(‘time,t in sec’); ylabel(‘unit step response,c(t)"); text(2.8,1-86,"\zeta=0") text(2.8,1.58,*\zeta=0.2') text(2.8,1.30,\zeta=0.4") text(2.8,1.12,"\zeta=0.6') text(2.8,0.95,’\zeta=0.8") text(2.8,0.72,’\zeta=1.0") OUTPUT The output waveforms are shown in fig p2.2. Unit step response, o(t) Q z 4 @ 8 10 12 time, in see Fig P2.1.: Unit step response off second order system for various values of damping-ratio.2.07 PROGRAM 2.2 Consider the standard closed loop transfer function of the second order system given below. M(s)=0,2/(s?4260,5+0,2) Write a MATLAB program to find the unit step response for various values of hatural frequency of oscillation, o,. Take, damping ratio,=0.4. Kunit step response for various natural frequency of oscillation,wn. XThe damping ratio, zeta=0.4. clc t=0:0.1:85 Xspecify a time vector wns[1 2 4 6]; “store wn as an array zeta=0.4; c=zeros(81,4); Xinitialize the response array as zeros for i=1:4; %for loop to compute c(t) 4 times b2=wn (i) *wn Cid; al=2*zeta*wn(i 0 0 b2] 1 al 62]; ¢(1:81,i)=step(num_cof,den_cof,t); end plot(t,cC:,1),*--k’,t,cC:,2),'xk’,t,c€:,3),%-k",t,0€:,4),7-.k7)5 grid; xlabel(‘time,t in sec’); ylabelC‘unit step response,c(t)’); text(4.25,1.25,'wn=1') text(1.5,1.30, text(0.7,1.30, text(0.1,1.25, OUTPUT The output waveforms are shown in Fig p2.2. Unit step response, oft) times in seo Fig P2.2 : Unit step response of second order system for various values of natural frequency of oscillation.(@ Ghapter2-Time Response Analysis) PROGRAM 2.3 write @ MATLAB program to find impulse response of the following systems. a) M,(s)=(2s+1)/(s#1)? b) Xerogram to find impulse cle syms s complex; M1=(2*s41)/((s+1) 42); disp(‘tmpulse response of m1=ilaplace(m1) M2=s/(s+1); disp(‘Impulse response of m2=ilaplace(m2) M3=1/(sA2+1)3 disp(‘Impulse response of m3=i laplace(m3) satf(‘s’)+ M1=(2*s+1)/((s+1)A2); M2=s/(s+1); M3=1/(SA241); 005:10; mlsimpulse(M1,t); m2=impulse(m2,t); m3-impulse(m3,t); plot(t,ml,’--k’,t,m2,* xlabel(*time,t in sec’ ylabel (‘Impulse text(0.4,1230, ’mi(t)’*) text(0.3,-0.30, ’m2(t)*) text(2.2,0.90,’m3(t)*) E z a 3 3 a E M,(s)=s/(s+1) ©) M(s)=1/(s*+1) response the systeml is,"); is.'); the system2 the system3 »t,m3,’-k’)igrid responses,mi(t),m2(t).m3(t)'); time,t in seo Fig P2.3 : Impulse response of systems given in program 2.3, 2.08Impulse response of the systeml 7s, mi = (2-0) *exp(-t) Impulse response of the system2 is, a2 = dirac(t)-exp(-t) Impulse response of the system? is, m= sinct) The output waveforms are shown _in_fig p2.3. PROGRAM 2.4 write a MATLAB program to find unit step response of the following systems. a) M,(5)=4/(s%5s+4) b) M,(s)=100/(s’+12s+100) ©) ™,(s)=600/(s?+70s+600) Sprogram to find unit step response clc syms s complex; R=1/s; %Laplace of unit step input M1=4/(5A2+5*5+4)5 sler*m1; %s-domain unit step response of systemi disp(‘unit step response of the system1 is,"); sl=iTaplace($1) Ktime domain unit step response of system M2=100/(sA2+12*s+100) ; S2=R*N2; Xs-domain unit step response of systemZ disp(‘unit step response of the system2 is,"); s2=iTaplace(s2) Xtime domain unit step response of system2 M3=600/(5A2+#70*5+600) ; S3=R*M3; %s-domain unit step response of systen3 disp(‘unit step response of the system3 is,'); s3=i Taplace(s3) Xtime domain unit step response of system3 se=tf C's"); Mi=4/(sA2+5*s+4); M2=100/(sA2+12*5+100) ; M3=600/(SA2+70*$+600) ; s2=step(m2,t); s3=step(m3,t)i plot(t,s1,*--k’,t,s2, xlabel('time,t in sec’) ylabel(‘unit. step responses,s1(t),s2(t),s3(t)")i text(2.2,0.85,7s1(t)") text(0.2,1.15,"s2(t)") text(0.5,0.95,’'s3(t)") k',t,s3,'-k')igrid2.70 OUTPUT unit step response of the systeml is, ae L/3*exp(-48t)41-4/F*expC-t) unit step response of the system2 is, ee t-exp(-64t) *c0s (8*t)-3/4*exp(-6*t) *s7n(8*t) unit step response of the system3 is, oe I+1/5*exp(-60*t)-6/5*exp(-10*0) The output waveform is shown in Fig p2.4. 14.-— Unit step responses, s i(t).s2tt). 6&0) $6 8 1. 8 ° 1 2 @ 4 6 6 7 B § times insoo Fig P24 : Unit step response of systems given m program 2.4. PROGRAM 2.5 Consider the closed loop transfer function of the following second order system, M(s)=16/(s*+4s+16) Write a MATLAB program to find the rise time, peak time, maximum’ peak overshoot, and settling time from the unit step response of the system. %set time vector numcof=[0 0 161; %store the numerator coefficients as an array den_cof=[1 4 16]; %store denominator coefficients as an array [c,x,t]=step(num_cof,den_cof,t); 3 %initialize count as 1 while c(n)<1.0001; %count the time index as along as c(t)<1 nent; end;2.71 rise_time=(n-1)*0.005 %rise time=(count-1)*time interval Comax,tpJ=max(c); %determine maximum value of c(t) & %corresponding time peak_time=(tp-1)*0.005 %peak time=(tp-1)*time interval max_overshoot=cmax-1 %compute peak overshoot n=1001; Kinivialize count as (5/.005)+1=1001 while c(n)>0.95&(n)<1.05; nzn-1; Keount time index between c(t)>0.958c(t)<1-05 end: settling_time_Sper_err=(n-1)*0.005 n=1001; Kinitialize count as (5/.005)+1=1001 while c(n)>0.98 & c(n)<1.02; %count time index between c(t)>0.98&c(t)<1-02 n=n-: end; settling_time_2per_err=(n-1)*0.005 OUTPUT rise_time = 0.6050 peak_time = 0.9050 max_overshoot = 0. 1630 settling time_Spererr = 1.3200 settling time_2per_err = 2.0150 PROGRAM 2.6 Consider the closed loop transfer function of the following second order system, M(s)=64/(s’+8s+64) Write a MATLAB program to find the response for unit step, unit ramp and unit parabolic input signals. wunit step/ramp/parabolic response cle num_cof=[0 0 64]; den_cof=[1 8 64]; Xunit ramp input signal Xunit parabolic input signal cl=step(numcof, den_cof,t); c2=Lsim(num_cof, den_cof,ri,t); c3=Lsim(num_cof, den_cof,r2,t);2.72 plot(t,cl,*--k’,t,¢2," xlabel(‘time,t in sec’): ylabel(‘Responses,ci(t),c2(t),c3(t)"); text(0.25,1.15,*c1(t)') text(1-45,1.5,’e2¢t)’) text(1.35,0.7,'c3(t)") k?,t,e3,'-k’); grid OUTPUT The output waveform is shown in fig p2.6. Responses, o7(t).02(), 234) £55 z 0 0. 02 of 06 o8 1 12 id 16 18 2 time, in seo Fig P2.6 : Step, ramp and parabolic response of system given in program 2.6. PROGRAM 2.7 Consider the closed loop transfer function of the following second order system, M(s)=5/(s*+5+5) write a MATLA8 program to find the response for the input signal, r(t)=2-2t+t’. Xprogram to find response for given input ele num_cof=[0 0 51; den_cof=[1 1 51; t20:0.005:3; %specify a time vector re2-24t4t.A2; %input signal c=Lsim(num_cof,den_cof,r,t); %compute response using tsim function plot(t,r,’--k’,t,c,'-.k'); grid xlabel(‘time,t in sec’): ylabelC‘iInput,r(t) and output,c(t)’); text(0.25,1.65,’r(t)") text(0.25,0.6,’¢(t)")2.3 Control OUTPUT The output waveform is shown in Fig p2.7. 3 z 5 = E ° 05 1 18 2 26 a time. in seo Fig P2.7 : Input and Output of the system given in program 2.7. 2.23 SHORT QUESTIONS AND ANSWERS 2.1. What is time response? The time response is the output of the closed loop system as a function of time. It is denoted by c(t). Itis given by inverse Laplace of the product of input and transfer function of the system. Ss) ___ ls) ‘The closed loop transfer function, =" Tasne . = RS) Gis)_ Response in s-domain, C(s) Ge) HS =f — £1] Ris) Gis) Response in time domain, o(t)=£YC(s)}= £ [Bese 02.2. What is transient and steaily state response? The transient response is the response of the system when the input changes from one state to another. The response of the system as t + «is called steady state response. 02.3 What is the importance of test signals? The test signals can be easily generated in test laboratories and the characteristics of test signals resembles, the characteristics of actual input signals. The test signals are used to predetermine the performance of the system. Ifthe response of a systems satisfactory for a test signal, then the system will be suitable for practical applications. 2.4 Name the test signals used in control system. The commonly used test input signals in control system are impulse, Step, Ramp, Acceleration and Sinusoidal signals.2.74 02.6 02.7 02.8 02.10 211 92.12 02.13 2.14 Define step signal. The step signal is a signal whose value changes from 0 to A and remains Wa constant at A for t> 0. The mathematical representation of step signal Is, HQ)=A,t20 =0,t<0 Define ramp signal. > A ramp signal is a signal whose value increases linearly with time from Fig Q2.5 : Step signal. an initial value of zero at t = 0. Mathematical representation of ramp signal is, r= At, t=O =0,t<0 Define parabolic signal. Itis a signal in which the instantaneous vaiue varies as square of the time from an initial value of zero at t= 0.The mathematical representation of parabolic signal is, > a Fig 02.6 : Ramp signal. y= t20 = 20. , ts oO What is weighing function? The impulse response of system is called weighing function. It is given by inverse Laplace transform of system transfer function. What is an impulse signal? : A signal which is available for very short duration is called impulse ! signal. Ideal impulse signal is a unit impulse signal which is defined as a a signal having zero values at all ime exceptatt=0.Att=Othe magnitude “rig Q2.4 : Parabolic signal becomes infinite. Its denoted by 8(t) and mathematically expressed as, s(t) = Tay at ao: and f ()dt=1 DAP fh Define pole. The pole of a function, F(s) is the value at which the function, F(s) becomes infinite, where F(s) isa function of complex variable s. Define zero. The zero of a function, F(6) is the value at which the function, F(s) becomes zero, where Fis) is.a function of complex variable s. What is the order of a system? The order of the system is given by the order of the differential equation governing the system. It is also given by the maximum power of s in the denominator polynomial of transfer function. The maximum power of s also gives the number of poles of the system and so the order of the system is also given by number of poles of the transfer function. Define damping ratio. ‘The damping ratio is defined as the ratio of actual damping to critical damping. Give the expression for damping ratio of mechanical and electrieal system. B The damping ratio of second order mechanical transtational system, Tho damping ratio of second order mechanical rotational system, ¢=—F=Manito Systemsiinggneering > 2.75 , R The damping ratio of second order electrical system, c= ping system, G= = 02.15 How the system is classified depending on the value of damping? Depending on the value of damping, the system can be classified into the following four cases. Case 1: Undamped system, ¢ = 0 Case 2 : Underdamped system, 0<¢<1 Case 3 : Critically damped system, ¢ = Case 4 : Over damped system, £>1 02.16. Sketch the response of a second order under damped system. cOa eth pg: |, 1 o-3 . > es Allowable error 7 2% or 5% ost fi: : oft et a? Fig Q2.16 : Response of under damped second order system. 02.17 What will be the nature of response of a second order system with different types of damping? For undamped system the response is oscillatory, For underdamped system the response is damped oscillatory. For critically damped system the response is exponentially rising. For overdamped system the response is exponentially rising but the rise time will be very large. Q2.18.. What is damped frequency of oscillation? In underdamped systom the response is damped oscillatory. The frequency of damped oscillation is given by, ag-,/1-2 Q2.19. Give the expression for natural frequency of oscillations of electrical and mechanical system. The natural frequency of oscillation of (a second order mechanical translational system| °° YM Tho natural frequency of esclaton of || IK second order mechanical rotational system) “" YJ The natural frequency of oscillation of] 1 second order electrical system nc 20) 10 | 2.20. The closed loop transfer function of second order system is ergy GF gy a7 + What is the spe of damping in the system?. Lot us compare the given transfer function with the standard form of second order transfer function of 10 20,8+02 s*+65+10 Zerg =6 Since ¢ <1, the system is underdamped.Time Re: 2.76 02.21 2.22 2.23 02.24 92,25 82.26 02.27 92.28 200 57+ 205 209° Petermine The closed loop transfer function of a second order system ts given by the damping ratio and natural frequency of oscillation. Let us compare the given transfer function with the standard form of second order transfer function Os) oh 200 TRIS) $?7+2t0,s+0% s*+20s+200 02 = 200 Lon = 20 20 20 = 200 =14.14 red ou en = 4200 redisec Pxon 2x1814 =0.707 Damping ratio, 6 = 0.707 Natural frequency of oscillation, @, = 14.14 rad/sec. A second order system has a damping ratio of 0.6 and natural frequency of oscillation is 10 rad/sec. Determine the damped frequency of oscillation. Damped frequency of oscillation, 04 =opy1-¢2 = 1Oy1-(06)? = 10x0.8=8 rad/sec 20 The open loop transfer function of aunty feedback system is Gis) = 2°. Whats the nature of sie response of closed loop system for unit step input. The closed loop transfer function, Se) __Gs)__ Bisis+10)__—20_ Rs) Gis) 4, _ 20 9(8#10)#20 $?+10s+20 e+ 10) The standard form of second order transfer function is, C8) ea Winossot RG) = 2lo,s+0F (On comparing system transfer function with standard form of second order transfer function we gel. 02 =20 20, =10 10 10 y= ¥20 = 447 rads att 420 sec Sao, bead? . Since damping ratio, ¢ >1, the system is overdamped and the response will be exponentially rising. List the time domain specifications. The time domain specifications are, () Delay time (i). Rise time (iii) Peak time (v) Maximum overshoot (v) Settling time. Define delay time. tt is the time taken for response to reach 50% of the final value, the very first time. Define rise time. Itis the time taken for response to raise from 0 to 100% , the vary first timo. For underdamped system, the rise time is calculated from 0 to 100%. But for overdamped system it is the time taken by the response to raise from 10% to 90%. For critically damped system, it is the time taken for response to raise from 5% to 95%. Define peak time. Itis the time taken for the response to reach the peak value, the very first time (or) Itis the time taken for the response to reach peak overshoot, M,. Define peak overshoot. It is defined as the ratio of the maximum peak value to final value, where maximum peak value is measured from final vatue. Alta) —e =) Let final value = c(@), Maximum value = c(t.) «. Peak overshoot, M, =2.77 02,29 02.30 02.31 02.32 02.33 92.34 02.35 02.36 02.37 Define settling time. itis defined as the time taken by the response to reach and stay within a specified error and the error is usually specified as % of final value. The usual tolerable error is 2% or 5% of the final value. The damping ratio of a system is 0.75 and the natural frequency of oscillation is 12 rad/sec. Determine the peak overshoot and the peak time. ars. Peak overshoot, M, 028; =, %M, =0.028x100=2.8% Damped frequency of oscillation, ag =any1-¢7 =12/1- (0.75) =7.94 rad/ sec Peak time, t, —_- 0396 sek time, ty = X= see sec The damping ratio of system is 0.6 and the natural frequency of oscillation is 8 rad/sec. Determine the rise time, Rise time, t, = 2=2 Og a = ton 208 sige = @g=Ony1-£? 28) 1-(06)? = 64 rad/sec 927 @=tan* xntad = 0.927 rad nl Rise time, t, = = 034 sec What is type number of a system? What is its significance? The type number is given by number of poles of loop transfer function at the origin, The type number of the system decides the steady state error. Distinguish between type and order of a system. ()) Type number is specified for loop transfer function but order can be specified for any transfer function. (apen loop or closed ioop transfer function), (ii) The type number is given by number of poles of loop transfer function lying at origin of s-plane but the order is given by the number of poles of transfer function, For the system with following transfer function, determine type and order of the system. K Gam 2065+) 5(5 +1) (5? + 6s +8) ®) 69) HO) = Sey +05 G(s) H6) = ay oe) He) = (©) 60) A)= sa 6-2) +025) (5? +254) Ans: (i) Type - 1, order-4 (i) Type -2, order-4 (i) Type -0, order -2 (v) ‘Type -3, order-5. What is steady state error? The steady state error is the value of error signal et), when t tends to infinity . The steady state error is a measure of system accuracy. These errors erise from the nature of inputs, type of system and from non-linearity of system components. What are static error constants? The K,, K, and K, are called static error constants. These constants are associated with steady state error in @ particular type of system and for a standard input Define positional error constant. The positional error constant K, = 1, Gls) H(a) .The steady state error in type-O system when the input is unit step is given byt/1+K,,Chapter 2=Fime Response Analy: > 2.78 92.39 02.40 dn44 Das Define velocity error constant. The velocity error constant K, = Lt, S46) His). The steady state error in type-1 system for unit ramp input is given by 1/K,. Define acceleration error constant. The acceleration error constant K, = Lt_ s* G(s) His). The steady state error in type-2 system for unit parabolic input is given by 1K, A unity feedback system has a open loop transfer function of G(s) =——2—_.. Determine the G+ DSt+Q steady state error for unit step input. The steady state error for unit step input, e55 = For unity feedback system H(s) = 1. 10 530 (61) (5*2) where, Kp = Lt G(s) H(s) = Ly G)= Lt and Coe = A unity feedback system has a open loop transfer function of G(s) =—22-S*9 _.. Determine S(8+0.5) (8+ 2) the steady state error for unit ramp input. The steady state efror for unit ramp inputis, @ss system H(s) = where, K, = at, § GIS) Hs). For unity feedback . . 25(s+4)_] _ 25x4_ stot. Kee e510 us, o[ -2e-8) Denn 100 ANd ss — Fy_~ OOF ; 5 20(5+5) A unity feedback system has a open loop transfer function of G(s)= —~—"""__ . Determine S(5+0.1) (5+ 3) the steady state error for parabolic input. 