Type Testing a 2000 MW Turbogenerator

K. Sedlazeck S. Lindholm
C. Richter J. Pipkin
S. Strack F. Fu
B. Humphries
L. Montgomery
Siemens Energy Siemens Energy
Mülheim, Germany Orlando, FL USA

Abstract— Type tests completed in August 2008 have validated all I. INTRODUCTION
electrical and mechanical design parameters of the 1992 MVA 4-
pole, 1500 RPM turbogenerator that Teollisuuden Voima Oy is The first EPR turbogenerator with its brushless exciter is
now installing in the Olkiluoto 3 nuclear power station. Olkiluoto shown fully assembled on the test field in Figure 1 at the
3 is the first European Pressurized Reactor (EPR) power station. conclusion of a 3 week type test that was completed in August
In response to TVO’s requirement that the generator capability 2008. A few days after this photograph was taken, the
be at least 10% above the 1992 MVA nameplate rating, Siemens generator was disassembled and shipped to the Olkiluoto 3
Energy designed the generator so that its capability is above 2222 power station, where it is now being installed on the turbine
MVA (2000 MW, 0.90 power factor). Besides being designed to deck.
have this high capability, the generator is also designed to be very
efficient (tested generator efficiency is nearly 99% at the 1992
MVA nameplate rating), and it is designed to exhibit very low
In this paper the authors outline design features that have
mechanical vibration levels. In this paper the authors report the been previously described in more detail in References 2 and 3,
following highlights of the electrical and mechanical and we also describe the main electrical and mechanical
characteristics that were observed during recent type tests: characteristics that were observed during the recent type test.
• Confirmation of expected steady state MVA capability
• Confirmation of expected dynamic stability parameters
Successfully completing the type test of this 2000 MW
turbogenerator marks a major milestone in a journey that began
• Confirmation of expected acceptable mechanical vibration
nearly 40 years ago with the initiation of nuclear
Index Terms— turbogenerator, nuclear power station, turbogenerator designs in this rating class. Some discussions
hydrogen cooling, water cooling, type testing, MVA capability, of the work of those earlier days are recorded in Reference 1.
efficiency, vibration, stability parameters

Figure 1 - Olkiluoto 3 turbogenerator with its brushless exciter at the conclusion of the type test

978-1-4244-4252-2/09/$25.00 ©2009 IEEE 465

 II. MAIN PERFORMANCE ATTRIBUTES
Main performance attributes are summarized in Table 1
below.
TABLE 1 - ATTRIBUTES OF EPR TURBOGENERATOR

Applied Standard IEC 60034

Class of Insulation Class 155, with Class 130
temperatures
Apparent Power 2222 MVA
Power Factor 0.90
Frequency and Speed 50 Hz, 25 sec-1
Rated Voltage 27 kV
Rated Current 47.5 kA (50 kA at 95% voltage)
Excitation Brushless Exciter
Figure 3 – Axial flow hydrogen-cooled rotor winding and water-cooled
stator winding in EPR turbogenerator

III. MAIN DESIGN FEATURES

IV. TYPE TEST OVERVIEW
Evolutionary design: As discussed in References 2 and 3,
the 2222 MVA EPR turbogenerator design is based on the The 3 week type test included the following activities:
1500-1700 MVA Siemens “Konvoi” turbogenerators that have
• Mechanical runs at rated speed (1500 rpm)
operated with high reliability and availability for decades in
nuclear power stations (see Ref 1). To meet requirements Balance and vibration (in air)
specified for the EPR generator, well proven features from Ventilation test (in air)
other Siemens generators have been incorporated. Friction and windage losses (in hydrogen)

Generator Cooling: As illustrated in Figures 2 and 3, this • Steady-state short circuit tests
hydrogen-cooled generator is equipped with an all water- Short circuit heat runs with collector set
cooled stator winding. The rotor winding is axially hydrogen- Short circuit losses and saturation curve
cooled. The stator core is primarily axially hydrogen-cooled,
with radial ducts at each end for supplemental cooling. Short circuit heat runs with brushless exciter
Hydrogen is circulated with one multi-stage, axial flow blower • Steady state open circuit tests
located at the turbine end of the generator, between two Open circuit heat runs at 27 kV
vertical coolers.
Open circuit saturation curve and core losses
Voltage wave shape analysis
• Rotor moment of inertia (retardation test)
• Sudden short circuit tests (with brushless exciter)

Accomplishing this extensive test program in only 3 weeks

was facilitated by the fact that all activities were controlled and
monitored by experienced test field personnel working in a
state of the art control room overlooking the test field. As
shown in Figure 4, throughout the test program key members
of the design team joined the test field colleagues in this
control room to compare test results with pre-calculated values
for key parameters (e.g. temperatures, reactances, vibration
levels) as those parameters were being measured.

