CSE39300 Analytical and Quantitative Methods for Civil Engineers
CSE39300 Analytical and Quantitative Methods for Civil Engineers
Mini Project
Lecturer: Ir. CL Kwan
Students Name: Cheung King Hin (16000458D)
Tang Chung Yue (15054117D)
1. The 28-day compressive strength in MPa of 60 concrete cubes (n=60) are tabulated
below:
1
CSE39300 Analytical and Quantitative Methods for Civil Engineers
(a) Determine the 95% two-sided confidence interval for the population mean.
n 60
∑ f ci ∑ f ci
x= i=1 = i =1 =37.74
n 60
( (∑ ) ) ( (∑ ) )
n 2 60 2
n
f ci 60
f ci
1 1
∑ f ci − ∑ f ci −
2 i=1 2 i=1
s= = =4.54
n−1 i=1 n 60−1 i =1 60
Since the population variance σ2 is unknown,
It is necessary to use t-distribution to construct the 95% two-sided confidence interval.
α =1−0.95=0.05
tα =t 0.025 ,59=2.0 0
,n−1
2
(
⟨ μ ⟩1−α = x−t α
, n−1
s
; x +t α
2 √n
, n−1
s
2 √n )
(
⟨ μ ⟩0.95= 37.74−2 4.54 ; 37.74+2 4.54 =( 36.57 ; 38.91 )
√ 60 √ 60 )
∴the 95% two-sided confidence interval for the population mean is (36.57 , 38.91).
(b) A normal or lognormal probability density function may be a plausible distribution to
model the compressive strength of concrete cubes. Estimate the parameters of the normal
and lognormal distributions by the method of moments respectively.
For normal distribution:
X : N (μ , σ )
μ=x =37.74
⇒
σ 2=s 2=20.58 σ =4.54
For lognormal distribution:
X : ln(λ , ζ )
( ) ( )
2 ⇒
2 σ 20.58 1 2 1
ζ =ln 1+ 2 =ln 1+ =0.014 ζ=0.1 2λ=ln μ− ζ =ln37.74− × 0.014=3.62
μ 37.74
2
2 2
2
CSE39300 Analytical and Quantitative Methods for Civil Engineers
(c) Plot the data on both the normal and lognormal probability papers, and construct the
least-squares regression line through the data points and determine the coefficient of
determination if a linear trend can be observed from each probability plot. From a visual
inspection of the plotted data points, which the underlying distribution is a better model
to represent the probability distribution of the compressive strength of concrete cubes.
Estimate the parameters of the normal and lognormal distributions respectively.
For normal probability paper:
Normal Probability Paper
50
f(x) = 4.75069753775842 x + 37.7416666666667
R² = 0.988648058655944 45
40
35
30
fc (MPa)
25
20
15
10
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
The least-squared regression line: f c =4.7507 s+37.742
The coefficient of determination: r 2=0.9886
∴ μ=37.742∧σ=4.7507
3
CSE39300 Analytical and Quantitative Methods for Civil Engineers
For lognormal probability paper:
Lognormal Probability Paper
4
3.9
f(x) = 0.12822981156936 x + 3.62344225736572
R² = 0.977296802087956
3.8
3.7
3.6
ln fc (MPa)
3.5
3.4
3.3
3.2
3.1
3
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
The least-squared regression line: ln f c =0.1282 s +3.6234
The coefficient of determination: r 2=0.9773
∴ λ=3.6234∧ζ=0.1282 2 gets closer to 1, the data set will result in stronger linear
relationship. Therefore, Normal distribution is a better model to represent the probability
distribution of the compressive strength of concrete cubes, since the value is closer to 1.
4
CSE39300 Analytical and Quantitative Methods for Civil Engineers
(d) Using the result in part (b), perform chi-square goodness-of-fit tests for both the normal
and lognormal distributions at the 5% significance level. Determine which of the two
distributions is more appropriate to model the compressive strength of concrete cubes.
[k=6 intervals of <32, 32-34.5, 34.5-37, 37-39.5, 39.5-42 & >42 can be used]
Result from part (b), μ=37.74 , σ=4.54 , λ=3.62 , ζ=0.12
Interval Observed ei 2
(n i−e i)
(MPa) Frequency, ni ei
normal lognormal normal lognormal
<32 7 6.1692 5.6251 0.1119 0.3360
32-34.5 7 8.0771 9.0820 0.1436 0.4773
34.5-37 14 11.8577 12.7646 0.3870 0.1196
37-39.5 12 12.9464 12.7320 0.0692 0.0421
39.5-42 11 10.5125 9.5706 0.0226 0.2135
>42 9 10.4370 10.2257 0.1979 0.1469
Σ 60 60 60 0.9322 1.3354
Z N =0.9322 ; Z ln =1.3354 ; f =k−1−m=6−1−2=3 ;C 0.95 ,3 =7.815
∵ Z N ∧Z ln ≤C 0.95 ,3
∴Both models substantiated
But normal distributions is more appropriate to model the compressive strength of
concrete cubes.
(e) Using the result in parts (b) and (d), determine the probability that the compressive
strength of a concrete cube is below the characteristic strength of 30MPa from the
appropriate distribution.
Let X is the compressive strength of a concrete cube.
P ( X <30 )=P Z< ( 30−37.74
4.54 )
=P ( Z ←1.705 )=0.044