1 ‘The steady state error for unit ramp input is eqs = 7 feedback system H(s) = 1. where, K,= Lt s?G(s) H(s) . For unity Pas _ 20(s+5) | 20x5 _100_ oto ae 5 [pees Oates 93 738953 and Os = = saR Gy = 1.008 What are generalized error coefficients? They are the coefficients of generalized error series. The generalized error series is given by, Saran . any O35 et) = corto +Cf)+ iy + B70) + The coefficients C,, C,, C,...uC, are called generalized error coefficients or dynamic error coefficients. a 1 The n® coefficient, Cy = Lt Fis) , where, Fis)= ee ay Glve the relation between generalized and static error coefficients. The following expression shows the relation between generalized and static error coefficient. 1, 1, Mention two advantages of generalized error constants over static error constants. (i) Generalized error series gives error signal as a function of time. (ii) Using generalized error constants the steady state error can be determined for any type of input but static error constants are used to determine steady state error when the input is anyone of the standard input.iain Systems kginsering 279 02:46. What are the basic components of an auiomattc control system ? The basic components of an automatic control system are, 1. Error detector 4. Plant 2, Amplifier and controller 5. Sensor or feedback system 3. Actuator (Power actuator) 02.47 What is automatic controller ? The combined unit of error detector, amplifier and controller is called automatic controller. 02.48 — What is the need for a controller? The controller is provided to modify the error signal for better control action. 02.49 — What are the different types of controllers? The different types of controller used in control system are P, Pl, PD and PID controllers. 02.50 What is Proportional controller and what are its advantages? ‘The Proportional controller is a device that produces @ contro! signal which is proportional {0 th input error signal. The advantages in the proportional controller are improvement in steady-state tracking accuracy disturbance signal rejection and the relative stability. italso makes a system less sensitive to paramete variations. 02.51 What is the drawback in P-controller? The drawback in P-controlier is that it devolop a constant steady-state error. 02.52 What is integral control action? in integral control action, the contro signal is proportional to integral of error signal. 02.53 What is the advantage and disadvantage in integral controller? The advantage in Integral controller is that it eliminates or reduces the steady-state error. Th disadvantage is that it can make a system unstable. 02.54 Write the transfer function of P, PI, PD and PID controllers, The transfer function of P-controlier, aoe K, :where, K, = Proportional gain. uUls) 1) The transfer function of Pl-controlier, a ‘o5) j where, T, = Integral.time constant. The transfer function of PD-controller, Boo =K,(1+T.s) _ : where, T, = Derivative time constant - US) 1 } The transfer function of PID-controiler, Bap teaeets | §) f 2.55 What is Reset rate? ‘ The Reset rate is the reciprocal of integral time or reset time. The reset rate is the number of time per minute that the proportional part of the control action is duplicated and it is measured in term of repeats/minute. 02.56 Why derivative control is not employed in isolation? A derivative control mode in isolation produces no corrective efforts for any constant errors. Because it acts only on rate of change of error. 02.57 — What is Pl-controiler? The Pl-controller is a device which produces a control signal consisting of two terms : one proportions 0 error signal and the other proportional to the integral of error signal._ GapierTTema Ri piso Analysis) 2.80 92.58 92.59 02.60 02.61 02.62 02.63 02.64 What is PD-controller? The PD-controller is @ device which produces a control signal consisting of two terms : one proportional to error signal and the other proportional to the derivative of error signal. What is PID-Controlier? The PID-controlier is a device which produces a contro! signal Consisting of three terms : one Proportional to error signal, another one proportional to integral of error signal and the third one proportional to derivative of error signal. Give an example of electronic PlD-controller The electronic PiD-controller can be realized by an op-amp amplifier with integral and derivative action followed by sign changer, as shown in figure Q2.60. rot R e(t) fey R rw SLA R, ud) Amplifier : ‘Sign changer Fig 02.60 Sketch the step response of a P and Pl-controller ? Let e(!) be the input signal to the controller and u(t) be the output signal to the controller. The input and output signals are shown in the figure 02.61. ) Unit step ut a! 1 : ox Pi Control action a P control action Fig Q2.61 ‘Sketch the ramp response of P, PD and PlD-controller? Let e(t) be the input signal to the controller and u(t) be the output signal to the controller. The input and output signals are shown in the figure Q2.62. PID, gon action a(t). ua ion we ack Pp con ee af es ue so Proportional control action 0 Fig Q2.62 What is the effect on system performance when a proportional controller is introduced in a system? The proportional controller improves the steady-state tracking accuracy, disturbance signal rejection and relative stability of the system. It also increases the loop gain of the system which results in reducing the sensitivity of the system to parameter variations. What is the disadvantage in proportional controller? The disadvantage in proportional controller is that it produces a constant steady state error.Caneel Systems Hngincering > 2.81 02.65 — What is the effect of PI-controlter on the system performance? . The PI - controller increases the order of the system by one, which results in reducing, the steady state error. But the system becomes less stable than the original system. 02.66 What is the effect of PD-controller on the system performance? The éffect of PD - controller is to increase the damping ratio of the system and so the peak overshoot, is reduced. 02.67. Why derivative controller is not used in control systems? The dorivative controller produces 6 control action based on rate af change of error signal and it does Not produce corrective measures for any constant error. Hence derivative controller is not used in control systems. 02.68 — Determine the impulse response of the feedback system governed by the closed loop transfer 2s+1 function, M(s)= . a s+ 1) By partial fraction expansion the given closed loop transfer function can be expressed as, 2s41_ A iB Ms) = S45 = 4 FO Ms) (s+* (s+7* s+1 = 2 Mis) = ar M8) = Coa ead The impulse response is given by inverse Laplace transform of closed loop transfer function. 2 -. Impulse response, m(t) = £7" +—=}= -te'+2e7 pl P vt) ter 3} ev +2e 22.69 Determine the impulse response of the feedback systems governed by the following closed loop transfer functions, 2) The impulse response is given by inverse Laplace transform of closed loop transfer function. Impulse response, m(t) =~! f S (s+1 b) The impulse response is given by inverse Laplace transform of closed loop transfer function. 1 :.tmpul t) = 2 } = sint Impulse response, m(t) way 92.70 Determine the impulse response of the feedback system governed by the closed loop transfer 2(¢+3) Ge3P4T ‘The impulse response is given by inverse Laplace transform of closed loop transfer function. a {_2s+3) earl function, Mis) : Impulse response, m(t) = = 2e* cost( Ghapier.2 - Time Response Analysis _) 2.8 EXERCISES E23 F210 What is the unit-step response of the system shown in figE2.1 RG) 10 Cs) ” Fig E21. Obiain the unit-step response of a unity-feedback system S(s+20) : ion is (3) =$———— whose open-loop transfer function is We 450) (4341541635) 100 s(8+2).6+ 5) For unit step input,find the time response of the closed loop system and determine % over shoot and the rise time. ‘The open loop transfer function of an unity feedback control system is given by G(s) = A Servomechanism has its moment of inertia ] = 10x10“ Kg-m’, retarding friction, B = 400x 10 N-m/ (rad/sec) and elasticity coefficient, K = 0,004 N-nvrad. Find the natural frequency and damping factor of the system. For a second order system whose open Joop transfer function G(s) = » determine the maximum 3(8+2) over shoot, the time to reach the maximum overshoot when a step Gisplavenent of 18° is given to the system, Find the rise time, time constant and the settling time for an error of 7% 25 s(5+5) Obtain the rise time, Peak time, Maximum overshoot and the settling time when the system is subjected to ‘a unit-step input. Consider the unity feedback closed loop system where the forward transfer function is G(s)= Consider the system shown in fig E2.7, where £=0.6 and @,= 0.5 rad/sec. Determine the rise time, peak time, maximum overshoot and Settling time, when the system is subjected to a unit-step input. RG) Eb) ct) Fig E2.7. For the system shown in fig E2.8, determine the values of K and K, so that the maximum overshoot in the unit step response is 0.2 and the peak time is 1 sec, With these values of K and K,, obtain rise time and settling times RG) E(s), K ls) Y sis+]) = | nik d | Fig £28 The system shown in fig E2.9 subjected to a unit-step input. Determine the values of K and T, where the Maximum overshoot of the system is 25.4% corresponding to ¢ = 0.4. Determine the values of K and T of the closed-loop system shown in Fig E2.10, so that the maximum overshoot in unit-step response is 25% and the peak time is 2 sec. Assume that J=1 Kg-m?.R6) RO ig x 1 Cs) ° Js? [" 1s E2.11 A unity-feedback system is characterized by the open-loop transfer function G(s) = L s(05s+1) (028+) a) Determine the steady-state errors to unit-step, unit-ramp and unlt parabolic inputs. b) Determine rise time, peak time, peak overshoot and settling time of “the unit-step respone of the system. E212 For a system whose (i(s) =——1¢__ s(S+1) (S42 find the steady state error when it is subjected to the input, r(t) = 1+2t+1.5 t. E213 A unity feedback system has. G(s) = The input to the system is described by r(t ) = 4+ 6t+-21°, Find the generalized error coefficients and steady state error. 10 (s+) Find the steady state error and the generalised error coefficient for r(t) = t. E2.14 A unity feedback system has the forward path transfer function G(s) = E215 Find out the position, velocity and acceleration error coefficients for the following unity feedback systems having forward loop transfer function G(s) as, 100 b x © man © Mensary « Kiet) (125) (2) FFTs 100) © 3545020) E216 The open loop transfer function of a unity feedback control system is G(s) = 9/(s+1), using'the generalized error series determine the error signal and steady state error of the system when the system is excited by, @ ()=2 Gi) =e Git) (= 38/2 (iv) r() = 1424302 K(s+2) Fela tesiy Determine, E217 For unity feedback system having open loop transfer function as G(s) = (type of system, (ii) error constants K,, K,and K, (ii) steady state error for parabolic input.2.84 ANSWER FOR EXERCISE PROBLEMS 2.1 18 401455 tS etcostt 16 singt E2.2 olt)=t+Fe'cosst- Te sinat |E2.3 c(ty= [t-o186 e745 _ 0.88 297 cos(3, 65t-22°)| As Lends to infinity, c(t) tends to infinity and so the:system is unstable. Therefore % over shoot and rise time are not defined. E2.4 Natural frequency, 13 3: - Se sini zo snt = 20 radisec, Damiping factor, ¢= 1. 2.5 Maximum overshoot = 0.16, when input is 18%; M,= 2.88% Peak time, t.=1.8isec. Rise time, t=1.21 sec } Time constant, T=1 sec, Settling time for 7% error = 2.66 sec. E2.6 Rise time, {= 0.55sec, %Peak overshoot, M.= 9.5% Peak time, t,= 0.785sec, Settling time, t,= 1.338ec (for 2% error); 1, =1 seo( for 5% error) E2.7 Rise time, ,=0.55sec, Maximum overshoot, —_-M,= 0.095 Peak time, t,= 0.785sec, Settling time, t, = 1 sec {for 5% criterion) |E2.8 K=125, Risetime, = 0.65sec K,=0.178 ; Settling time, t,= 2.48 sec (for 2% error); = 1.86 sec (for 5% error) E2.9 K=142, T=1.09 E210 K=2.95N-m T=0.471 sec = (b) Rise time, =1.91 sec eo. ©) salen ° Peaktime, ip =2.79 sec | SSiuntrem Peak overshoot, Mp =0.1265 | gre parton = © Settlingtime, ts =54sec £2.12 The total steady state errors = }£2.13 C,=0; C.=1; C50; C=-6; eso JE2.14 Cet; C1012; ee, j£2.18 Question K, K, Ky | @ 1000 0 ) x K 0 © ~ co — KI100 @ @ ~ K/20 @ et) =02 58, =02 £2.17 (i) _Itistype-2 system (li) e@) = 0.11+0.09 re, =o. @) Kee Kz at KK (il) e(t) = 0.15'+0.27t-0.054 ;e = (v) eft) = 0.152+0.77+0.226 se, Gil) _e,= 6K } | I tCHAPTER 3 FREQUENCY RESPONSE ANALYSIS : 3. 1 SINUSOIDAL TRANSFER FUNCTION AND FREQUENCY RESPONSE The response of a system for the sinusoidal input is called sinusoidal response. The ratio of sinusoidal response and sinusoidal input is called sinusoidal transfer function of the system and in general, it is denoted by T(j@). The sinusoidal transfer function is the frequency domain representation of the system, and so it is also called frequency domain transfer function. The sinusoidal transfer, function T(ja) can be obtained as shown below. 1. Contruct a physical model of a system using basic elements/parameters. 2. Determine the differential equations governing the system from the physical model of the system. 3. Take Laplace transform of differential equations in order to convert them to s-domain equation. Determine s-domain transfer function, T(s), which is ratio of s-domain output and input. 5. Determine the frequency domain transfer function, T(ja) by replacing s by jo in the s-domain transfer function, T(s). | Note : If the s-domain transfer function, T(s) is known, then frequency domain transfer | function, Ta) can be obtained directly from T(s) by replacing s by ja. ie., T(s) =2> T(jo) Consider a linear time invariant system with frequency domain transfer function, T(j@) shown in fig 3.1. Let the system be excited by a sinusoidal signal frequency @, amplitude A, and phase 6. Now the tesponse or output will also be a sinusoidal signal of same frequency @, but the amplitude and phase of response will be modified by amplitude and phase of the transfer. function respectively. Now, the amplitude of the response is given by the product of the amplitude of the input and transfer function. The phase of the response is given by the sum of the phase of the input and transfer function. Let, T(jo)=|TGo)| 2 Téo) where, [TGe)| = Magnitude of T(ja), and, 2 T(j@) = Phase of T(ja). Let, Input, r(t)=Asin (@t+@)=A 40 where, A = Amplitude of input , @ = Frequency of input, and © = Phase of input. Now, Response, c(t) =r(t)xT(ja) = AZO x [T(ja)| Z Te) =A x{TC jo)| 2(8+ ZT§o)) =B Zo where, B= Ax|T(jo)| = Magnitude of response, and, 6= @+2T(j@) = Phase of response. r(t) , ; ; c(t) o(t)=B ZO ; r()=Asinfat +0)=AZ0 | SU =!TUoX 4u,-4u7-827u, =0 6 4ubs4u,-8Cu, => uw =1-2¢? > u=fi-2¢? sen(3.13) Therefore, the resonant peak occurs when u, = ¥ 1-2¢? 4u,(1-u7) ~ 8C7u, = 0 Put this condition in the equation for M and solve for M. 1 I fa—weyPeactag]? [a-c-2e2y? +4222} 1 [anes ache - . 1 usup ‘ 1 I 1 = [acteac? act? [ac?—actP faa vt ~.|Resonant peak, My wn(3,14) 26y1-C RESONANT FREQUENCY (o,) Normalized resonant frequency, u, = = i (3.15) A sasus(3.16) The resonant frequency, o, =@,\1—2C? 2M, = BANDWIDTH (@,) Let, Normalized bandwidth, u, = 2 oy When u = u,, the magnitude M, of the closed loop system is 1/./2 (or -3db). Hence in the equation for M (equation 3.9), put u =u, and equate to 1/ v2. 1 1 Mes ae ge ye (3.17) 22 22 is vz [id uy? + 407g? On squaring and cross multiplying we get, (l-up)?+4C? uf =20 => Lut 2ue 4 ac? a2 =2 Let, x=ug; 0 -.x?-2(1-20?)x-1=0 xe 21-207) 4 y4—207)? +4 _ 20-20) s 244g = 402) 41 2 => uf-2u? (1-¢?)-1 =0 2 Let us take only the positive sign, 2x= 1-20? #4f2-4C? +44 But,uy=vx: uy =vx=[1-20 + 2-40 +4c4 oO Also, u, =—2 aKS 2.| Bandwidth, o, =o, uy = [1-20 442-402 +40! I “eal 3.18) PHASE MARGIN (y) The open loop transfer function of second order system, 2 @: 62 —_s sep caguell 3.19 ® s(s+2Co,) ( ) The sinusoidal transfer function G(ja) is obtained by letting s = jo. 2 2 G(jo) = — = =——— = i sne(3.20) eee o{2}o{xi2] i2(axi2| oO, Oy , ®y Let normalized frequency, u = w/o, On substituting u = w/o, in equation (3.20) we get, i GG) $e (3.21) GO) = eee 1 1 Magnitude of G(jeo) = {Gao )) = a (3.22) uy4tur ut +4c7u? 3.23) Phase of G(jo) = -90° tan! ase of GGjw) tan 2 At the gain cross-over frequency ©,., the magnitude of G(ja) is unity. Let normalized gain cross over frequency, u,, = o,/o, On substituting u by Ye in the equation (3.22) and equating to unity, we get, - 4 22 “ ALU=Uge, |G(ja)} = Hl = ut4eu=1 = ul+4@u-1=0 1 A Use +4c2 uz, Let, x=uz. 3 ox? 440?x-1=0 24 4 nog Sy ays £4 eas fi Let us take only the positive sign, axe 204 f4ctel 1 But, uje2 Ve jot = ve-[-2¢ + fac? vif se(3.24) ee 3.25) gee The phase margin, y = 180+ ZG(jo)| utg¢ Substituting for “G(j@) from equation (3.23) in equation (3.25) we get, es ni |-2e2 +4 acti F 7 =180+| -90° — tan? 8] =90- tan L vnse(3.26) 2G 26 | Note : The gain margin of second order system is infinite |3.7 3.4 CORRELATION BETWEENTIME AND FREQUENCY RESPONSE The correlation between time and frequency response has an explicit-form only for first and second order systems. The correlation for second-order system is discussed here. Consider the magnitude and phase of-a closed loop second order system as a function of normalized frequency, as given by equations (3.9) and (3.10). i a Magnitude of closed loop system, M=|M(ja)|= Yew? + guy? Phase of closed loop system, a= ZM(jo)=— tan! = -w The magnitude and phase angle characteristics for normalized frequency u, for certain values of C are shown in fig 3.2 and 3.3. The frequency at which M has a peak value is known as the resonant frequency. The peak value of the magnitude is the resonant peak M, . At this frequency the slope of the magnitude curve is zero. The frequency corresponding to M, is u, which is the normalized resonant frequency. From equations (3,14) and (3.15) we get, Resonant peak, M, = ! 