Figure 5 shows the generator with the test field collector

Figure 2 - General assembly of EPR turbogenerator set, and Figure 6 shows the generator with its brushless exciter.

466
 Figure 7 - Summary of test results for steady state thermal capability

Details of the winding temperatures are described in

Figure 4 - Control room of test field Figures 8 and 9. As shown in Figures 7 and 8, temperatures in
the rotor winding are well below IEC and IEEE Class 130
limits.

Figure 5 - EPR generator and test field collector set on the test field at the
beginning of the test program

Figure 8 - Temperatures in rotor winding of EPR turbogenerator

As shown in Figures 7 and 9, stator winding temperatures

are well below limits for water-cooled stator windings.

Figure 6 - EPR generator and its brushless exciter on the test field at the
end of the test program

V. TEST CONFIRMATION OF THERMAL CAPABILITY

As outlined in Figure 7, tested generator winding
temperatures were well below IEC 60034-1 and IEEE C50.13
Class 130 limits. As can also be seen in this figure, tested
temperatures closely matched calculated temperatures. Figure 9 - Stator winding temperatures in EPR turbogenerator

467
 As shown in Figures 7, 8 and 9, from the point of view of
IEC Class 130 thermal limits for the windings, the tested
capability exceeds the target maximum apparent power, 2222
MVA, by over 20%.

All additionally measured temperatures (e.g. core, frame,

bearing brackets, terminal box) were lower than calculated and
lower than any temperatures that would prove injurious to
generator components during operation of the generator at the
design target maximum apparent power, 2222 MVA.

One of the major design challenges to achieving this 2222

MVA rating is in the design of the terminals. As noted in
Figure 7, at this rating, the terminal current approaches 50 kA.
Indeed, the numbers recorded in Table 1 in Section II of this
Figure 11 – Dodecagon flexible connector assemblies between main leads
paper show that at the lower extreme of the ±5% voltage range and bushings
specified in IEC and IEEE standards the required terminal
current slightly exceeds 50 kA. Designing the terminals Because each assembly of 12 flexible connectors must
needed for so high a current was a major challenge in carry 50kA, and because cooling is by thermal conduction to
developing the EPR turbogenerator design. Confirming the adjacent water-cooled components, even distribution of current
proper functioning of these conductors was an important among the flexible connectors is important. Designing this
objective of the type test. Results from the short circuit factory assembly for even distribution of the connector currents around
tests at 40kA (refer to Figure 7) revealed that temperatures of the circle of conductors shown in Figure 11 included three
outlet water leaving the high voltage bushings would be less dimensional electromagnetic finite element analyses. As
than 60oC when the generator current reaches 50KA. This shown in Figures 10 and 11, the 6 connector assemblies are
temperature is well below the 90oC limit specified by IEC and much larger than a depth of penetration, and the 6 assemblies
IEEE standards. are close to each other. Consequent electromagnetic coupling
induces uneven current distribution. By working with 3D FEA
A very important step in validating the design of the 50 kA models (and by experimentally confirming these models by
terminals of the EPR generator was to verify the effectiveness favorable comparisons of calculations with tests on a smaller
of thermal conduction cooling of the braided copper flexible generator with similar flexible connector assemblies) the
conductors which are used to electrically connect the 6 water design team achieved a calculated ±5% variation among the 12
cooled main leads to the 6 water cooled bushings. These flexible conductors. As reported in Reference 4, observations
connector assemblies, shown in Figures 10 and 11, are used to made with Rogowski coils wrapped around each connector
facilitate connecting the bushings to the main leads during during the type tests showed good agreement between
assembly of the terminal box to the generator frame in the calculations and tests. Comparing calculated distributions of
power station. They also serve to isolate bushing vibration currents in the flexible connectors (the 12 white bars in Figure
from main lead vibration. Cooling of these flexible conductors 12) with results from two separate tests (the blue and red bars.)
is by thermal conduction to the water-cooled bushings and shows this good agreement and confirms the ±5% variation
main leads to which the flexible conductors are bolted. design goal.

Figure 10 - Terminals of EPR turbogenerator, showing 6 main leads, 6 Figure 12 - Comparison of tested and calculated current distributions in
flexible connector assemblies and 6 bushings flexible connector assemblies between main leads and bushings

468
 Validating calculated magnetic saturation and calculated VI. CONFIRMATION OF DYNAMIC CHARACTERISTICS
synchronous reactances was an important aspect of the test
As shown in Figures 15 and 16, all tested reactances were
program. As shown in Figures 13 and 14, the tested saturation
within the calculated min/max range, thus validating grid
curves closely matched the curves calculated by finite element
stability studies completed with reactances calculated during
analysis (FEA) of the generator electromagnetic performance.
the development of this generator design.