2EyI-C Resonant frequency, o, = @,y1-22 Whenf=0, @, =o,¥1-207 =o, Bo ge ES eel 27) Wheng=0, Mi=—ee— ei (3.28) for C =C,< 0.707 e for o> 0.707 08 06 04 180 027 0 uy u ® 4 a Fig 3.2: Magnitude, M as a function of u. Fig 3.3 : Phase, « as a function of u. From equations (3.27) and (3.28), it is clear that as ¢ tends to zero, @, approaches @,, and M, approaches infinity. When 1-2¢?= 0, @, = 0, which means there is no resonant peak at this condition. Let, 1-22=20:2-0-+ = c= et, 20.5 © 5 a For 0 172 , the condition (dM/du) = 0, will not be satisfied for any real value of . Hence when €> WW2 the magnitude M decreases monotonically from M = | at u = 0 with increasing u. It follows that for C >-1/-/2 there is no resonant peak and the greatest value of M equals one. The frequency at which M has a value of 1//2 is of special significance and is called the cut-off frequency @.. The signal frequencies above cut-off are greatly attenuated on passing through a system. For feedback control system, the range of frequencies over which M > y 4/2 is defined as bandwidth ,. Control system being low-pass filters (at zero frequency M = 1), the bandwidth @, is equal to cut-off frequency ,. “In general the bandwidth of'a control system indicates the noise-filtering characteristics of the system. Also, bandwidth gives a measure of the transient Tapa The normalized bandwidth, u, ="! = 2+ -[1- ae? +42-4¢ “EP From the equation of'u, it is sind that u, is a function of ¢ se The graph between u, and ¢ is shown in fig 3.4. The expression for the damped frequency of oscillation @, and peak overshoot M, of the step response, for 0 << 1 are, ~te Damped frequency, ®,=o,J1-@2 and _ Peak overshoot, M,= ele Comparison of the equation of M_ and M, reveals that both are functions of only ¢. The sketch of M, and M, for various value of € u, are shown in fig 3.5. The sketches reveals that a system with a given value of M, must exhibit a corresponding 1.5 value of M, if subjected to a step:input. For C>W2, the resonant peak M_ does not exist and the correlation 1.0 breaks down. This is not a serious problem as for this range of G, the step response oscillations are well damped 9.5 and M, is negligible. The comparison of the equation of w, and ©, s reveals that there exists a definite correlation between ° 05 0707 1 c fee sketch of @ /a, with respect to £ is shown in Fig 3.4: Normalised bandwidth as a fimetion of. MM, Be @q 1.0 o, Og 05 0 05° 0.707 1 $ oO 05 0.707 t Fig 3.5: Mand M, as afunetionofC. — - Fig 3.6 : w,/@, as a function of 63.9 3.5 FREQUENCY RESPONSE PLOTS Frequency response analysis of control systems can be carried either analytically or graphically. The various graphical techniques available for frequency response analysis are, 1... Bode plot 4. Mand N circles 2. Polar plot (or Nyquist plot) 5, Nichols chart 3. Nichols plot The Bode plot, Polar piot and Nichols plot are usually drawn for open loop systems. From the open loop response plot, the performance and stability of closed loop system are estimated. The M and N circles and Nichols chart are used to graphically determine the frequency response of unity feedback closed loop system from the knowledge of open loop response. The frequency response plots are used to determine the frequency domain specifications, to study the stability of the systems and to adjust the gain of the system to satisfy the desired specifications. 3.6 BODE PLOT The Bode plot is a frequency response plot of the sinusoidal transfer function of a system. A Bode plot consists of two graphs. One is’a plot of the magnitude of a sinusoidal transfer function versus log a. The other is a plot of the phase angle of a sinusoidal transfer function versus log @. The Bode plot can be drawn for both open loop and closed loop system. Usually the bode plot is drawn for open loop system. The standard representation of the logarithmic magnitude of open loop transfer function of G(ja) is 20 log |G(jm)| where the base of the logarithm is 10. The unit used in this representation of the magnitude is the decibel,usually abbreviated as db. The curves are drawn on semilog paper, using the log scale (abcissa) for frequency and the linear scale (ordinate) for either magnitude (in decibels) or phase angle (in degrees). The main advantage of the bode plot is that multiplication of magnitudes.can be converted into addition. Also a simple method for sketching an approximate log-magnitude curve is available. Consider the open loop transfer function, G(s) = Gs stp s(1+sT,) (1+sT K (1+ jot) jo (1+ joT,) (1+ joT) & K20° f1+07T? Aan of; 290° yltort? Ztan @T,J1+07T? Ztan* oT; K +0212 The magnitude of G(ja) = |G(jo)|= iter — ole’ j1+0’t The phase angle of the GGa) = ZG(ja) = tan” oT, — 90° — tan” wT, tan”! eT, G(jo) = The magnitude of G(jw) can be expressed in decibels as shown below. |GGo) in db = 20 log |G(jo)/ = 20 log K o Jie? Ji+07EK “1 1 =20 log] >xJi+o™1 x x : Vi +07 Vi +07? =20 tog © +20 108 1+0°T) +20 log L +20 log 4 © itot : 1+@71? = 20 log S420 log f1+@2T? —20 log ji+or? —20 log jitor aa (3.29) From the equation (3,29) it is clear that, when the magnitude is expressed in db, the multiplication is converted to addition. Hence in magnitude plot, the db magnitudes of individual factors of G(ja) can be added. Therefore to sketch the magnitude plot, a knowledge of the magnitude variations of individual factor is essential. The magnitude plot and phase plot of various factors of G(j@) are explained in the following section. " BASIC FACTORS OF G(ja) The basic factors that very frequently occur in a typical transfer function G(jo) are, 1. Constant gain, K 2. Integral factor, K or a jo (jo) 3.. Derivative factor, K x jo or K x (jo)* 1 1 4, First order factor in denominator, > ofr ——__— 1+joT (1+ joT)™ 5. First order factor in numerator, (1 +joT) or (1 + joT)" 6. Quadratic factor in denominator, Fexomecoces io /a@, jo / , On On , am 7. Quadratic factor in numerator, I. 2B (2}-(2) | Constant Gain, K Let, G()=K tb K>4 1. Gia) =K=K 20° , (joo) A K=4 A=|G(jo)| in db =20 log K in 0 db $= ZG(ja) =0° db O 1, 20 log K is positive. When 0>1; -. A=~20 log Vito" T? 2-20 tog Yw?T? =-20 log oT Ato=2, A=-20log 1=0 Ato 2 =~20 log10=~20 db The above analysis shows that the magnitude plot of the factor 1/(1+joT) can be approximated by two straight lines, one is a straight line at 0 db for the frequency range, 0<@ < W/ T, and the other is a straight line with slope -20 db/dec for the frequency range, 1/T'< @ < @. The two straight lines are asymptotes of the exact curve. The frequency at which the two asymptotes meet is called corner. freqivency oF break frequency. For the factor 1/(1+j@T) the frequency, o = 1/T is the corner frequency, ,. It divides the frequency response curve into two regions, a curve for low frequency region and a curve for high frequency region. The actual magnitude at the comer frequency, , = uy is, A=-20 logy +1 =-3 db. Hence by this approximation the loss in db at the comer frequency is —3 db. The phase plot is obtained by calculating the phase angle of G(j) for various values of @ Phase angle, 6 = 2G(jo) =— tan”! oT At the corner frequency, o =o, = x, =-tan eT =- tan! 1=—45° As@—>0, 0° Asa >on, 690°3.13 The phase angle of the factor, 1/(1+ j@T), varies from 0°-to —90° as @ is varied from zero to infinity. The phase plot is a curve passing through —45° at ®,. When the first order factor in the denominator has a multiplicity of m, then, G(s) = G(jo) pre he ps 1 1 Tap?" a (l+sT) (1+ joTy ( fi+o* ) Zm tan oT =|GGo)|in db = 20 log ——!_ = -20 m togv+0*1 ( +07) $= ZG(jo) =-m tan oT [ ae i A _20m| : : rt : : ’ ' 0. . : 4-90 i =i 10 e : 1 10 2 Qo = T Tr (log scale) rs - (log scale) Fig 3.12 : Bode plot of the factor Fig 3.13 : Bode piot of the factor 1/(1+joT)”. 1 1+joT* Now the magnitude plot of the factor 1/(1+ jaT)" can be approximated by two straight lines, one is a straight line at zero db for the frequency range, 0 < @ < 1/T, and the other is a straight line with slope -20 m db/dec for the frequency range, 1/T <@ <~. The corner frequency, o, = 1/T and the loss in db at the corner frequency is —3m db. The phase angle of the factor 1/(1+ jaT)" varies from 0° to -90m° as @ is varied from zero to infinity. The phase plot is a curve passing through —45m° at o,. FIRST ORDER FACTOR IN THE NUMERATOR G(s)=1+sT G(je) =14 joT =V1l4+0?T? Ztan oT = |G(ja)| in db =20log ¥1+071? $= ZG(jo) = tan! oT By an analysis similar to that of previous section it can be shown that the magnitude plot of the factor (1+j@T) can be approximated by two straight lines, one is a straight line at zero db for the frequency Tange 0< @ < 1/ T and the other is a straight line with slope +20 db/dec for the frequency range WT < @ <«, The two straight lines are asymptotes of the exact curve. The frequency at which the two asymptotes meet is called the corner frequency or break frequency. For the factor (1+jmT), the frequency, (@ = 1/T) is the corner frequency, w,. By this approximation the loss in db at the comer frequency is +3 db. The phase angle of the factor (1+joT) varies from zero to +90° as @ is varied from 0 to. The phase plot is a curve passing through +45° at o,.3.14 When the first order factor in the numerator has a multiplicity of m, then, G(s) =(1+sT)” Gijo)=(1+ jot)!” = (i +o ) ZmtanoT . eo A= |G(ja)| in db =2010g (Vi+0* ) =20mlogy 1+ 071 $= ZG(jo) = m tan oT a 10 @ Ome Tr (log scale) Fig 3.15 : Bode plot of the factor (1+jaT)” 10 @ Om FT (log scale) Fig 3.14 : Bode plot of the factor (1+joT). Now the magnitude plot of the factor (1 + jaT)" can be approximated by two straight lines, one is a straight line at zero db for the frequency range 0 < w < W/T and the other is a straight line with a slope of +20m db/dec for the frequency range 1/T<@ tar —Ss oO 2 + j26— ee. 2 0 Ia ) On iC o +46 z 2 Let, A=|G(jo)| in db. 2 2 at A=20log =-201 (1-25) wee On On 4 -: 2 2 4 =-20i0g | 1+ 8p -2% oe =~20log | 1-20-4042, OQ) On OF @r a,3.5 At very low frequencies when @ << @, , the magnitude is, 2 >» of A=~20 log [1-—> (2-4 )+—7 =—20 log 1=0 a @y At very high frequencies when @ >> @, , the magnitude is, 2. 4 4 gi 2 A=-20 log fi-S> (2-407) +25 2-20 log 2 =~20 log gr 204[ 2.) On a, Oy On Oy i A=—40 log? 2 Ato=o, A=-40 logl =0 db At w= 100,, A=—40 logl0 =— 40 db From the above analysis it is evident that the magnitude plot of the quadratic factor in the denominator can be approximated by two straight lines, one is a straight line at 0 db for the frequency range 0 <@ < oa, and the other is a straight line with slope -40-db/dec for the frequency range @,< @ HiAS @=0,, 9=-tan™ Baton «o=—90° As@30, 630 AS @ 20, 6-180" The phase angle of the quadratic factor varies from 0 to -180° as @ is varied from 0 to «0. The phase plot is a curve passing through —90° at @,. Atthe comer frequency phase angle is -90° and independent of G, but at all other frequency it depends on €. QUADRATIC FACTOR IN THE NUMERATOR 2 2 e G(s) = 22260808 _ 14 ¢ (2 (=) @ Oy ry Oy . 42 ay 3 xe G(jo) = 14 j26 2.{2] =f [1-25] +4? 2 ean? —2en. ®n (On 3 Oy y- 2 5 a +db) Approximate plot Slope = +40 db/dec =a, a(log scale) Fig 3.17 : Bode plot of quadratic factor in numerator. Based on an analysis similar to that of denominator quadratic factor, the magnitude plot of the quadratic factor in the numerator can be approximated by two straight lines, one is a straight line at 0 db for the frequency range 0 < <@, and the other is a-straight line with slope +40 db/dec for the frequency range @, <<, The corner frequency is , . Due to this approximation the error at the corner frequency depends on ¢. The phase angle varies from 0 to +180°, as @ is varied from 0 to ©. At the corner frequency the phase angle is +90° and independent of C, but at all other frequency it depends on €.3.17 PROCEDURE FOR MAGNITUDE PLOT OF BODE PLOT From the analysis of previous sections the following conclusions can be obtained. 1. The constant gain K, integral and derivative factors contribute gain (magnitude) at all frequencies. 2. In approximate plot the first, quadratic and higher order factors contribute gain (magnitude) only when the frequency is greater than the corner frequency. Hence the low frequency response upto the lowest corner frequency is decided by K or K / (ja) or K(jo) term. Then at every corner frequency the slope of the magnitude plot is altered by the first, quadratic and higher order terms. Therefore the magnitude plot can be started. with K or K/§jo) or KGo)' term and then the db magnitude of every first and higher order terms are added one by one in the increasing order of the comer frequency. This is illustrated in the following example. 2 Let, G(s) -= Kes s’ (1+sT,) (1+sT;) sant 2 6Ga)=_—- kate (jo) (A+ joT,) (1+ jo) Let, T2 +db Odb -db +db Odb Odb —db Odb -db +db Odb —db 3.18 : Magnitude plot. of the term ae is Go} wo : 2 2 Magnitude plot of thetterm (1 + joT,) Magnitude plot of the term 1 (14 joT3) z Magnitude plot of the term ae oe 1 O db/dec (14 joTe) a ®s ®2 ea(log scale) K(1+ joT,Y Fig 3.18 : Magnitude plot of bode plot of, Gjo )=— > __. 's ignated plot of bode plot of GOGO)" a jam Ml JOB)5.19. Step 4: Then calculate the gain (db magnitude) at every corner frequency one by one by using the formula, Gain at @, = change in gain from @, to o, + Gain at o, ® =| Slope from @, to 0 xhe2t| + Gain ato, Ox Positive slope O, a, Step 5: Choose an arbitrary frequency o, which is greater than the highest corner frequency. Calculate the gain at @, by using the formula in step 4. Step 6: Inasemilog graph sheet mark the required range of frequency on x-axis (log scale) and the range of db magnitude on y-axis (ordinary scale) after choosing proper units. Step 7: Mark all the points obtained in steps 3, 4, and 5 on the graph and join the points by straight lines. Mark the slope at every part of the graph. Note : The magnitude plot obtained above is an approximate plot. If an exact plot is needed then appropriate corrections should be made at every corner frequencies. PROCEDURE FOR PHASE PLOT OF BODE PLOT The phase plot is an exact plot and no approximations are made while drawing the phase plot. Hence the exact phase angles of G(j@) are computed for various values of @ and tabulated. The choice of frequencies are preferably the frequencies chosen for magnitude plot. Usually the magnitude plot and phase plot are drawn in a single semilog - sheet on a common frequency scale. Take another y-axis in the graph where the magnitude plot is drawn and in this. y-axis mark the desired range of phase angles after choosing proper units. From the tabulated values of @ and phase angles, mark all the points on the graph. Join the points by a smooth curve. DETERMINATION OF GAIN MARGIN AND PHASE MARGIN FROM BODE PLOT The gain margin in db is given by the negative of db magnitude of G(jo) at the phase cross-over frequency, o,.. The ©, is the frequency at which phase of G(j@) is -180°. If the db magnitude of GG) at @,. is negative then gain margin is positive and vice versa. : Let be. be the phase angle of G(ja) at gain cross over frequency ,,. The @,, is the frequency at which the db: magnitude of G(j@) is zero. Now the phase margin, y is given by, y = 180° + Oy: If 9,, is less negative than —180° then phase margin is positive and vice versa.3.20 +db | Negative A o GM. in db —db Positive PM Negative - PM { -180 6 —270 9 @, — @(log scale) eo a(log scale) ee Fig 3.19: Bode plot showing phase margin (PM) and gain margin (GM). GAIN ADJUSTMENT IN BODE PLOT In the open loop transfer function G(j@) the constant K contributes only magnitude. Hence by changing the value of K the system gain can be adjusted to meet the desired specifications. The desired specifications are gain margin, phase margin, o . and ©,,, Ina system transfer function if the value of K required to be estimated to satisfy a desired specification then draw the bode plot of the system with K = 1. The constant K can add 20 log K to every point of the magnitude plot and due to this addition the magnitude plot will shift vertically up or down. Hence shift the magnitude plot vertically up or down to meet the desired specification. Equate the vertical distance by which the magnitude plot is shifted to 20 log K and solve for K. Let; x= change in db __ (x is positive if the plot is shifted up and vice versa). Now, 20logK=x ; logK=x20; «. K=10% Note : A point in complex plane can be represented by rectangular coordinates or by polar coordinates. Consider a point, z = atjb in complex plane. Now, |{=va? +b? and Z2= tan" Wa, If the point lies in first or fourth quadrant then the argument as calculated by tan’’ b/a will be the correct values. But if it lies either in second or third quadrant then a correction should be made in the calculated values of argument, because the calculator will always give the values of tan” bla either from 0 to + 90° or from 0 to-90°. The corrections to be made while converting from rectangular to polar coordinates is shown below. A point in I* quadrant, a+jb=Va?+b* Ztan”'b/a A point in I™ quadrant, -a+jb= va’ +b? Z(x— tan 'b/a) A point in III" quadrant, —a- jb= a? +b? 2(n+ tan“b/a) A point in TV" quadrant, a—jb =va? +b? Z~tan™'b/a)3.21 EXAMPLE 3.1 ‘Sketch Bode plot for the following transfer function and determine the system gain K for the gain cross over frequency to be 5 rad/sec. Ks 2 tS) = Tres) (1 0025) SOLUTION ‘The sinusoidal transfer function G(jo) is obtained by replacing s by jo in the given s-domain transfer function. K(jo)" (1+ 0.2j0)(14 0.02j0) (On (1+ }0.20)(1+ j0.020) +. Gje)= LetK=1, ©. G(jo)= MAGNITUDE PLOT The comer frequencies are, o,)= 215 =5 tad /sec and Oa = hy 50 rad / sec ‘The various terms of G(ja) are listed in Table-1 in the increasing order of their comer frequency. Also the table shows the slope contributed by each term and the change in slope at the comer frequency. TABLE-1 Term Corner frequency Slope Change in slope rad/sec db/dec db/dec of 7 +40~., 1 cdg 14 }0.20 a= a9 “a 1 1 => => =50 1+ 0.020 Oe = O02 Choose a low frequency a, such that o,<@,, and choose a high frequency ‘©, Such that o,> 0, Let, o,= 0.5 rad/sec and o,=100 rad/sec, Let, A=|G(jo)| in db. Letus calculate A ato, ,0,,,0,. and o,. Ato=a,, A=20log(ja)"|= 20 log (0)*= 20109(0.5)" = 12 db Ate =o, A= 20 log(jo)"|=20 tog (w)” = 20109(5)* = 28 db Ato=am, A -|sooe from @¢; 10 Og x log 2a|-a, )= 20xlog 2+ 28 =48 db ct erecy, =Oxlog 3? + A8=48 db Ato =on, A=|slone from 0,2 to xtog 2 | +4 2 | ete-oe2)3.22 Letthe points a, b, cand d be the points corresponding to frequencies o,, o,,,@,, and a, respectively on the magnitude plot. Ina semilog graph sheet choose a scale of tunit= 10db on y-axis. The frequencies are marked in. decades from 0.1 to 100 rad/sec on, logarithmic scales in x-axis. Fix: the points a,b, cand don he graph. Join.the points by straight lines and mark the slope on the respective region. PHASE PLOT The phase angle of G(jo) as a function of » is given by, = ZG(jo) = 180° - tan“'0.20 — tan”'0.020 The phase angle of G(ja) are calculated for various values of and listed in table-2. TABLE o tan 0.20 tan? 0.020 o = Z2G(jo) Point in rad/sec deg deg deg phase plot 0.5 5.7 06 173.7 2174 e 1 11.3 Wa 167.6 168 f 5 45 57 129.32 130 g 10 63.4 13 105.3~106 h 50 84.3 45 50.7 = 50 i 100 87.1 63.4 29.5 =30 j On the same semilog graph sheet choose-a scale of 1unit = 20°, on the y-axis‘on the right side of semilog graph sheet. Mark the calculated phase angle on the graph sheet. Join the points by a smooth curve... CALCULATION OF K Given that the gain crossover frequency is 5 rad/sec. At@=5 rad/sec the gain is 28 db. If gain crossover frequency is 5 rad/sec then at that frequency the db gain should be zero. Hence to every point of magnitude plot a db gain of -28db should be added. The addition of -28db shifts the plot downwards. The corrected magnitude plotis obtained by shifting the plot with K = t by -28db downwards. The magnitude correction is independent of frequency. Hence the magnitude of -28db is contributed by the term K. The value of K is calculated by equating 20 logK to—28 db. *. 20 log K =-28 db 8 2B fog K :K =1018) = 0.0398 20 The magnitude plot with K = 1 and 0,0398 and the phase plot are shown in fig 3.1.1 Note : The frequency @ = 5 rad/sec is acomer frequency. Hence in the exact plot the db gain at o =5rad/sec will be 3db less than the actual plot. Therefore for exact plot the 20 jog K will contribute a. gain of -25ab. -. 20 log K=-25 db 25.7 log K=—— ; K=10 \” eae oO = 0.0562C0z0 of + I@col +) O08 99s/ped UL <— © 001 os Ol 5 I 50 ro = (ODD Uopounf safsunis fo jojd apog + rT" Buy oe- z1- 4P Ol- wy Vv ved oe aS | 1689S & Y T6896 € Z 5.23 a 6 ©.i 3.24 ere Eee 8 EXAMPLE 3.2 ‘Sketch the bode plot for the following transfer function and determine phase margin and gain margin. 75 (1+0.2s) s)o SFE +16s +100) SOLUTION On comparing the quadratic factor in the denominator of G(s) with standard form of ‘quadratic factor we can estimate Gando,. .8?+168+100 = s?+2t0,8+0? On comparing we get, o2=100 = o,=10 16 16 oO, = epee is 2o,=16 = ¢ 20, 2x10 0.8 Letus convert the given s-domain transfer function into bode form or time constant form. Ge) 75(1+0.28) 75(1+0:2s) “0:75 (4402s) ~ s (s?+16s +100) (3? 16s s (1¥0.0is?+0.16s) sx 100 (100 +> 44 100. 100 The sinusoidal transfer function G(jo) is obtained by replacing s by je. in G(s). 2 0.75(1+0.2jo) 0.75 (1+ j0.20) + Gio) = ————_-—___ = - eo jo (140.01(j0)° +0.16j0) jo (1-0.0» + j0.160) MAGNITUDE PLOT i The comer frequencies are; ©. “O07 Srad/sec and @)=0,=10rad/sec Note - For the quadratic factor the commer frequency is ®,,. i EU oe Soe The various terms of G(ja) are listed in table-t inthe increasing order of their comer frequencies. Also the table shows the slope contributed by each term and the change in slope at the comer frequency. TABLE-1 Term Corner frequency Slope Change in slope rad/sec db/dec dbidec 0.75 5 jo 7 14j0.2 oaeae=b +j0.20 a= Go 1 1-0.019? + /0. 160 BF =a0 L Choose a low frequency @, such thato,< ®., and choose a high frequency w, such that o,> o_. Let, @,= 0.5 rad/sec and w,=20 rad/sec. Let, A=|G(jo)| in db,3.25 Letus calculate A ata, o,,,0, and ,,. 0.75 0.75 om |-20 0-75 _ 3.5 db jo 805 0.751 _ 59 tog 922 =-16.5db At, a=, A=20log| AL @=@,,, A=20 log je AL @= 0g, A= [sone from 0,1 10.02 x log 22h Nasi ct =0x log" +(-16.5)=-165 db At © =, A'=| slope from @ .2 too, xlog-2- +N o=, i 2 F0Op @y| ere) =—40 x tog 20 +(-16.5) =-28.5 db Letthe points a, b, cand d be the points correspanding to frequencies o, ,,, 0,2 and , respectively on the magnitude plot. Ina semilog graph sheet choose a Scale of tunit = 5 db on y-axis. The frequencies are marked in decades from 0.1 to 100 radisec on logarithmic scales in x-axis. Fix the points a,b,cand don the graph. Join the points by straight lines and mark the slope One respective region: Note : In quadratic factors the PHASE PLOT phase varies from 0° to 180°. But calculator calculates tar’ only The phase angle of G(jm) as a function of « is given by, between 0° to 90°. Hence a $= 2G (jo) = tan“'0.20 — 90° wan foro <0, correction of 180° should be added 20.088 to phase aftero,, 0.160 $= ZG(ja) = tan“'0.20 - 90° {ten + 180") for @>@, . 1-0.0t0? ‘The phase angle of G(jo) are calculated for various values of w and listed in Table-2. ‘TABLE-2 ® tan“'0.2 0 tan $= 2G(o) Points in rad/sec deg deg deg -_ phase plot 0.5 57 46 —88.9~-88 e 1 11.3 9.2 —87.9%-88 f 5 oe 45 468 =91.8%-92 Q 10 63.4 90 —116.6 =-116 h 20 79 —46.8+180 =133.2 -147.3 »-148 i 50 84.3 —18.4+180 =161.6 -167.3 ~-168 j 100 87.1 ~92+180=170.8 -173.7 = -174 k ‘On the same semilog graph sheet choose a scale of {unit = 20° on the y-axis on the right side of semilog graph sheet Mark the calculated phase angle on the graph sheet. Join the points by a smooth curve. The magnitude plot and the phase plot are shown in fig 3.2.1.3.26 Response Analysis) ler 3 - (ogrof +070 0-Dol pie aera g Her =(ofy ‘uopounf 4 Oor os Oz upd fo jojd apog : Peg Buy 998/PRi Ul <— © ud : al | i i | | ! oslo s bp £ Z losc9s bh € F 1ORLIOS PET TosL9S yp € TC TO8L9S bp € OF I i3.27 Let 4, be the phase of G(jo) at gain cross-over frequency, Oe From the fig 3.2.1, we get, d= 88° -. Phase margin, g = 180° + 6, = 180° -88° = 92° The phase plot crosses — 180° only atinfinity. The |eGo)} atinfinity is— =2tad/sec anda. =~ = 8 rad/ sec The comer frequencies are, 9c} 05 = 0405 The various terms of G (jo) are listed in table-1 in the increasing order of their comer frequencies. Also the table shows the slope contributed by each term and the change in slope at the comer frequency. TABLE-1 Term Corner frequency Slope Change in slope rad/sec db/dec db/dec 0.0625 je - 20. 1 1+ j0.50 2 1 a“ 1+ j0.1250 —40~20=—60 Choose a low frequency «, such thato, o,. Let, a, =0.5rad/secand o, = 50 rad/sec. Let, A=[G(jo)| in db. Letus calculate A at o,.0,,, Oo and a,3.28 Ato =a, A=20 og2-2625| - 2o19g°-9825 _ seg | jo 0.5 Ato =0,, A=20 lo — = 20109 ee =~30db At@ =@q9, A=| Slope from @,1t0 ay xlog 22 Aaron y= 40xtog +(-30)= -54db Oy eect 2 50 . (at e=exg) "-80%l0og $+(-84) =—102db Ato =p, A=| Stone from og to o, ~iog 22 +4 2 Letthe points a,b,c,d be the points corresponding to frequencies o,, o,,, 0,, and w, respectively on the magnitude plot. Ina semilog graph sheet choose a scale of junit = 10db on y-axis. The frequencies are marked in decades from 0.01 to 100 rad/sec on logarithmic scale in x-axis. Fix the points a, b, c, and d onthe graph. Join the points by straight line and mark the slope on the respective region. PHASE PLOT The phase angle of G(ja) as a function of a is given by, 180° $=-0.20 x —90° — tan“'0.Se - tan'0.1250 The phase angle of G(jo) are calculated for various values of o and listed in table-2. TABLE-2 | « =0.2 @ (180%) tan 0.5 @ tan 0.125 o @= ZG(jo) Point in rad/sec deg deg deg deg phase piot 0.01 -0.1145 0.2864 0.0716 -90.4= -90 e 0.1 1.145 2.862 0.716 94.7 = -94 f 05 5.7 14 3.6 -113.3 = -114 9 1 -11.4 26 712 -134.4 = -134 h 2 +22.9 45 14 ~171.9 = +172 i 3 — 34.37 56.30 20.56 201.2 = -202 j 4 — 45.84 63.43 26.57 225.8 »- 226 k Onthe same semilog graph sheet choose a scaie of {unit=20° on the y-axis on the rightside of the semilog graph sheet. Mark the calculated phase angle on the graph sheet. Join the points by smooth curve. The magnitude and phase piotare shown in fig 3.3.1. CALCULATION OF K Phase margin, y= 180°+ $,,, where ¢,. is the phase of G(ja) ata =a... When y= 45°, 6,, = 7-180? = 45° — 180° = -135°. With K=, the db gain atg=-135° is -24 db. This gain should be made zereto have to PM of 45°. Hence to every point of magnitude plot a db gain of 24 db should be added. The corrected magnitude plot is obtained by shifting the plot with K=1 by 24 db upwards. The magnitude correction is independentof frequency. Hence the magnitude of 24 db is contributed by the term K. The value of Kis calculated by equating 20lagK to 24 db. 2 20logK=24 ; K=10% > K=15.84 With K = 1, the gain margin =—(-32)= 32 db. But the required gain margin is. 2 db. Hence to every point of magnitude plot a.db gain of 30 db should be added. This addition of 30 db shifts the plot upwards. The magnitude correction is independent of frequency. Hence the magnitude of 30 db is contributed by the term K. The value of Kis calculated by equating 20logK to. 30 db.tems Engineerin: Conk 3.29 (@Szr of + Lac of + Der . : a = (OND uOHOUN Jafsubin fo Jord apog + ["E"S Bh, wey? ¥SC900 (oo rf dafsund fo royd opog + 1°g"¢ Mey 998 /per Ul <— oO 09- vs os- 16849 $ ¥3.30 + 20 log K = 30 > K= 10% 7 K=31.62 The magnitude plot with K = 15.84 and 31.62 are shown in fig 3.3.1. EXAMPLE 3.4 Plotthe Bode diagram for the following transfer function and obtain the gain and phase cross over frequencies. gi AOL s (1+ 0.4s) (140.18) SOLUTION The sinusoidal transfer function of G(ja) is obtained by replacing s by jain the given transfer function. Eko 10 * Glo) = 7 x j0da) (+ 50.10) MAGNITUDE PLOT. ‘The comer frequencies are, 1 1 et oas 2.5 rad/sec and 0.9 = oA The various terms of G(ja) are listed in table-1 in the increasing order of their comer frequencies. Also the table shows the slope contributed by each term and the change in slope atthe comer frequency. =10 rad/sec TABLE1 Term Corner frequency Slope Change in slope radisec db/dec db/dec r 40 : ie = -20.. 1 1+ j0.40 1 4+ j0.10 =| Choose a low frequency @, Such that o, < @,, and choose a high frequency @,, such thata, > @,,. Let, @, =0.1 rad/sec, and.o, = 50 rad/sec. Let, A=|G(ja)| in db. Letus calculate A at o,, 0,,, ©, ando,,. 10 10 Ato =o, A=20log Hal 20854740 i 10 40 Ato =0,4 A=20 log a =20log55 = 12 db © 10 Ato =Oy, A= tne from @,;t0 ©. xtog2| *Aata = on) =-40x bog, tlas-12 db Ato =O, A -| sone ffOM 9 t0 Op x bo | *Acat ~60x tog 25+ (-12) =-54db a 2 =e)”a * 16829 ¢ 9a8/pel Ul <— © v i} £ G COL OF +r 0+ DOr Tossa oy € Or & = (ODD ‘uojounf sefsuny fo jojd apog : T-p'E By 2408 oor as 0% ols 8% I 1O8L9S bE t 16819 § pet (68L90 & © c 09~ os- Or Of - or ZL 0z Of Orr3i- Res} sts . * 3.32 eS Lethe points a, b, cand d be the points corresponding to frequencies @,, o,,, 24nd o, respectively on the magnitude plot. Ina semilog graph sheet choose a scale of 1unit = 10 db on y-axis. The frequencies are marked in decades from 0.1 to 100 rad/sec on logarithmic scales in x-axis. Fixthe points a,b, candd onthe graph. Join the points by a straight line and mark the slope in the respective region. PHASE PLOT The phase angle of G(jo) as a function of wis given by, $=-90° - tan“ 0.4@- tan 0.10 The phase angle of G(ja) are calculated for various values of @ and listed in table-2. TABLE-2 | ® tan’ 0.4 o tan’ 0.10 $ = ZG(jo) Points. in rad/sec deg 2 deg deg phase plot 0.1 2.29 0.57 792.86 =-92 e 1 21.80 §.71 117.5 2-118 f 25 45.0 14.0 149 x-150 g 4 57.99 218 169.79 =-170 h 10 75.96 45.0 210.96 =—210 i 20 | 82.87 63.43 ~236.3 = 236 j On the same semilog graph sheet choose a scale of 1unit= 20° on the y-axis on the right side of: ‘semilog graph sheet. Mark the calculated phase angle on the graph sheet. Join the points by a smooth curve. The magnitude and phase plots are shown in fig 3.4.1. From the graph, the gain and phase cross over frequencies are found to be 5rad/sec. RESULT Gain cross-over frequency = 5 rad/sec. - Phase cross-over frequency = 5 rad/sec. EXAMPLE 3.5 For the following transfer function draw bode plotand obtain gain cross-over frequency. 20 as) = 8 (143s) (1+ 4s) SOLUTION The sinusoidal transfer function of G(jo) is obtained by replacing s by je in the given transfer function. st. 20 Sle) = Si Flo) Cieza) MAGNITUDE PLOT A The corer frequenciesare, @,,= 0.25 rad/ sec, OQ = 3 = 0.333 rad/ sec. The various terms of G(jo) are listed in table-1 in the increasing order of their frequencies. Also the table shows the slope contributed by each term and change in slope at the comer frequency.3.33 TABLE-1 Term Corner frequency Slope Change in slope rad/sec db/dec db/dec 20 i = -20 : =4=025 1+ j40 De 1 ele 0.33 1+ Bo aang” Choose a frequency @, such that a, < ,, and choose a high frequency @, such thata, > @,. Let, @,= 0.15 rad/secand o, = 1 rad/sec. Let, A=|G(o)lindb. Letus calculate A at o;, o,,,0,,and o,. Ato =o, A=|Glo), =20log|-22] = 425 ab 015 Ato = 0, A=|G(jo)| = 20'log 20 | 38 ap sh 0.25) Ato =O, A+| sperm O4t005 noo 22 + Ate eq) 94 =40xlog 232 +38=33 db Ato=o,,A -| sere from to @, xlog 22 | + Agte=oc2) Oo =-60 xlog 0 5 +33=4db Letthe points a, b, cand d be the points corresponding to frequencies @,, o,,. 0,, and @, respectively on the magnitude plot. Ina semilog graph sheet choose a scale of 1unit= 10 db on y-axis. The frequencies are marked in decades from 0.01 to 10 rad/sec on logarithmic scales on x-axis. Fix the points a, b, cand don the graph sheet. Join the points by a straight line and mark the slope in the respective region. PHASE PLOT The phase angle of G(ja), = —90° tan” 3m - tan 40 The phase angfe of G(ja) are calculated for various values of » and listed in table-2. TABLE-2 o, tad/sec | tan“' 30, deg tan 40, deg 6 = 2G(jo), deg Points in phase plot 0.15, 24.22 30.96 145.18 *—-146 e 02 30.96 38.66 ~159.61=—160 f 0.25 36.86 45.0 ~171.86 =—172 g 0.33 447 52.8 —187.5 ~—188 h 06 60.14 67.38 —218.32=-218 i 1 71.56 75.98 —237.56 »—238 jf f+pot a eee ee = (OND uopounf safsuoa sof jord apog + 1°9°€ Buy ia) 99S/PRI Ul <= © 2 + asperpt MO . g 849 8 5 = AMAT ] I Response Analysi 0€ ef BE 4% +O | TTC AT LUTE i T68L9S 4. € c 16829 © t T68L9S b € 09 Chapter 3 =F: L68L9S bE t3.35 Onthe same semilog graph sheet choose a scale of 1unit = 20° on the y-axis on the right side of semilog graph sheet. Mark the calculated phase angle on the graph sheet. Join the points by a smooth curve. The magnitude and phase plots are shown in fig 3.5.1. From the graph the gain cross-over frequency is found to be ©,,= 1.1 rad/sec. EXAMPLE 3.6 5(1+ 2s) Forthefunetion, G(s) = 775) 1058s)" draw the bade plot. SOLUTION The sinusoidal transfer function G(jco) is obtained by replacing s by jo in G(s). sete B(1+i20) “Gle) = Gria6) (is 10250) MAGNITUDE PLOT = 025 rad/sec, 0g =F=05 rad/ 00, 0,5 = >be = 4 ad /sec The corner frequencies are, o., = oe 4 The various terms of G(ja) are listed in table-1 in the increasing order of their corner frequencies. Also the table shows the slope contributed by the each term and the change in slope atthe comer frequency. Choose a low frequency a, such that o,<@,, and choose a high frequency @, such thata,> o,,. Leta,=0.1 rad/sec and o,=10 rad/sec. Let A= |G(jo)| in db and let us calculate Aato,, 0,,: 0, 0,,ando,: TABLE-1 Term Corner frequency Slope Change in slope | tad/sec db/dec db/deg 5 - ect lee 1+ j4o 1 +20. | a Nas 17 j0.250 Ato =o, A =|G(jo)] = 20 log 5=+14 db At 0 = ay A=|GGa)| = 20 log 5=+14 db Mo=o4, A= [Shope rom Oytoos xlog2e |-Aroeo =+20 xlog 22 4+ 14 = 48db et 025 ee: = a a AL@ = 0.5, A=! Slope from @ 210 a, xlog—> |+ Ag oo.g) =O wg Gg eee a ; Qn Ato =p, A= Stone from © 310, x log J+ =-20.log Br8= Odb 3 Letthe points a, b, c, dand € be the points correponding to frequencies ©, 0, © 4, @, and @, respectively on the . magnitude plot. Ina semilog graph sheet choose a scale of junit = 5 db ony axis. The frequencies are marked in decades from 0.1 to 100 rad/sec on logarithmic scales on x-axis. Fix the points a, b, c, d and e on the graph. Join the points by a straight line and mark the slope in the respective region.(OST Of + DEO +) (0 +D¢ 3.36 =(OND ‘uoyouns safsuv.s fo joyd apog 2179'S Ita oas/pel UL <— © 001 0s OL v zc I $0 seo 10 Se TO8lL9S pb € c 168L9 S$ & t t 168L9S ¢ €£ z T68L9S F £ t 168L9S F € Tf 13257 PHASE PLOT The phase angle of G(ja), 9 = tan“'(20) - tan“ (40) — tan’ 1(0,250) The phase angle of G(ja) are calculated for various values of w and listed in the table-2. TABLE-2 o tan! 20° |. tan-! 4o tan“! 0.250 $= ZG(o) Points in deg deg deg phase plot 04 113 21.8 1.43 -11.93~-12 f | 0.25 26.56 45.0 35 -21.94%-22 g | 05 45.0 63.43 7A -25.53=-26 h 2 78.96 ~ 82.87 26.56 -33.47 =-33, i 4 82.87 86.42 45.0 ~48.55*-49- | j 10 87.13 88.56 \ 68.19 69.62 =-70 k 50 89.42 89.71 | 85.42 | -85.71~-86 | ! On the same semilog graph sheet choose a scale of 1unit= 10° on y-axis on the right side of the semilog graph sheet. Mark the calculated phase angle on the graph sheet, Join the points by a smooth curve. The magnitude and phase plots are shown in fig 3.6.1. 