SGen5-4000W for Olkiluoto 3 SGen5-4000W Generator

Open Circuit Saturation Curve
Unsaturated Reactances
40
3.5

Reactances P.U. at 1992 MVA

35
3.0 calc unsat max
line to line voltage [kV]

30 test unsat
2.5
25 calc unsat min

20 2.0
Rated Voltage
15 Calculated by FEA 1.5
10 Test Points
1.0
5
0.5
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0
field current normalized by xd x’d x’’d
27 kV open circuit field current
Figure 13 - Calculated and tested open circuit saturation curves Figure 15 – Calculated and tested unsaturated reactances

SGen5-4000W for Olkiluoto 3 SGen5-4000W Generator

Short Circuit Saturation Curve Saturated Reactances
50
3.5
Reactances P.U. at 1992 MVA

45
stator phase current [kA]

40 3.0 calc sat max

35 test sat
2.5
30 calc sat min
25 2.0
20
1.5
15
Rated Current
10 Calculated 1.0
5 Test Points
Test Based If(rated current) 0.5
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0.0
field current normalized by
xd x’d x’’d
42.6 kA short circuit field current
Figure 16 – Calculated and tested saturated reactances
Figure 14 - Calculated and tested short circuit saturation curves

469
 VII. TEST CONFIRMATION OF ACCEPTABLE VIBRATION VIII. CONCLUSIONS
During the type test the stator was extensively instrumented Results from the August 2008 type test of the 2222 MVA,
with vibration sensors. These sensors included fiberoptic 2000 MW turbogenerator designed and manufactured for the
vibration sensors on the high voltage windings, piezoelectric first EPR nuclear power station showed excellent agreement
sensors on the core and frame and bearing brackets, and between design values and tested values of the generator
standard sensors to measure rotor shaft vibration. All sensors parameters. All results met or exceeded requirements set by
were monitored and recorded with computer data acquisition Siemens and by IEC standards.
systems that provided both real time output for immediate
inspection by the engineers monitoring the test and also orderly Successfully completing the type test of this 2000 MW
storage of the data for post processing. As shown in Figure 17, turbogenerator has marked a major milestone in a journey that
all tested stator and rotor vibrations were less than specified began over 40 years ago with the initiation of nuclear
limits. The Rigi-Flex stator winding vibration levels were turbogenerator designs in the rating class above 1200 MW.
quite low, less than 60% of allowable levels. Nearly all All people who conducted and supported this type test, all
vibration levels were less than 80% of specified limits. people who contributed to the previous years of development
of the generator design, and all people who contributed to
manufacturing the generator have good reason to be pleased
with the excellent results after this years-long and very difficult
work.

Figure 17 - Summary of vibration test results

In regard to the fact shown in Figure 17 that vibration of

the rotor shaft approached its permissible limit, two points are REFERENCES
important to note. First, the shaft vibration limit for the
Olkiluoto 3 generator rotor was specified to be very low, about [1] A. Abolins, H. Achenbach, D. Lambrecht, “Design and
75% of previously acceptable limits. Second, the bearing performance of large four-pole turbogenerators with semi-
bracket and rotor shaft vibration levels reflect the nature of the conductor excitation for nuclear power stations”, CIGRE
foundation support for the generator in the test field. As can Report 11-04, 1972 Session
be seen in Figures 1, 5 and 6, in the test field the generator was
supported on steel parallels that were nearly 2m tall. Because [2] K. Sedlazeck, W. Adelmann, H. Bailly, I. Gahbler, H.
that support is different from (and less stiff than) the support Harders, U.Kainka, U. Schuberth, H. Spiess, R. Chianese,
that is to be provided by the power station foundation, the P. Hugh Sam, R. Ward, L. Montgomery, “Influence of
generator rotor was balanced only to the extent needed to get customer’s specifications upon design features of the EPR
vibration levels to be acceptable for the type test. When the turbogenerator”, CIGRÉ Report 11-106, 2002 Session.
generator is installed in the power station and connected to the
steam turbines, rotor balance can be refined further and the [3] R. Gray, R. Nelson, L. Montgomery, J. Pipkin, S. Joki-
rotor shaft and bearing bracket vibrations can be reduced Korpella, F. Caguiat, “Designing the world’s most
further below the limits specified for Olkiluoto 3. powerful turbogenerator – Olkiluoto Unit 3”, Conference
Paper no. 0167, IEEE PES, Montreal, 2006
All generator vibration levels observed during the test were
below specified limits, and it is anticipated that generator [4] K. Sedlazeck, L. Montgomery, “Generators > 1200 MW”,
vibration levels in the power plant will be less than observed Giga-Watt Generation Panel Session, CIGRÉ Group A1
during the factory test. (Rotating Electrical Machines), 2008 Session

470