3.7 POLAR PLOT The polar plot of a sinusoidal transfer function G(§ja) is a plot of the magnitude of G(j@) versus the phase angle of G(j@) on polar coordinates as w is varied from zero to infinity. Thus the polar plot is the locus of vectors |G(j@)| ZGG@) as @ is varied from zero to infinity. The polar plot is also called Nyquist plot. The polar plot is usually plotted on a polar graph sheet. The -270° ~ polar graph sheet has concentric circles and radial lines. The circles s represent the magnitude and the radial lines represent the phase angles. Positive Each point on the polar graph has a magnitude and phase angle. The a: magnitude ofa point is given by the value of the circle passing through+180" As a eference. that point and the phase angle is given by the radial line passing through 180" that point. In polar graph sheet a positive phase angle is measured in anticlockwise from the reference axis (0°) and a negative angle is measured clockwise from the reference axis (0°). Negative angle In order to plot the polar plot, magnitude and phase of Ga) . are computed for various values of @ and tabulated. Usually the choice Fig 3.20: Polar graph. of frequencies are corner frequencies and frequencies around corner frequencies. Choose proper scale for the magnitude circles. Fix all the points on polar graph sheet and join the points by smooth curve. Write the frequency corresponding to each point of the plot. Alternatively, if GG@) can be expressed in rectangular coordinates as, G§jo) = G,(jo) + jG,Ge) where, G,(j@) = Real part of G(ja) ; G,Ge) = Imaginary part of Go), then the polar plot can be plotted in ordinary graph sheet between G,(j@) and G,j@) by varying @ from 0 to «In order to plot the polar plot on ordinary graph sheet, the magnitude and phase of G(ja) are computed for various values of @. Then convert the polar coordinates to rectangular coordinates using P— R conversion (polar to rectangular conversion) in the calculator. Sketch the polar plot using rectangular coordinates.3.38 For minimum phase transfer function with only poles, type number of the system determines the quadrant at which the polar plot starts and the order of the system determines the quadrant at which the polar plot ends. The minimum phase systems are systems with all poles and zeros on left half of s- plane.The start and end of polar plot of all pole minimum phase system are shown in fig 3.21 & 3.22 respectively. Some typical sketches of polar plot are shown in table-3.1. The: change in shape of polar plot can be predicted due to addition of'a pole or zero. 1.. When a pole is added to a system, the polar plot end point will shift by -90°. 2. When a zero is added to a system the polar plot end point will shift by +90°. Start of type-3 system End of 3° order system End of 4" order system Start of type-2—> +— Start of type-0 system system End of 2" End of 1* order system order system Start of type-tt system Fig 3.21 : Start of polar plot of all pole Fig 3.21 : Start of polar plot of all pole minimum phase system. minimum phase system. TABLE-3.1: Typical Sketches of Polar Plot Type: 0, Order: 1 G(s) = 1 = 1+sT 1 1 1 1 L SZ -tan oT _ygg 1+joT ito? AantoT Jl+0?T? As@ 0, G(jo) > 120° GGe)= o increasing Asa>0, G(jo) > 02-90" Type: 1, Order : 2 o~-—— i+sT) 180° G(jo) =~ 1 ——= ! = u Z(-90° - tan @T) jo(l+joT) » 00° ite??? Ztan lot ovl+ oT? . As@ 30, G(jo) > 02-90" As @ 300, .G(ja) > 02-180° 1 Type: 0, Order : 2 G(s) = (l+sT,)(i+st) :, 1 1 G(jo) © (1+ joF) (+ jot) " itor? Z tan oT, i+ 07? Ztan oT, oa 4a 0 - tan) j (1+07 T?\1+0* 7) Aso 30, G(jo) > 120° As@—>0, G(jo) > 02-180"3.39 TABLE-3.1 : Typical Sketches of Polar Plot Type: 0. Order : 3 G(s) = L 6(jo) (I+sT))(1+ sl )(1+sTy) i 1 ~ (1+ joh) (1+ jot) (1+ jot) 1 qi+07t? Ztan ot, (+071? Ltanot, {140713 Ztan oT, _ I As@ > 0, (140°?) (14 0742) (1+0719) Z(-tan oT, — tan“ oT, - tan oT) G(jo) > 120° As @ >, G(jo) > 02-270° Type : 1, Order : 3 e@seee le s(1+ sT,)(1+ 81) | 1 | Gjo)= 270° ‘ . 1 180! jo+ jo) 0+ 50%) goo" f1+0°T? Ztan oT; 1+ 0°T? Ztan“o®, 1 o (1+071?) (+0717 As @ > 0, ne G(jo) > ~Z-90° As @ > 0, G(jo) > 02-270° Type: 2, Order : 4 G(s) L “#(+sh)d+sE) G(jo) 1 - _ 1 Gay’ (+ joT) (1+ 50%) 22-180" f1+07T? Zan of, J1+07 3 Ztan lob, 1 Z(—180° — tan'wT, - tan“ oT) 0° ror f-or) As@ 30, G(ja) > 02-180" As @ 0, G(jo) > 02-360" Type: 2, Order : 5 G(s) A “Fd +sh)d+sh)d+sh) 1 (ja) (1+ jot) (1+ jot) (1+ jot) Gia) = =180° S; 1 07 £- 180° J1+0°T? tan ot, f1+07F? Ztan ot; fi +077? ZtanoT, 1 (i+ (1+074}) G(jo) > 22-180" — As @ +0, Z(-180° ~ tan oT; — tan“loT, - tan oT;) AS @ > 0, G(jo) > 02-450" =02-90°‘TABLE-3.1 : Typical Sketches of Polar Plot 3.40 Type: 1, Order: 1 c@=! s Gtjo)= + =—+ = 1 2-99" As@—>0, Go) > 22 -90° As@ >, G(jo) > 02-90° a@=ltt i 1+joT 1 1 i = G(jo) = =F] =—— 41 = — 2-0 +1 Go) joT joT oT 290° oT As@>0, G(jo) > 04-90" +1 AS@ >, G(ja) > 02-90° +1 G(s) =s G(jo) = jo =0 290° As@>0, G(ja) > 090° AS@ >, G(jo) > 0290" G(s) =1+sT | G(j@)=1+ joT=1+0TZ90° Asa@—->0, G(ja) > 1+0290° As@ >, G(jo) >14+0290° | | L DETERMINATION OF GAIN MARGIN AND PHASE MARGIN FROM POLAR PLOT The gain margin is defined as the inverse of the magnitude of G(ja) at phase crossover frequency. The phase crossover frequency is the frequency at which the phase of G(jo) is 180°. Let the polar plot cut the 180° axis at point B and the magnitude circle passing through the point B be G,. Now the Gain margin, K, = 1/G,. If the point B lies within unity circle, then the Gain margin is positive otherwise negative. (If the polar plot is drawn in ordinary graph sheet using rectangular coordinates then the point B is the cutting point of G(j@) locus with negative real axis and K, = 14G,| where G, is the magnitude corresponding to point B). The phase margin is defined as, phase margin, y = 180° + 6,, where 9,, is the phase angle of G(ja) at gain crossover frequency. The gain crossover frequency is the frequency at which the magnitude of G(ja) is unity.3.41 Let the polar plot cut the unity circle at point A as. shown in fig 3.23 and 3.24. Now the phase margin, y is given by ZAOP, i.e. if ZAOP is below —180° axis then the phase margin is positive and if it is above -180° axis then the phase margin is negative. |-270° 5 7 1 j-90° Gain margin, K, = 2 & Phase margin, y =180° +4, . Phase margin, y =180° +4, Fig 3.23 : Polar plot showing positive gain Fig 3.24 : Polar plot showing negative gain margin and phase margin. margin and phase margin. GAIN ADJUSTMENT USING POLAR PLOT. To Determine K for Specified GM @, 270 Draw G(ja) locus with K =1. Let it cut the -180° axis at point B corresponding to a gain of G,. Let the specified gain _ 480° margin be x db. For this gain margin, the GG@) locus will cut ~180° at point A whose magnitude is G,. ty 1 1_x 1 120 ae Now, 20log—= x = log—=—- = —=10" b ‘ ac eG, 20 CG ¥ |x! 1 G, = 10°" @ Fig 3.25 : Polar plot for different Now the value of K is given by, K= a values of K B If, K>1, then the system gain should be increased. pare": If, K <1, then the system gain should be reduced. To Determine K for Specified PM Draw GG) locus with K = 1. Let it cut the unity circle at ey o° point B. (The gain at point B is G, and equal to unity). Let the ee specified phase margin be x° ; For a phase margin of x°, let ex be the phase angle of G(jq) at gain crossover frequency. 90° x8 = 180° + => by = K° — 180° Fig 3.26 : Gain adjustment for In the polar plot, the radial line corresponding to $,., will velpitréd phase inargint cut the locus of G(j@) with K = 1 at point A and the magnitude corresponding to that point be G,G, 1 Now, K=—%=— “2 Gp = OW, G, G, (~ Gp =1) EXAMPLE 3.7 ‘The open loop transfer function ofa unity feedback systemis given by G(s) = 1/s(1+s) (1+2s). Sketch the polarplotand determine the gain margin and phase margin. SOLUTION Given that, G(s) = 1/s(1+s) (142s) Puts =jo, 1 Ce) Fe Cfo) CAB) 2 The corner frequencies are w_, = 1/2=-0.5 rad/sec and ©.,= 1 rad/sec. The magnitude and phase angle of G(ja) are calculated for the comer frequencies and for frequencies around comer frequencies and tabulated in table-1. Using polarto rectangular conversion, the polar coordinates listed in table-1 are converted to rectangular coordinates and tabulated in table- 2. The polar plot using polar coordinates is sketched on a polar graph sheet as shown in fig 3.7.1. The polar plot using rectangular coordinates is sketched on an ordinary graph sheetas shown in fig 3.7.2. 1 1 G(jo) = (jo) (jo) (120) gra0° Vito? Ztan'o ¥1+40? Aan 20 = 4-90 tanto tan 20 @ (1+ 07) (1# 40?) . 1 1 i “AGGe)= = * @ ylito?) (1¥402) ~oVtt402+0%+40% 0 14507 +404 ZG(jo) = -90°-tan“@ - tan“'20 TABLE-1 : Magnitude and phase of G(j@) at various frequencies @ radisec 0.35 04 0.45 05 06 07 1.0 [GGe)| 2.2 18 15 12 0.9 07 0.3 ZG(jo) “144 ~150 -156 -162 “171 “179.5 -198 deg I =-180. TABLE-2 : Real and imaginary part of G(j) at various frequencies © rad/sec 0.35 04 0.45 05 0.6 07 1.0 Go) 1.78 -1.56 =1.37 -1.14 -0.89 0.7 0.29 Gio) -1.29 0.9 -0.61 -0.37 0.14 0 0.09 RESULT Gain margin, K= 1.4286 Phase margin, y =+12°3.83 30° 20° 10° 350° 340° 330° —330° =340° -350° o° ~10° 20° -30° 320° -40° 310° -50° 300° yf Sf 60° 60° V//L/P XK ~300° Yi Vf) 290° ~70° 280° 80° 270° ~90° 260° -100° -210° —200° —190° 180° 170° 150° 160° 170° 180° 190° Fig. 3.7.1:.Polar plot of Giia).= Mie (i+ia) (1 +i2@). (using volar coordinates).3.83 10° p -250° 1 120° -240° 130° 230° 140° ~220° ~170° & 190° ! "ss 340° ~20° 330° 30° ~-150° 210° 320° —40° 310° ~50° 300° 60° 290° ~70° 280° 80° 270° 90° 260° -100° 220° 140° Fig. 3.7.1:.Polar plot of Gia@).= Mie (+i) (+i2a). (using volar coordinates).(oz? +1) (Of +1) of + ‘(Soqutpsoos swnn8unjaas Bursn) = (OND ‘fo jojd wyog :7°L'E Buy3.45 CGGRIPOL Systems Engineering ) EXAMPLE 3.8 The open loop transfer function of a unity feedback system is given by G(s) = 1/ s*(1+8) (1+2s). Sketch the polar plotand determine the gain margin and phase margin. SOLUTION Given that, G(s) =1/ s%(1+s) (142s) 1 (jo)? (1+ jo) (1+ 20) Puts =ja, -. G(jo) = The comer frequencies are @,, = 0.5 rad/sec and ,, = 1 rad/sec. The magnitude and phase angle of G(jo) are calculated for the corner frequencies and frequencies around comer frequencies are tabulated in table-1. Using the polar to rectangular conversion, the polar coordinates listed in table-1 are converted to rectangular coordinates and tabulated in table- 2. The polar piot using polar coordinates is sketched on a polar graph sheet as shown in fig 3.8.1. The polar plot using rectangular coordinates is sketched on an ordinary graph sheetas shown in fig 3.8.2. 1 layer dle) {io = Tay (iF 0) (Fe) 1 022180° ito? Ztan“'o V1+40? Ztan'20 1 G(jo) = ——————— ¢(-180- tan“ -tan™*20) 0714+ 0? ¥1+407 iGinie ae 0? Vit 0? ¥14+407 0? (140?) (1+407) 1 0? \1+50? +404 ZG(jo) = -180°-tan“'o - tan!20. TABLE-1 :. Magnitude and phase plot of G(j@) at various frequencies oO radisec |. 0.45 O05 O58 | 06 0.65 07 0.75 1.0 T Go) 3.3 25 18 15 12 | 0971} 08 03 ep) | deg 246 -251 256 -261 ~265 —269 273 288 TABLE-2 : Real and imaginary parts of G(j@) | @ | |rad/sec | 0.45 05 0.55 0.6 065 | O07 0.75 10 | {G.go) | -1.34 -0.81 | -0.46.) -0.23 | -0.1 | ~0.02 0.04 0,09 | G(e) 3.01 236 | 1.84 1.48 1.2 1.0 0.8 0.29 | RESULT Gain margin, K, =0 Phase margin, y=-90°3.46 30° 20° 10° 350° 340° 330° =330° =340° -350° o° =10° =20° 30° 320° 40° 40° -320° 310° “50° 50° ~310° 300° 60° 60° 300° 290° =70° 70° —290° 280° sor -80° 280° 90° 270° -270° 90° 260° 100° 100° —260° 250° 110° ) —110° 250° “ig 3.8.1: Polar plot of G(jo) = 1/G@)* (1+j@) (14j20), (using polar coordinates) 240° 20 120° & —240° 230° 130° -130° 230° 220° 140° cA 140! 220° =210° 200° 190" 180° =170° =160° “150° 150° 160° 170° 180° 190° 200° 210°(sarpuyp009 snjnBunyoad Butsn) (oxly-1) (Of+-1) AOD/1 = (ODD fo sord anyog se°R'e Sty 3.47onse Analysis Frequency 3.48 EXAMPLE 3.9 The open loop transfer function ofa unity feedback systemis given by, (1+.0.2s) (1+.0.025s) qs)= As} s°(1+0.005s) (1+0.00ts) - ‘Sketch the polar plot and determine the phase margin. SOLUTION Given that Gis) =~ 028) (1+ 0.0258) _ s*(1+.0.005s) (1+ 0.001s) (1+ j0.20) (1+ j0.0250) OO) 3 4sci OGRE GOOEY (jo) (jo)® (1+ J0.0050) (1+ j0.00%) Jit (0.20)? Atan'0.20 1+ (0.0250)? Zan 10.0250 9) £270" {1+(0.0050)* Atan0.0050 1+ (0.0010)? tan"0.001e IeGoyl= t+(0.20)? f+ (0.0250)? @® J1+(0.0050)? /1+(0.00%a)? ZG{(jo) = tan“'0.20 + tan“'0.025@ — 270°-tan™'0.005e —tan™'0.00%0 The magnitude and phase angle of G(jo) are calculated for various frequencies and listed in table-T. Using the polar to rectangular conversion, the polar coordinates listed in table-1 are converted to rectangular coordinates and tabulated in table-2. The polar plot using polar coordinates is sketched on a polar graph sheet as shown in fig 3.9.1. The polar plot using rectangular coordinates is sketched on an ordinary graph sheet as shown in fig 3.9.2. TABLE-1 : Magnitude and phase of G(jo) a,radisec o9 | oss | 10] 41 | 12 14 | 47 IGto)} 14 | 12 1.0 08 06 | O04 0.2 ZGlo),deg | 259. | -258 | 257] -256 | -255 | -253 | -249 ko | | L TABLE-2 : Real and imaginaty part of G(jo) loradsec | o9 | 095 | 10 | 14 | [Goo -027 | -0.25 |-022 | -0.19 | | ao , Glo) 137 | 147 | 097 078 | RESULT Phase margin, y=-77°3.49 ~-190° 170° —150° 210° ) 280° 320° —4a° 310° 50° inates) i polar coordi 300° ~60° | ing 0010), (usi 290° eS 0050) (1+0. -80° S (140. 210° > 90° 025e)/ 260° = —100°* 0 450.20) (144 250° © —110° T ( of G(ja) 240° > -120°8 Polar plo 230° cs ~130°R Fig 3 220° 140°(sappusp4o os anjnduenjoo4 3upsn) (OL 90+ 1) (SOOO DC) /(SZO'O+1) (OTOL) = (ND Jo 1014 wrod :2°6"E Ft 3.50 Response Analysis Chapier 3.3.51 EXAMPLE 3.10 The open loop transfer function of a unity feedback system is given by G(s) = 1/s(1+s)’. Sketch the polar plot and determine the gain and phase margin. SOLUTION Given that, G(s) = 1/s(1+s)?, Puts =jo, ecy 1 7 1 00) eae Jo U¥e) (Fo) ‘The comer frequency is ©,,= 1 rad/sec. The magnitude and phase angle of G(jo) are calculated for corner frequency and frequencies around comer frequency and tabulated in table-1. Using polar torectangular conversion the polar coordinates listed in table-1 are converted to rectangular coordinates and tabulated in tabie-2. The polar plot using polar coordinates is sketched ona polar graph sheet.as shown in fig 3.10.1. The polar plot using rectangular coordinates are sketched on an ordinary graph sheet as shown in fig 3.10.2. 1 cL jo(1¥j@) jo (1+ je) (1 jo) 1 0290 yi+e? Ztan a ¥14+02 Ztan-to =—— 4-90" 2tan“v) 2 of vi+0?) poet Ha a(1+o?) o+o0° £G(jo) = -90° - 2tan'o , TABLE-1:; Magnitude and phase of GG) at various frequencies Glia) = n IG(jo)|= Psa | T | ! | } rad/sec | 04 05 06 | 07 08. | 09 1.0 11 | ee [22 | 16 | 12 | 7 [os | o6 | 05 | 04 | 2a) i | | | | | | deg | -134 | -143 | -151 | -159 | -167 ~174 | 180 | -185 i L TABLE-2 : Real and imaginary parts of. G(j@) at various frequencies Ta ] 1 | a | | ;. {| rad/sec 04 05 | 06 07 | 08 | 09 | 10 11 | G60) 153 | -128 | -1.05 | -093 | 078 | 06 | 05 | 04 | Gfe) -1.58 | -0.96 | -0.58 | -0.36 | 0.18 | 006 | 0 0.03 | RESULT Gain margin, K,=2 Phase margin, y=21°50° 310° 60° 300° 110° -250° 120° ~240° 130° 230° 140° -220° 200° 160° —190° 170° —180° 180° 350° —10° 170° 190° 160° % 200° ©, ve 330° -30° CS -150° 210° 220° —40° 310° -50° 300° -60° 290° mea: 280° -80° 270° -90° 260° ~100° 250° 110° 240° —120° 230° —130° 220° —140° Polar plot of G(jw) = 1/[ja (1+jo)"] (using polar coordinates) Fig 3.10.13.53 Fig 3.10.2: Polar plot of GG@) = 1[j@ O4j@)"] (using rectangular coordinates)chapter 3 ~ Frequency Response Analysis ) EXAMPLE 3.11 Consider a unity feedback system having an open loop transfer function G(s) = s Sketch the polar plotand determine the value of K so that (i) Gain margin is 18 db (ji) Phase margin is 60°. SOLUTION 3 K a Given that, G(s) = ——-—~_____. TI Yt ing K=1. iven that, G(s) 5 (iF0.2s) (140.088) he polar plotis sketched by taking K =7 1 «.PutK=1 ands =jo in G(s). -. G(jo) = ———_——_____—_ ul ands=joinG(s). -.G(jo) jo G*]0.2a) (i ].050) ~ (1+ 0.28) (1+ 0.05s) ° The comer frequencies are o,, = 1/0.2= 5 rad/sec and o , =1/0.05 = 20 rad/sec. The magnitude and phase angle of G(jo) are calculated for various frequencies and tabulated in table-1. Using polar to rectangular conversion the polar coordinates listed in table-1 are converted to rectangular coordinates and tabulated in table-2, The polar plot using polar coordinates is sketched on a polar graph sheet as shown in fig 3.11.1. Polar plot using rectangular coordinates is sketched on an ordinary graph sheet as shown in fig 3.11.2. 1 © j@ (14 j0.20) (1+ j0.05e) i @Z90° {1+(0.20)? Ztan'020 1#(0.050 ¥ Ztan'0.05a 1 a 7-90" ~ tan "0.205 — tan '0.050) 0 ¥1+(0.20)* 1+ (0050)? 1 + [G(jo)= ———_—_——_— © 1+(0.20)? 1+ (0.050)? TABL Magnitude and Phase of G(jo) at Various Freque: S G(jo) and £G(jo)=—90° — tan™'0.20 — tan“'0.050 T i o 1 | i { | radisec | 06 08 1 | 2 3’ | 4 | (G0 [165 | 123 | 10 | 08 | 03 | 02 | 2G) | | | | |. deg | 98 101 ~to4 | 1175 | 1294 | -140 @ | T rad/sec 5 6 7 9 10 te od, oe 1 G60) 0.14 01 0.07 0.05 0.04 0.03 | 0.02 | ZGho) | | deg | 149 | 157, | -t64 | -176 | -180 | 184 | 195 | inary Parts of G(j@) at Various Frequencies 1 o | | rad/sec 0.6 08 1 | 2 | 3 | 4 | t 1 Gey 0.23 0.23 024 | 023 | 019 | 015 | P | mA el | Ge) 1.63 =a) | -0.97 0.44 0.23 | 0.13 |3.55 a | | | | radisec 5 | 6 7 9 | 10 Mo, 4 Gwe) | -0.120 | -0.092 | -0.067 | -0050 | -004 | -o.030 | -oo19 | Gf) |_0.072 | -0.099 | -0.019 | -00034 | 0 0002 | 0.005 | In the polar plot shown in fig 3.11.1 and 3.11.2 there are two plots, marked as curve-| and curve-ll. These two loci are sketched with different scales to clearly determine the gain margin and phase margin. From the polar plot, with K= 1, Gain margin, K, = 1/0.04 = 25. Gain margin in db = 20 log 25 = 28 db. Phase margin, y= 76°. With K = 1, let G(@j@) cut the -180° axis at point B and gain corresponding to that point be G,. From the polar plot G,=0.04. The gain margin of 28 db with K = 1 has to be reduced to 18 db and so K has to be increased toa value greater than one. Let G, be the gain at-180° for a gain margin of 18 db. 1 1 18 1 19/20 Now, 20 log —=18 log — =—— ——=10 3G, = Ea FG] 1 Gq = sora = 0125 G 0.125 f Kis given by, K = —* =—— = 3.12: The value of Kis given by, G, 004 5 Case (ii) With K = 1, the phase margin is 76°. This has to be reduced to 60°. Hence gain has to be increased. Let: a2 be the phase of G(jo) for a phase margin of 60° 1, 60" = 180°+6,., qcq = 60° — 180° = -120° Inthe polar plot the -120° line cut the locus of G(jc) at point C and cut the unity circle at point D. Let, G,= Magnitude of Ga) at pointC. G, = Magnitude of G(ja) at point D. From the polar plot, G, = 0.425 and G, = 1. Cy 4 => 1 __ 9353 Nowe GP ae RESULT @ When K=1, Gain margin, K, = 25 Gain marginindb = 28db ) When K = 1, Phase margin, y = 76° © Fora gain margin of 18db,K = 3.125 @ Fora phase margin of 60°, KK = 2.353(sajpurpuooa dpjod Bujsn) (oso'of+1) (7 0F+1) Of = (DD fo1ojd jog : VII's Ba 2 ee 29 20 ad ae as 6. Be eS =a mit pa 8 omy a Vie Oe ee Nelsen Ee Seal VAN sate bs os (I I fH TOSS i PROSE" BAS 9 eI VGN3.57 Fig 3.11.2: Polar plot of G(j@) = 1/jo (1+j0.2e) (1+j0.050), (using rectangular coordinates)EXAMPLE 3.12 Consider a unity feedback system having an open loop transferfunction, G(s) = ed . Sketch the polar plotand determine the value of K so that (i) Gain margin is 20 db and (ii) Phase margin is 30°. SOLUTION Given that, G(s) = K/s (1+0.5s) (1+4s) ‘The polar plotis sketched by taking K =1. PutK =1 ands =jain G(s). sen”. 1 * Se) =o GF faba) Ho) The comer frequencies are ,, = 1/4 = 0.25 rad/sec and w_, = 1/0.5 = 2 rad/sec. The magnitude and phase angle of G(j@) are calculated for various frequencies and tabulated in table-1. Using polar to rectangular conversion the polar coordinates listed in table-1 are converted to rectangular coordinates and tabulated in table-2. The polar plot using polar coordinates is sketched ona polar graph sheetas shown in fig 3.12.1. The polar plot using rectangular coordinates is sketched onan ordinary graph sheet as shown in fig 3.12.2. 1 Gti oe a Noy lo) = io (iF ]050) (Hay _ 1 0290° 1+ (0.50)? Ztan“0.5a J1+(40)? Ztan~ 4 1 = 4-90 ta 0.50 - tan 40) oV1+0.250? V1+1602 : 1 “/G(jo ) = o V1+0.250? 141602 4G(jo) =-90° - tan"'0.5@ - tan 4o TABLE-1 : Magnitude and Phase of GG) at Various Frequencies T T o i rad/sec 0.3 04 05 06 08 10 | 12 Glo 241 13 0.87 0.61 0.35 o22 | 0.15 2G(p) deg 149 -159 | -167 -174 | 184 ~193 199 | TABLE-2 : Real part and Imaginary parts of G(j@) at Various Frequencies a r.1.aTs? al |_tad/sec 0.3 04 05 | 06 | 08 | 10 | 12 | | Gao) -18 | -121 | -085 | -o81 | -035 | -021 | -o14 3) 109 | -047 | 02 | ~0.08 | 0.92 0.05 From the polar plot, with K= 1, Gain margin, K, = 1/0.44=2.27 Gain margin in db= 20 log 2.27 = 7.12 db Phase margin, y= 180°+ $,, = 180°— 165° = 15°LE == no Menai LTRS LE ti a Z NG Sr SSFig 3.12.2: Polar plot of Gj@) = 1/fja (1+)0.5@) (14j40)], (using rectangular coordinates).3.61 Case (i) . With K =, let G(j) cut the -180° axis at point B and gain corresponding to that point be G,. From the polar plot, G, = 0.44, The gain margin of 7.12 db with K= 1 has to be increased to 20 db and so Khas to be decreased to a value less than one. Let G, be the gain at—180° for a gain margin of 20 db. 1 Now, 20 log —=20 low, 20 log a, fog 2 28.24 5G," 207 1 1 —=10'=10 Gy 1 Geo a Gy=h=04 04 G The value of K is given by, K=—4 = = 0.227 ee SEM VAL Case (ii) With K = 1, the phase margin is 15°. This has to be increased to 30°. Hence the gain has to be decreased. Let x0 be the phase of G(j@) for a phase margin of 30°. + 30°= 180° +6,, yg = 30° — 180° = -150° . In the polar plot the -150° line cuts the locus of G(jo) at point C and cut the unity circle at point D. Let, G, = Magnitude of G(ja) at point C. G, = Magnitude of G(ja) at point D. From the polar plot, G, = 2.04 andG, = 4 Now, Kage a 0.49 RESULT @ When K = 1, Gain margin, kK, = 2.27 Gain margin in db = 7.12 db (b) When K = 1, Phase margin, y = 15° (©) Fora gain margin of 20 db,K = 0.227 d@ Pov’ dhasa margin of 30°,K = 0.49 3.8 NICHOLS PLOT The Nichols plot is a frequency response plot of the open loop transfer function of a-system. The Nichols plot is a graph between magnitude of G{je) in db and the phase of G(ja) in degree, plotted on a ordinary graph sheet. In order to plot the Nichols plot, the magnitude of G(ja) in db and phase of G(ja) in deg are computed for various values of @ and tabulated. Usually the choice of frequencies are corner frequencies. Choose appropriate scales for magnitude on y-axis and phase on x-axis. Fix all the points on ordinary graph sheet and join the points by smooth curve, and mark frequencies corresponding to each point.© Ghapler Fz Frequency Response Analysis) 3.62 In another method, first the Bode plot of G(ja) is sketched. From the Bode plot the magnitude and phase for various values of frequency, @ are noted and tabulated. Using these values the Nichols plot is sketched as explained earlier. DETERMINATION OF GAIN MARGIN AND PHASE MARGIN FROM NICHOLS PLOT The gain margin in db is given by the negative of db magnitude of G(ja) at the phase crossover frequency, ©,.-The o,. is the frequency at which phase of G(ja) is -180°. If the db magnitude of G(ja) at @,. is negative then gain margin is positive and vice versa. Let ,, be the phase angle of G(j@) at gain cross over frequency @,,. The @,. is the frequency at which the db magnitude of G(ja) is zero. Now the phase margin, y is given by y =1 80°+$,.. Ifo, is less negative than —180° then phase margin is positive and vice versa. The positive and negative gain margins are illustrated in fig 3.27. +db +db t Positive GM t t pete |GGoy |Gde)| 0 in. in db ar db : ~db #—* | Positive db | ; + 1PM i Renae : PI -360° -270° -180° %,-90° 0° -360" -270° ¢. —180° -90° o° Fig 3.27 : Nichols plot showing phase margin (PM) and gain margin (GM). GAIN ADJUSTMENT IN NICHOLS PLOT In the open loop transfer function, G(jw) the constant K contributes only magnitude. Hence by changing the value of K the system gain can be adjusted to meet the desired specifications. The desired specifications are gain margin and phase margin. Ina system transfer function, if the value of K required to be estimated, in order to satisfy a desired specification, then draw the Nichols plot of the system with K=1. The constant K can add 20logK to every point of the plot. Due to this addition, the Nichols plot will shift vertically up or down. Hence shift the plot vertically up or down to meet the desired specification. Equate the vertical distance by which the Nichols plot is shifted to 20logK and solve for K. Let, x = change in db (x is positive if the plot is shifted up and vice versa). Now, 20logK=x => logK= => sK=108 EXAMPLE 3.13 K(1+ 10s) s? (1#5) (1+ 2s) plotand determine the value of K so that (i) Gain margin is 10db, (ii) Phase margin is 10°. SOLUTION Consider a unity feedback system having an open loop transferfunction G(s) = . Sketch the Nichols K(1+ 10s) Given that G(s) = Sess 28)5.8 1GGoy et ® : i | {to po foci ft a 200° -180°_179° -160° —, -140° —120° -100° ZG (i K(1+ j10) indeg Go) (1+ joj + j20)° Per yy 20° Fig 3.13.1 : Nichols plot of G(jo) = 0. 240The sinusoidal transfer function G(ja) is obtained by Jetting to s = jo. Also put = 1. “ehi6} 2 (14100) . H+(100)° Ztan-10e a fee % s (oy"(1+ Jo) 20) @?2180°V1+0? ctanwy1+ (20) Ztan"'20 ieio)|= 441000? vfi+1000? ————S <.|G(Qja))_ = 20 tog | —~ orfteo? V1440? ee of1+0? 1+ 407 2G(jo) =tan-10 - 180° - tan"o -tan™‘20 ‘The magnitude of G(jo) in db and phase of G(jo) in deg are calculated for various values of @ and listed in the following table. The Nichols plot of G(jw) with K = 1 is sketched as shown in fig 3.13.1 T oO | | | 1 | radisec | (02 | 04 | 08 | 08 | 10 | 15 | 20 | 30 [40 | | ee | | | db | 341 25.4 19.3 143 cae | 14 5.3 1. 15.2 | -225 | ep) | | | | | j deg | —150 | -164 | -181 | -194 | -204 | -222 | -232 | -244 | -250 | | From the Nichols plot the gain margin and phase margin of the system when K=1 are, Gain margin Phase margi Gain adjustment for required gain margin For a gain margin of 10 db, the magnitude of G(jo) should be -10db, when the phase is -180°. When K = 1, the magnitude of G{jo) is +19.5db corresponding to phase angle of -180°. Hence if we add-29.5 db to every point of G(ja) then the plot shifts downwards and it will cross —180° axis ata magnitude of -10db. The magnitude correction is independent of frequency and So this gain can be contributed by the term K. Let this value of K be K,. The value of K, is calculated by equating 20logk,to-29.Sdb. f. -295 = © Dlogky=295db = egKy=—S> > Ky=10 ® = 00335 Gain adjustment for required phase margin Leto,..= phase of G(ja) at gain crossover frequency for a phase margin of 10° ..Phase margin, y,= 180° Fy 2 gca= Yp7180° = 10°-180° =-170° . When K = 1, the magnitude of G(jo) is +23 db corresponding to a phase of—-170°. But fora phase margin of 10°, this gain should be made zero. Hence if we add —23db to every point of G{ja) locus then the plot shifts downwards and it will cross -170° axis at magnitude of 0 db. The magnitude correction is independent of frequency and so this gain can be contributed by the term K. Let this value of K be K,. The value of K, is calculated by equating 20 logK, to—23db. j BB 1-20 log Kz=-23 = logK,=-23/20 =» k,=107 =007 RESULT @ When K=1, Gainmargin = -19.5db Phase margin -45° ) Fora gain margin of 10db, K=K, 0.0335. (©) Fora phase margin of 10°, K=K, = 0.073.6 3.9 CLOSED LOOP RESPONSE FROM OPEN LOOP RESPONSE The closed loop transfer function of the system is given by, @___ 6) RG) 1266) He The sinusoidal transfer function is obtained by replacing s by jo. ~ MGa)= Senay Let, M(jo) = MZa where, M = Magnitude of closed loop transfer function a = Phase of closed loop transfer function. The magnitude and phase of closed loop system are functions of frequency, @. The sketch of magnitude and phase of closed loop system’ with respect to @ is closed loop frequency. response plot. The magnitude and phase of closed loop system for various values of frequency can be evaluated analytically or graphically. The analytical method of determining the frequency response involves tedious calculations. Two graphical methods are available to determine the closed loop frequency response from open loop frequency response. They are, 1. Mand N circles 2. Nichols chart. 3.10 MAND NCIRCLES The magnitude of closed loop transfer function with unity feedback can be shown to be in the form of circle for every value of M. These circles are called M-circles. If the phase of closed loop transfer function with unity feedback is a, then it can be shown that tan & will be in the form of circle for every value of a. These circles are called N-circles. The M and N circles are used to find the closed loop frequency response graphically from the open loop frequency response G(j) without calculating thé magnitude and phase of the closed loop transfer function at each frequency. The M and N circles are available as standard chart. The chart consists of M and N circles superimposed on ordinary graph sheet. Using ordinary graph the locus of G(ja) (Polar Plot) is sketched. The locus of G(jo) will cut the M-circles and N-circles at various points. The intersection of G(ja) locus with M and N circles gives the magnitude and phase of the closed loop system at frequencies corresponding to the cutting point of G(§ja). The M and @ for various values of « are tabulated. The magnitude and phase response of closed loop system are sketched on semilog graph sheet by taking @ on the logarithmic scale-on x-axis. [The closed loop frequency response has two plots. They are M Vs @ and a Vs @]3.66 M-CIRCLES Consider the closed loop transfer function of unity feedback system, M(s) = aaa Puts = jo, 2 M(jay= 7-3 oem Let, GG) =X+jY where, X= Real part of G(jo). Y = Imaginary part of G(ja). : PY? tan Y “.M(o)= 7 = B= (1S tat 2S) 14+X+jY (+x 2 tant (i+xy+y¥? + Let, M = Magnitude of M(jo) vx? + ¥? i+ X)?+¥? On squaring the above equation we get, “M= po Ys M(ixfey)=x ey Ss M(14X742X4Y2)=x? + ¥? (1+ X)° + Y° M? + M?X? + M?2X + M?Y?—X?—Y?=0 X? (M71) + M22X4+MP+¥2?(ME1)=0 ae (3.30) When M = 1, the equation (3.30) represents a straight line. When M = 1, the equation (3.30) is, “X?(1-1)+2X+1+Y2(1-1)=0 => 2X+1=0 “> X=-1/22 Hence when M = 1, equation (3.30) represents a straight line passing through X =-1/2 & Y=0. When M # 1, the equation (3.30) represents a family of circles. When M = 1, equation (3.30) can be rearranged in the form of equation ofa circle as shown below. XX? (M21) + M?22X + M? + Y? (M?-1)=0 Divide the above equation throughout by (M?- 1).3.67 2 Add ey on both sides of the above equation. -1 Me X+ M?-1 a? + 2ab + b? = (a+b? mw YP M2 X+—3 +¥ =——_ “ soee(3.31) M*-1 (m2 a 1) The equation of circle with centre at (X,, Y,) and radius r is given by, (X-XYPRQAY Pr (3.32) On comparing equation (3.31) and equation (3.32), it can be concluded that the equation (3.31) represents a family circles with centre at (-M?/M?-1), 0) and with radius, r = M/(M2-1) for various values of M. The circles given by equation (3.31) are called M-circles. When M=0 When M = 0 Centre = (X,.¥,) Centre = (X,,¥,) 2 Mi. 6 =0 Radius r=-—_M_.M 11 a M’-1 M° M © Hence when M = 0, the magnitude Hence when M =~, the magnitude circle becomes a point at (0,0). circle becomes a point at (—1,0). From the above analysis it is clear that the magnitude of closed loop transfer function will be in the form of circles when M #1 and when M = 1, the magnitude is a straight line passing through (-1/2,0). For values of M less than 1, the magnitude is a circle to the right of the straight line corresponding to M = 1. It is observed that the circles for M<1 passes through (—1/2,0) and (0,0) on the negative real axis. For decreasing values of M, the radius decreases and the circle, becomes a point at (0,0) when M = 0. For values of M greater than 1, the magnitude is a circle to the left of the straight line corresponding to M = 1. It is observed that circle passes between the points (—1,0) and (—1/2,0) on the negative real axis. For increasing values of M the radius decreases and the circle becomes a point at (1,0) when M=~. The family of M-circles are shown in fig 3.28.iv Complex plane Rectangular coordinates Fig 3.28 : The family of constant M-circles. N-CIRCLES Consider the closed loop transfer function of unity feedback system. Gs) __G@) Ro 10 = --__GUia) Put s=ja, MU) = TG Gay Let G(j@)= X + jY, where, X = Real part of G(jw). . Y = Imaginary part of G(jo). 2,2 ak . X+i¥ tee AS Vx?+¥? aY¥ 4 Y¥ “. M(Gja)= —= y= Z| tan —tan 14+X+jY (+x) +Y? Ztan™ +xy+¥? x 1+X 1+X Let, a = Phase of M(ja) ; sont ag x 14+X Let, N= tana Nota! —tar-t 2) x 14+X san( tn)? | ¥ 2¥ YG+X)— XY uae x 1+X) __X 14x ___X(1+X) iste a ¥ yY “a 2 14 tanf tan~!Y tan! tan UY) Ye Xe x 1+x/) X 14#X X(1+X) xen +Y X+X°4+¥ tanA — tanB an tan(A -B)=$—————— = — 1+tanA x tanB X+X°+¥3.69 On rearranging the above equation we get, Ke yy22% ox4xeyviteo N 2 In the above equation add the term (4) on both sides. 2 2 xox2e yi yy Es ai, a N 4 \2N 4 \2N (e+fex}[e+ ! ahs 4 Guy ON cc 2 a xd) +(y-b) 4+ 8 IN) 4" ae The equation of circle with centre at (X,,Y,) and radius r is, sexed (3.34) (XX, +(Y-¥,F = On comparing equation (3.33) and (3.34), it can be concluded that the equation (3.33) represents a 1 1 family of circle with centre at (-1/2,1/2N) and with radius 14" any for various values of N. The circles given by the equation (3.30) are called N-circles. : For any value of N, the equation of N-circles is satisfied at two points (0,0) and (—1,0). Hence the N-circles passes through these two points for all values of a. (N = tan «). Consider the equation of N-circle, ‘When X =~1 and Y= 0, .---(3.33) Consider the equation of N-circle, When X = 0 and Y= 0, 2 2 4 2 1 1 11 1 1 1, 1 (xe + v-) =a X+=| +|/ Y-——| =—+——} \ 3] ( 2N) 4° (Ny ( 2 2N) 4° QNY 2 e 2 a. 1 1 ro Gy+-Ay-t+ aed) of EP 2. 2N 4 (2N) 2 2N 4° (4Ny? t,t Bd. dg le dy 1 4 4N? > 4. 4N? The above analysis shows that the equation of N-circle is satisfied at points(0,0) and (-1, 0) When a = 180° the circle becomes a straight line passing through real axis. It is also observed that the circle for a = 8°-180° above the real axis will be a part of circle for a = 6 ° below the real axis, as shown in fig 3.29. The family of N circles are shown in fig 3.30. a = 0,-180° Fig 3.29.3.70 Complex plane Rectangular coordinates ie a=-20° Fig 3.30 : The family of constant N-circles. 3.11. NICHOLS CHART N.B Nichols transformed the constant M and N circles to log-magnitude and phase angle coordinates and the resulting chart is known as Nichols chart. Nichols chart consist of M and N contours, superimposed on ordinary graph. The M contours are the magnitude of closed loop system in decibels and the N contours are the corresponding phase angle locus of closed loop system. The ordinary graph consist of magnitude in db marked on the Y-axis and the phase in degrees marked on the X-axis. The Nichols plot of open loop system can be plotted on the ordinary graph. The Nichols plot is a graph between magnitude of G(j@) in db and the phase of G(jw) in degree,plotted on an ordinary graph sheet. To draw the Nichols plot the magnitude and phase angle of G(jo) are calculated for various values of @. Alternatively the Bode plot of G(j@) is sketched and from Bode plot, the magnitude and phase of G(jo) for any frequency can be obtained. Using Nichols chart the closed loop frequency response can be determined graphically from the locus of open loop frequency response. When the Nichols plot of GG) is sketched on Nichols chart, the locus of G(@j@) will cut the M and N contours at various points. The cutting point of locus of G(ja) with the M-contour gives the magnitude of closed loop frequency response corresponding to a frequency same as that of G(jq) at that point. ‘ The cutting point of locus of G(j@) and N contour gives the phase of closed loop frequency response corresponding to a frequency same as that of G(ja) at that point. The magnitude M and phaseFah angle a (N = tana) of closed loop system are tabulated . The closed loop frequency response consists of two plots. They are magnitude M Vs @ and phase angle « Vs @. Hence using the tabulated values the bode plot of closed of closed loop system can be drawn. The'frequency domain specifications can be determined from Nicols chart. Fig 3.31, shows various frequency domain specifications of a typical G(ja) locus. Also the Nichols plot drawn on a Nichols chart can be used for gain adjustment. ESTIMATION OF FREQUENCY DOMAIN SPECIFICATIONS USING NICHOLS CHART Resonant Peak (M) and Resonant Frequency (@) The resonant peak is given by the value of M-contour which is tangent to G(j@) locus. The resonant frequency is given by the frequency of G(ja) at the tangency point. Bandwidth The Bandwidth is given by frequency corresponding to the intersection point of G(j@) and -3 db M -contour. Gain Margin The gain margin is given by negative of magnitude of G@ja) in db at phase crossover frequency, @,,. At phase crossover frequency the phase of G(j@) is —180° Gain Margin, K, indb= -GGo,)),, ia Phase Margin The phase margin,y is given by y =180° +, where o is the phase of G(ja) at gain crossover frequency. At gain crossover frequency the magnitude of G(ja) is-zero db. GAIN ADJUSTMENT USING NICHOLS CHART Determination of K for Specified Gain Margin Draw the G@j@) locus with K=1. Determine the amount of gain to be added at 6 =—180°, so that db magnitude of G(j@) locus at —180° is negative of the specified gain margin. Let the db gain to be added be x db. The gain contribution is independent of frequency and so it can be achieved by choosing proper value of K. The value of K is obtained by equating 20logK to x db. Now, 20logK = x 1K =10% Determination of K for Specified Phase Margin Draw the GG@) locus with K= 1. The phase margin, y =180° *O,, where acs is phase of G(ja) at gain crossover frequency. ..$,.= y -180°. For specified phase margin, calculate $, and from the Nichols plot determine the db gain at 6... Let this gain be y db. For the specified phase margin, this gain should be made zero. Hence —y db should be added to every point of G(j). This is achieved by choosing proper value of K. The value of K is obtained by equating 20logK to -y db. a Now, 20lgk=-y => logk =~ => K=10%Gain Margin -12db 180° Se go? o k—————| Phase Margin Fig 3.31 : Determination of frequency domain specification from Nichols Chart. Determination of K for Specified Resonant Peak, M, Draw the G(jw) locus with K = 1. Using a tracing paper, trace the locus of G(j@). A standard tracing paper, called Nichols overlay is available). Then shift the locus vertically up or down, so that M = M_ contour is tangent to GGa) locus. Measure the vertical shift in db. Let the shift be + x db. (+ for up and —for down). Now, 20logK=tx => logk = i= > K-10» Determination of K for a Specified Bandwidth Draw the G(ja) locus with K=1, Determine the open loop gain G(ja) at @ = @, where, o, is the specified bandwith. Determine the point of intersection of -3db M-contour and this open loop gain on the Nichols ‘chart. Let this point be point A. Trace the G(j@) locus. Shift the G(j@) locus vertically up or down, so that it passes through point A. Measure the vertical shift in db. Let the shift be + x db Gfor up and —for down). : Now, 20logK=#x => logk =25° > K-10 EXAMPLE 3.14 The open loop transfer function of unity feedback systemis, G(s) = Ke®/s(1+0.25s) (1+0.1s). Using Nichols chart, determine the following. @. The value of K so that the gain margin of the system is 4 db. (0) The value of K so that the phase margin of the system is 40° () The value of K so that resonant peak M_ of the system is 1 db. What are the corresponding values ofa, ando,? (@® The value of K so that the bandwidth o, of the system is 1.5 rad/sec.3.73 SOLUTION First the actual bode plat of G(jo) with K=1 is plotted on semilog graph sheet. The magnitude of G(ja) in db and phase of GC) for various frequencies are calculated and listed in Table-1. The choice of frequencies are chosen such that the magnitude plot extends in the range of 40 db to -14db and the phase plotextends in the range of 0° to—180°. Ker025 (1+ 0.25s)(1+ 0.18) Let, K=1 and put, s=jo. Given that, G(s)= eee 12-0.20 x 180° + GGw) = = x jo(1+0.258)(140.1s) ©290°V1+ 0.062507 tan10.250 71+ 0.010? Ztan“'0. 10 1 AGG) = >————5 ==; --. IGi@) ng = 20!0g} wT 0.0625071+ 0.070" ° oy 1+. 0.062502J12 ZG(o)=-0.20 x 82 _ 96° —tan0.250 — tan—10-0 TABL Calculated values of |G(j@)! and “G(j@ ® rad/sec | 0.01 | 0.02 | 005 | 01 | -02 | o5 | 10 | 20 | 40 eo) | db 40 | 3% | 26 | 20 4 6 0 a_| -16 44) | -90 | -91 | -91 | -93 | -96 | —106 | 121 | 151 | 203 deg ! . ‘The magnitude and phase plot of Bode plot of G(jo) are shown in fig 3.14.1 From the bode plotthe phase and frequency for various values of magnitudes are noted and tabulated in table-2. (The choice of magnitudes are 20,16, 12,.... i.e, in steps of 4-db, which is convenient for Nichols plot on Nichols chart). Using the values listed in table-2 the locus of G(jo) on Nichols chart is sketched as shown in fig 3.14.2. TABLE-2: Values of |G(jo)| and ZG(j@) Noted from Bode Plot o fad/sec | 0.1 0.16 0.25 0.4 0.64 1.0 15 22 3.0 [GGo)} 20 16 12 8 4 9 -4 -8 -12 db ——_|—. ZG (0) -90 +92 96 102 | -110 | -120 | -136 | -156 | -180 deg i Gain Margin and Phase Margin when K=1 When K = 1, the G(jo) locus cuts the -180° axis at —12db. Hence the magnitude atphase crossover frequency is—12db. ~.Gain Margin, K,=AGG0, Ming =—H12) = +12db. When K = 1, the phase of G(ja) is -120° corresponding to magnitude of 0 db. Henee the phase at gain crossover frequency is -120°, --Phase margin, y= 180° +0,,= 180°— 120° = 60° To find K for a gain margin of 4db ‘When gain margin is 4 db, the locus of G(ja) should cross the 180° axis at—4 db. When K= 1, the magnitude of G{ja) is -12db corresponding toa phase of 180°. Hence ifwe add, —4— (-12)=8 db to every paint of G(jo) then the plot shifts upwards and crosses -180° axis at —4 db. This magnitude correction is achieved by choosing appropriate value of K. The value of Kis obtained by equating 20 log K to 8 db.a er 3 = Fre Re Analysis: 3.74 8 <. 0logK = 20db => look = 5 > K=10% =25 The G(ja), when K =2.5 is shown in fig 3.14.2. To find K for phase margin of 40° Let $4, be the phase of G(jw) at gain crossover frequency when the phase margin is 40°. Phase margin, y,= 180° + a2 10g g~ 180° = 40*— 180° = ~140° From the above calculation it is evident that for a phase margin of 40°, the magnitude of G(jo) should be 0 db corresponding to a phase of -140°. When K = 1, the magnitude of G(ja) is -Sdb corresponding to a phase of -140°. Hence ifwe add +5 db to every point of G(ja) locus then the plot shifts upwards and crosses ~—140° axis at 0 db. This magnitude correction is achieved by choosing appropriate values of K. The value of Kis obtained by equating 20 logK to 5 db. 5 “20lgK=5 => logk == > K=107 =178 ‘The G(ja) locus, when K = 1.78 is shown in fig 3.14.2 To find K for a resonant peak of 1 db The resonant peak, M. is given by M-contour which is tangent to: G(je) locus. When K = 1, the G(ja) locus is tangentto M= 0.25 db contour. Hence when, K = 1, resonant peak is 0.25 db. For a resonant peak of 1 db, the M=1 db contour should be made tangent to G (jo) locus. For this, G(ja) locus can be Shifted vertically up or down so that it becomes tangentto M= 1 db contour. In this Problem the G(ja) locus is shifted vertically up to make ittangent to M= 1 db contour. The shifted G(ja) locus is shown in fig 3.14.3. Note: Trace the G(jw) locus when K = 1 ona tracing paper and shift the traced locus over the Nichols chart vertically | so thatitis tangent to required M-contour. By keeping the. tracing paper at the shifted position darken the traced locus, so that} | itmakes an impression on nichols chart. The vertical shiftis equivalentto adding a magnitude of 20 log K to every point of G(jo) locus. From the shifted locus of Gj) itis observed that +2db is added to every point of ‘G(ja) locus. Hence the value of K is obtained by equating 20 log K to +2 db. 2 2logK = 2db => logk = 2 > K=10% = 126 The resonant frequency, @, is given by the frequency of G(ja) atthe ‘tangency point. The magnitude of G(ja) is 0 db at the tangency point of M= 1 db contour. The corresponding frequency is noted from the bode Plot of G(jw). From the bode plot the frequency at 0 db is 1.0 rad/sec. Hence the resonant ‘frequency, @,= 1.0 rad/sec. To find K so that o, rad/sec Se ene Ee The bandwidth, a, is given by the frequency of G(ja) corresponding to the meeting point of G(ja) locus and M =—3db contour. From the bode plot find the magnitude of G(ja) when @ = 1.5 rad/sec. From fig 3.14.1 itis observed that magnitude of G(jo) is — 4db when @ = 1.5 rad/sec. In the Nichols chart, find the point where.the M = —3db contour passes through —4db line. Let this point be P. Now the G(ja) locus with K = 1 is shifted vertically down so thatis passes through point P. The shifted G(jo) locus is shown infig 3.1.3. Note ; Trace the G(jo) focus when K = 1 ona tracing paper and shift the raced locus over Nichols chartso thatis} passes through point P. By keeping the tracing paper at the shifted, (position, darken the traced locus, so that is makes an émpression on Nichols chart. ‘The vertical shiftis equivalent to adding a magnitude of 20logK to every paint of G(jo) locus. From the shifted locus of G(jo) itis observed that-6db is added to every point of G(ja) locus. Hence the value of Kis obtained by equating 20logK to-6 db. 20logK =-6db => log K2-5 > K=10-° -05‘Cont 3.75 @r0l+ D @scOl + D Gory.? = @ND Jo tod apog + r91'€ Stat A98/PRE UE a or OCT SLOT pororo S70 oto TO. . $00 coo 100 168L9 S$ £ t 168L9S ¥ t c 1o8L9 5 F t a 16829 $ £ t 168.9 $s ¢ Of- Oz 1 oe 1 CASe TS FHOSN 0e— 180° ~ 160° ‘ Y h = a KY ZEL tS poaeaeee FOS y if UY J Hy y Y OS ~ 140° = 120° 1, Oe ES | MES DY Sc Sa KASS SOK, YUN be CTs V = - 20° o ZGQq) in deg > Fig 3.14.2 : Nichols plot of Ga) = Ke#”?go(1 + j0.25a) (I + j 0.10)- 180° — 160° ZG(ja) in deg > Ke¥?°Gja(T + 70.250) (1 +5 0.10) Fig 3.14.3 : Nichols plot of Gia)EXAMPLE 3.15 2 , 20 Aunity feedback system has open loop transfer function, G(s) = 3rd) G5) . Using Nichols chart, détermine the closed loop frequency response and estimate M,, o, and ,. SOLUTION Given that, 20 =a ) = Sey eH t ‘The transfer function G(s) is converted to time constant or bode form. Gls) = 20 _ _201(2x5) 2 sx S +1) xa( +1} s(i+3) (143) So(ThOSS)) (120-25) Put, S=jo, in G(s) to get G( jo). 2 jo(1+ j0.50)(1+ j0.20) 2 0290? ¥1+ 0.250? /tan-"0.50¥1+ 0.040 Ztan0.20 2 So 4-90-1050 — ta 0.20) oy 140.250? 71+ 0.0407 2 oV1+0.250? 1+ 0.040? : 2 Cle), 4 = 20leg| Iota | oJiv0250? Vi 0040? ZG(jo) = — 90°-tan“*0.5@ - tan*0.20 - G(je) = +: [@Go)| " ‘The magnitude of G(jc) in db and phase of G(ja)for various frequencies are calculated and listed in table-1. The choice of frequencies are chosen such that the magnitude plot extends in the range of 40 db to-14 db and the phase plot extends in the range of 0° to—180°. Using table-1, the actual bode plot of G(ja) is plotted on semilog graph sheet, as shown in fig 3.15.1. TABLE-1: Calculated values of |G(jo)| and ZG(jo) a, rad/sec 02 05 1.0 20 3.0 40 [G()), db 20 12 5 4 -10 —15 £G(jo), deg 98 -110 | +128 | +157 | -177 | -192 ‘The magnitude and phase plot of bode plot of G(je) are shown in fig 3.15.1 From the bode plot, the phase and frequency for various values of magnitudes are noted and tabulated in table-2. (The choice of magnitudes are 20, 16, 12... ., Le., insteps of 4 db, which is convenient for Nichols plot on Nichols chart). Using the values listed in table-2, the locus of G(ja) is sketched on the Nichols chart as shown in fig 3.15.2.[@Z0!+ D ES Of+ Dole = @Do-fo rod epog : VST" Sta 99S/peLUl << © oor TENTS SOLPLOSO 160 70 “10 ap (us 4] @D5| oc SI- or a oT 0t It tl | | T68i9.6 b © ©. T6sioo b € t 168L9S b c 168495 6 ct T6O8L9S bp < tFig 3.15.2 : Nichols plot of Gia) = 2/{j@ (1 + j0.5a) (1 + j 0.20). va | ‘ a7 gla 3tg—4a—r¢ A ea i BITES ¥ bea aa A ds — 420° = 100" - 80° 60 = 40" - 20° ° Phase ZG(jo) indeg >5 6 as/pet Ul <— o loseo sh & 7% [@c0f+ D @S0f+ DOME = @)O fo asuodsas fouanbarf dooy posed + ESTE Sta oor % ‘@ 01 10 1 16895 + 168L9s ¢ £ ct 16849 S$ 7 € c 16O8L9% fF £ a i ap uyChapter 3-- Frequency Response Analysis 2 3.82 TABLE-2: Values of |G(j@)| and ZG(j@) Obtained from Bode Plot @ iadisec 0.2 0.31 O5 0.74 1.05 5S | 2.0 26 3.2 Sb 20 16 12 8 4 0 4 -8 11.5 db | 2Go) -98 | -102 | -110 | -118 | -130 | -146 | -156 | -168 | -180 | deg \ | The locus of G(ja) drawn on the Nichols chart cuts the M-contour and N-contour at various points. The meeting points ‘of G(ja) locus and various M-contours are noted. The phase a corresponding to the meeting point are noted from N-contours passing through the meeting point. (if a meeting pointlies between two N-contours, then choose an approximate value of a). ‘The frequency corresponding to the meeting pointare noted from bode plot by transferring the |G(j@)| corresponding to meeting point to bade plot. The values of o, Manda are listed in table-3. - ‘The values of M and « are the magnitude and phase of closed loop frequency response of Gfjo) with unity feedback. TABLE-3: Values of M and a from Nichols Chart o 7 ] rad/sec [0.36 |0.62 | 1.0 | 12 | 16 | 18 | 20 | 21 | 23 | 24 | 25 | 28 | 30 m -los]a[2{3]4/3 211 ]o |-2/-3|-6 |-e db i i | a 2 | 21 | 35 | 50 | 78 | 102 | 120 | 430 | 140 | 151 | 155 | 165 175 | 1 deg | j ! i { a Using the values listed in table-3, the closed loop frequency response piots are sketched as shown in fig 3.15.3. The closed loop frequency response consists of two plots and they are magnitude piot, MVS. and phase plot, a Vs.«. From the closed loop frequency response the values of M,, 0, and @, are noted. Resonant peak,M, = +4 db Resonant Frequency o, = 1.6 rad/sec Bandwidth, o, = 2.5 rad/sec 3.12 FREQUENCY RESPONSE ANALYSIS USING MATLAB In general, the open Joop or closed loop transfer function of a system is denoted as T(s). Let, T(s) be a rational function of'"s", as shown below. voy. DoS + bys“! + bys? +, 8)=— ye, Ags ays ays 24, For frequency response analysis, the transfer function T(s) is declared as a function of s using the following commands. s=tf('s"); Ts=(bO*SAM+b1*sA(M-1)+...+bM)/(aO*SAN+al*sA(N-1)+...+aN); The coefficients of numerator and denominator polynomials of the transfer function are determined using the following command. reece [num.cof den_cof]=tfdata(Ts); | alain ab Sl3.83 The gain margin, phase margin, gain crossover frequency ‘and phase crossover frequency can be determined using the following command. [GM PM wgc wpc] = margin(Ts); BODE PLOT In order to draw Bode plot the frequecy range can be specified using following commands. {w = logspace(ds, de, n); | jwhere, ds represents Start decade as 10“ de represents end decade as 10 nh represents number of points to be calculated between 10° & 10° Method 1: The Bode plot can be plotted using any one of the following command. Tee | bode(Ts,'k'); |bode(Ts, w); | bode(num_cof, den_cof); | bode(num_cof, den_cof,w); Method 2 : The Bode plot can also be plotted using semilog plot command as shown below. [Mag Phase w] = bode(Ts,w); MagdB = 290*10g10(mag); subplot (2,1,1);semilogx(w,MagdB, 'k"); | subpTot(2,1,2);semiTogx(w,Phase, "k"); In this method the magnitude and phase can be scaled by drawing two lines at two specified upper and lower values. For drawing these lines, one dimensional arrays consisting of same values has to be created by multiplying the specified value with one. The length of the array should be same as number of frequency points for which the magnitude and phase are computed. (Refer program 3.3). POLAR PLOT The polar plot can be plotted using the following commands. W = wistart : wistep : wend ; [re, im, w] = nyquist(num_cof,den_cof,w); z= re + i*im; r= abs(z); theta = angle(z); |polar(theta,r, 'w') NICHOLS PLOT The Nichols plot of open loop transfer function, G(s) can be plotted using the following commands. [Tnumcof den_cof ] =. tfdata(s); nichols(Gs); axis([ph_start, ph_end, mag_start, mag_end]);<< Ghiaphers “irequency Response Analysis) 3.84 PROGRAM 3.1 write a MATLAB program to draw the Bode plot for the open loop system governed by the following transfer function. G(s)=s?/(1+0.2s)(1+0.02s) program to plot Bode plot clear all clc setf('s"); z disp(‘the given transfer function is,’); Gs=(SA2)/((14+0.2*5)*(14+0.02*5)) w=logspace(-1,2,200); %Zspecify the frequency range bode(Gs,w) grid OUTPUT The given transfer function 7s, Transfer function: 0.004 sA2 + 0.22 541 Bode Diagram ‘Magnituda (4B) Phase (clog) Frequency (ranisec} Fig P3.1 : Bode plot of the open loop system given in problem 3.1. The Bode plot of program 3.1° is -shown in Fig p3.1. PROGRAM 3.2 Write a MATLAB program to draw the Bode plot and to calculate gain margin, phase margin, gain crossover frequency & phase crossover frequency for the open loop system governed by the following transfer function. G(s)=10/(0.04s340.Ss?+s)3.85 Chonipal ysiomsEngineering ~) OOOO %program to find gain & phase margins using bode plot clear all ele s=tfC's’); disp(‘the given transfer function is,’): Gs=10/((0.04*543)+(0.5*5A2)+5) bode(Gs,’k’) grid (GM, PM,wgc,wpc]=margin(Gs); GMdB=20*10g10(GM); disp(‘Gain margin in dB,’); GMdB disp(‘Phase margin in deg,’);PmM disp(‘Gain cross over frequency in rad/sec,');wgc disp(‘Phase cross over frequency in rad/sec,”);wpc OUTPUT The given transfer function 7s, Transfer function: Gain margin in dB, Phase margin in deg, PM 5.2057 Gain cross over frequency in rad/sec, Wwgc = Magnitude (o1B) Phase (deg) 10° fa io! ae 10 Frequency {redisex) Fig P3.2 : Bode plot of the open loop system given in problem 3.2.3. 86 Phase cross over frequency in rad/sec, . WwpC = 4.4629 The Bode plot of program 3.2 is shown in fig p3.2. PROGRAM 3.3 write a MATLAB program to draw the Bode plot for the open loop system governed by the following transfer function. The program should take care of drawing magnitude plot in the range +10 to -50 dB, and phase plot in the range -60 to -180 deg. G(s)=(75(1+0.2s))/(s(s?#16s+100)). “Bode plot with magnitude and phase scaling clear all clc s=tf("s'); disp(‘The given transfer function is,’); Gs=(75*(1+0.2*s))/(Cs* (sA2+16*5+100)) [num_cof den_cof]=tfdata(Gs); %determine numerator and denominator scoefficients of G(s) w=logspace(-1,2,200); [Mag, Phase,w]=bode(num_cof,den_cof,w); MagdB=20*10g10 (mag); Mscalel=10*ones(1,200); Mscale2=-50*ones(i, 200); subplot(2,1,1);semi logx(w,MagdB, ’-k’,w,Mscalel, ’k’,w,Mscale2,’k’) grid; xlabel(‘Frequency in rad/sec’); ylabel(‘Magnitude in dB’); Pscale 60*ones(1, 200); Pscale2=-180*ones(1,200); subplot(2,1,2);semilogx(w, Phase, ’k’,w,Pscalel,’-k’,w,Pscale2,’-k") grid; xlabel(‘Frequency in rad/sec'); ylabel('phase in deg’); Magnitude in d8 Frequency in rat/seo Phase in deg 107 1 10 1c Frequency in rad/seo Fig P3.3 : Bode plot of the open loop system given in problem 3.3.3.87 OUTPUT The given transfer function is, Transfer function: 543 + 16 sA2 + 100 5 The Bode plot of program 3.3 is shown in fig p3.3. PROGRAM. 3.4 consider the transfer function of open loop system given below. G(s)=20/(s?+7s7+10s) Write a MATLAB program to determine the transfer function of closed loop system with unity feedback, to plot the sede plot of closed loop system, and to calculate resonant peak, resonant frequency and bandwidth. %Bode-plot of unity feedback closed Toop system clear all cle s=tf(‘s’); disp(‘The given open. loop transfer function G(s) is,’); Gs=20/(sA34(7*sA2)4+10*5) disp(‘The closed Joop transfer function M(s) is,’); Ms=feedback(Gs,1) [num_cof den_cof]=tfdata(mMs); %determine numerator & denominator %coefficients of M(s) w=logspace(-1,1); %specify frequency range bode(Ms,w) grid [Mag, Phase ,w]=bode(Ms,w); [Peakmag,k]=max (Mag); disp(‘Resonant peak in dB,"); Mp=20*logl0(PeakMag) disp(‘Resonant frequency in rad/sec,’);wWr=w(k) n=1; while 20*log(mag(n))>=-3;n=n+1; end disp(‘sandwidth in rad/sec,’);wb=w(n) OUTPUT The given open loop transfer function G(S)- is, Transfer function: SAR + 7 5424+ 105 The closed Joop transfer function M(s) vis, Transfer function: GAZ #7 SA2 + 105 + 20 .Resonant peak in dB, Mp = 4.3953 Resonant frequency in. rad/sec, Wr= 1.6768 Bandwidth in rad/sec, Wb = 2.4421 The Bode plot of program 3.4 is shown in fig p3.4. Macyntuca (8) 3 & 8 Phase (dag) Frequency. {racisec) Fig P3.4 : Bode plot of the closed loop system of program 3.4. PROGRAM 3.5 write a MATLAB program to draw the polar plot and to calculate gain margin and phase margin for the open loop system governed by the following transfer function. G(s)=1/(s(1+s)*) %Program to draw polar plot and compute gain & phase. margins clear all clc s=tfC's"); dispC(‘the given transfer function is,’); Gs=1/(s*(1+s)*(1+s))} [num_cof den_cof]=tfdata(Gs); %determine numerator and . “denominator coeff. of G(s) w=0.4 : 0.01 : 4; “specify frequency’ range [re,im,wl=nyquist(num_cof ,den_cof,w); &determine the real. and Simaginary parts of G(jw) [em pM]=margin(num_cof, den_cof);: dispC‘Gain margin,’);GM disp(‘Phase margin in deg,’);PM %compute gain & phase margins3:89 , is the phase of G(ja). at the gain cross over frequency.Chapter Re: Ana; 3.94 03.10 Q3.11 03.12 03.13 03.14 03.15 03.16 03.17 Phase margin, y = 180° + oe 3 where, og, = ZG(jo) =a ge What is phase and Gain cross-over frequency? The gain cross over frequency is the frequency at which the magnitude of the open loop transfer function is unity. The phase cross over frequency is the frequency at which the phase of the open loop transfer function is 180°. Write the expression for resonant peak and resonant frequency. 1 fase Resonant peak, M, = ON eared : Resonant frequency, ©, = @,¥1— 2¢? 2641-6 - “Write a short note on the correlation between the time and frequency response? Correlation exists between time and frequency response of first or second order systems. The frequency domain specifications can be expressed in terms of the time domain parameters € and @,. For a peak overshoot in time domain there is a corresponding resonant peak in frequency domain. For higher order systems, there is no explicit correlation between time and frequency response. But if there is a pair of dominant complex conjugate poles, then the system can be approximated to second order system and the correlation between time and frequency response can be estimated. The damping ratio and natural frequency of oscillation of a second order systern is 0.5 and 8 rad/ sec respectively. Calculate the resonant peak and resonant frequency? Resonant peak, M. = i 26 ge Meee Vi=@ 2x(05)/1-(05)2 Resonant frequency, w, = 0, 41-20? =8x y1-2x 05* =5.657 rad/ sec What is Bode plot? The bode plot is a frequency response plot of the transfer function of a system. It consists of two plots : Magnitude plot and Phase plot. The magnitude plot is a graph between magnitude of a system transfer function in db and frequency, @. The phase plot is a graph between the phase or argument of a system transfer function in degrees and the frequency, o. Usually, both the plots are plotted on a common x-axis in which the frequencies are expressed in logarithmic scale. What is approximate bode plot? . In approximate bode plot, the magnitude plot of first and second order factors are approximated by {wo straight lines, which are asymptotes to exact plot. One straight line is at Odb, for the frequency range 0 to @, and the other straight line is drawn with a slope of + 20n db/dec for frequency range o, to «©. Here a, is the comer frequency. Define corner frequency? The magnitude plot can be approximated by asymptotic straight lines. The frequencies corresponding to the meeting point of asymptotes are called corner frequency. The slope of the magnitude plot changes at every corner frequency. What are the advantages of Bode Plot? 4. The magnitudes are expressed in db, and so, a simple procedure is available to add magnitude of each term one by one.3.95 Control Systems Engineering 2. The approximate bode plot can be quickly sketched, and the corrections can be made at comer frequencies to get the exact plot. 3. The frequency domain specifications can be easily determined. 4. The bode plot can be used to analyse both open loop and closed loop system. 03.18 What.is the value of error in the approximate magnitude plot of a first order factor at the corner Srequency? The error in the approximate magnitude plot of a first order factor at the corner frequency is + 3m db, where m is multiplicity factor. Positive error for numerator factor and negative error for denominator factor. 03.19 What is the value of error in the approximate magnitude plot of a quadratic factor with €=1 at the corner frequency? ‘The error is + 6db, for the quadratic factor with C=1. Positive error for numerator factor and negative error for denominator factor. K 03.20 Draw the bode plot of, G(s) = —. s Lets =ja, i K + G(ja) = —— (jo) The magnitude of G(ja) is unity when @ = K™*. The magnitude plot is a straight line with slope of -20n db/dec and passing through » = K"". The - Phase plot is straight fine parallel to x-axis at -90n°_ af I , O.1K™ KT 10KA o(log scale) Fig Q3.20 = Bode plot of integral factor, Kia)". Q3.21 Sketch the bode plot of G(s) = I/(1+sT). Let s = jo, - G{ja) = 1+joT 1 The corer frequency, 0, = T The magnitude plot is approximated by two straight lines : one straight line at Odb in the frequency range 0 to @, and the other straight Jine with the slope of -20db/dec in the frequency range «, to «. The phase of G(ja) varies from 0 to -90° as w is varied from 0 to «. Hence, the phase plotis a curve passing through -45° at the corner frequency.3.96 03.22 Q3.23 03.24 Q3.25 03.26 1 10 ° cry T (log scale) Fig 03.21 : Bode plot of the factor —_—. I+ jeT What is polar plot? The polar plot of a sinusoidal transfer function G(jw) is a plot of the magnitude of G(ja) versus the phase angle/argument of G@jo) on polar or rectangular co-ordinates as © is varied from zero to infinity. What is minimum phase system? The minimum phase systems are systems with minimum phase transfer functions. In minimum phase transfer functions, all poles and zeros will lie on the left half of s-plane. What is All-Pass systems? The all pass systems are systems with all pass transfer functions. In all pass transfer functions, the magnitude is unity at all frequencies and the transfer function will have anti-symmetric pole zero pattern (i.¢., for every pole in the left half s-plane, there is a zero in the mirror image position with respect to imaginary axis). What is non-minimum phase transfer function? A transfer function which has one or more zeros in the right half s-plane is known as non-minimum phase transfer function. In minimum phase system, how the start and end of polar plot are identified? For minimum phase transfer functions, with only poles, the type number of the system determines the quadrant in which the polar plot starts, and the order of a system determines the quadrant in which the polar plot ends. . Start of type-3 system t End of 3” order system End of 4° order system Start of type-2 —> “<— Start of type-0 aysiem Systenl End of 2“ End of 1* Start of type-I order system order system system "Fig Q3.26a : Start of polar plot of all pole Fig Q3.26b : End of polar plot of all pole minimum phase system minimum phase system.3.97 “Control Systems Engineeri 03.27 03.28 03.29 Q3.30 03.31 O#32 03.33 Q3.34 -270° Draw the polar plot of G(s)= (1487). 1 Lets =jo, .. G(jo) = -——— = —————___—- —180° 0° 14 joT Vt402T? Ztan oT e So © = Toone tantoT 90° 45 increasing Le Fig 03.27: Polar plot of G(s) = I(1+s). As @ >0, G(jo) > 120° . 0 Aso 30, G(jo) > 02-90° nese’ -270' 9 a Sketch th de bot. 6, CS) = SS =0 a etch the polar plot ef GO) = Sar iasE)l4sT) ° = + O° The given system is all pole minimum phase system. The type =180 3” > number of the system is 2 and the order is 5. Hence, the polar 1-90 plot starts in second quadrant and ends in fourth quadrant. Fig Q3.28 : Polar plot of type-2, Sth order system. What is Nichols plot? The Nichols plot is a frequency response plot of the open loop transfer function of a system. It is a graph between magnitude of G(ja) in db and the phase of G(ja) in degree, plotted on a ordinary graph sheet. What are M and N circles? The magnitude, M of closed loop transfer function with unity feedback will be in the form of circle in complex plane for each constant value of M. The family of these circles are called M-circles. Let N = tan a, where a is the phase of closed loop transfer function with unity feedback. For each constant value of N, a circle can be drawn in the complex plane. The family of these circles are called N-circles. How closed loop frequency response is determined from. open loop frequency response using M and N circles? The G(jw) locus or polar plot of open loop system is sketched on the standard M and N circles chart. The meeting point.of M circle with G(ja) locus gives the magnitude of closed loop system. (the frequency being same as that of open loop system). The meeting point of G(jo) locus with N-circle gives the value of phase of closed loop system, (frequency being same as that of open loop system). What is Nichols chart? The Nichols chart.consists of M and N contours superimposed on ordinary graph. Along each M-contour the magnitude of closed loop system, M will be a constant. Along each N-contour, the phase « of closed loop system will be constant. The ordinary graph consists of magnitude in db, marked on the y-axis and the phase in degrees marked on x-axis. The Nichols chart is used to find the closed loop frequency response from the open loop frequency response. How the closed loop frequency response is determined from the open loop frequency response using Nichols chart? The G(ja) locus or the Nichols plot is sketched on the standard Nichols chart. The meeting point of N-contour with G(jo) locus gives the magnitude of closed loop system and the meeting point with N-circle gives the argument/phase of the closed loop system. Wheat are the advantages of Nichols chart? 1. It is used to find closed loop frequency response from open loop frequency response. 2. The frequency domain specifications can be determined from Nichols chart. 3. The gain of the system can be adjusted to satisfy the given specification.Chapler 3 = Frequency Response Analysis 3.98 3.14 EXERCISES E3.1 E3.2 E3.3 E3.4 E3.5 E3.6 E3.7 E3.8 Sketch the Bode plot of the following open loop transfer functions and from the plot determine the Phase margin and gain margin. 4) G(s) = 100(1+0.1s)/s(I+0.2s) (1+0.5s) dD) G(s) = 846+ 10/4545}? (5+0.1) 5) G(s) = 50(1+0.1s)(1+0.01s) (14s) €) Gs) = 40 (145)A145s) (9242544) o) G(s) = 30(140.1s)4(1+0.01s) (1+5) DGG) = 10(1+5) €*Ys(1+0.2s) The open loop transfer function of a system is given by G(s) = K/s(1+0.5s) (1+0.2s). Using bode plot Jind the value of K so that (i) The gain margin of the system is 6db and (ii) The phase margin of the system is 25° Sketch the polar plot of the following transfer functions and from the plot, determine the phase margin and gain margin. a G(s) = 10 (st L)Ast10P od &°s (s+) (st+35) Bb) G(s) = 200 (s +2)/s(s? +10s +100) d) Is (5 +4) (5 +8) The open loop transfer function of a system is given by G(s} = K/s (s? + s + 4). Using polar plot, determine the value of K, so that phase margin is 50°. What is the corresponding value of gain margin? A unity feedback system has G(s) = K/s(1+0.1s). Using Nichols chart find the value of K so that resonant peak, M=1.4. Find the corresponding valie of @, The open loop transfer function of unity feedback system is, G(s) = K/(1+0.05s) (1+0.1s)(1+0.35). Using Nichols chart find the value of K so that gain margin of the system is 10db. What is. the corresponding value of phase margin. Using Nichols chart determine the closed loop frequency response of the unity feedback system, whose open loop transfer function is, G(s) = 200 (s+1)/s(s+10)°. A unity feedback system has open loop transfer function G(s) = 54/(1+0.1s) (s? + 88 + 25). Using Nichols chart determine the closed loop frequency response. From the closed loop response determine, the resonant peak, resonant frequency and bandwidth.CHAPTER 4 CONCEPTS OF STABILITY AND ROOT LOCUS 41 IMPULSE RESPONSE AND STABILITY DEFINITIONS OF STABILITY The term stability refers to the stable working condition of a control system. Every working system is designed to be stable. In a stable system, the response or output is predictable, finite and stable for a given input (or for any changes in input or for any changes in system parameters), The different definitions of the stability are the following 1. A system is stable, if its output is bounded (finite) for any bounded (finite) input. 2. A system is asymptotically stable, if in the absence of the input, the output tends towards zero , (or to the equilibrium state) irrespective of initial conditions. 3. A system is stable if for a bounded disturbing input signal the output vanishes ultimately as t approaches infinity. 4, A system is unstable if for a bounded disturbing input signal th the output is of infinite amplitude or oscillatory. 5. Fora bounded input signal, if the output has constant amplitude oscillations then the system may be stable or unstable under some limited constraints. Such a system is called Jimitedly stable, 6. Lfasystem outputis stable for all variations of its parameters, then the system is called absolutely Stable system. 7. Ifa system output is stable for a limited range of variations of its parameters, then the system is called conditionally stable system. IMPULSE RESPONSE OF A'SYSTEM Let, M(s) = Closed loop transfer function of a system. C(s) = Gutput / Response in s-domain. R(s) = Input in s-domain Cs) R(s) “. Response or Output in s-domain, C(s) = M(s) R(s) Now, M(s) = Now, Response in time domain, ¢(t) = £"{C(s)} Input in time domain, r(t) =L°{R(s)} For an impulse input, r(t) = 8(t) ; -. R(s) = £[S(t)]=1i 42 ~, Impulse response =L£-{C(s)} = £°'{M(s) R(s)} = £°{M(s)} = m(t) svoe(4.1) Hence, impulse response of a system is the inverse Laplace transform of system transfer function. The importance of impulse response is that, the output of a system for any arbitrary input can be obtained by convolution of input and impulse response. ie.; Response, c(t)=m(t) « r(t) where * is the symbol for convolution. Mathematically the convolution operation is defined as, e(t)= fre ett (4.2) where t is the dummy variable used for integration. BOUNDED - INPUT BOUNDED - OUTPUT (BIBO) STABILITY A linear relaxed system is said to have BIBO stability if every bounded (finite) input results in a bounded (finite) output. A condition for BIBO stability can be obtained from convolution operation defined by equation (4.2). For a relaxed system the equation (4.2) can-be written as, Response, c(t) =f m(t) (t=) de (43) 0 Note : A relaxed system is one in which the initial conditions are zero. Hence the limits of integration is from 0 to «. If the input r(t) is bounded then there exists a constant A,, such that {r(t)| < A,< 0. The condition for bounded output for this bounded input condition can be derived as follows. On taking the absolute value on both sides of equation (4.3), we get, la J mo rt-n ar weal. A) lo Since the absolute value of an integral is not greater than the integral of the absolute value of the integrand the equation (4.4) can be written as, . let) < J fm(x) t-a] de => fot] < j |m(x)| [nt-Tid] => lett] < J |m(t)) Aydt 0 0 0 For bounded input, a constant [oto] = Arf ime) ae exists such that, |r(t—1) < A;. 9 If the output c(t) is bounded then there exists a constant A, such that |c(t)| < A, < 0. no Ay free acSApe el (45) 043 The above condition is satisfied if, Ji dt a o Root on negative real axis m4 m(t)= ofA} =A A sta t Impulse response is exponentially decaying. Stable system. M(s) ss jOa s-a Root on positive real axis -1{_A at m(t) M(s)=L£ tot = Ae’ t Impulse response is exponentially increasing. Unstable system. +jb |}-———« Complex conjugate roots on right half of s-plane A A’ : m(t) = £7; ———— + Mo) =—*—+—4 — fw m(t ) ok — statjb sta—jb J ' op ane = Ae GTO), Aten(aribit | 2Ae-* cosbt = 2Ae™ sin(bt +90") vat ze -jb x. Complex conjugate roots Impulse response is damped sinusoidal on left half of s-plane (i.e., Damped oscillatory). Stable system g A Ee s-at+jb s—a-—jb M(s)= —* ~A_ = Ae CHB, ate Crib s-a+jb | s—a—jb 2 Baas enc anal aa race m(t) =2Ae™ cosbt = sin (bt +90 ) Impulse response is exponentially increasing sinusoidal (ie., Amplitude of oscillations exponentially increases with time). Unstable system.