ASSI Statistics PDF
ASSI Statistics PDF
ASSI Statistics PDF
Solution
1. Calculate the means (or expected value) E(X), E(W) and the variances V(X), V(W)
We have the probability density function of Gaussian and Weibull’s distribution
1 (𝑥 − 𝜇)2
𝑓𝑁 (𝑥) = 𝑒𝑥𝑝 (− ) −∞<𝑥 <∞
√2𝜋𝜎 2 2𝜎 2
𝑚 𝑥−𝑥0 𝑚−1 𝑥−𝑥0 𝑚
𝑓𝑊 (𝑥) = ( ) 𝑒𝑥𝑝[− ( ) ], 𝑥 ≥ 𝑥0
𝛼 𝛼 𝛼
𝑤ℎ𝑒𝑟𝑒 − ∞ < 𝜇 < ∞, 𝜎 > 0, 𝛼 > 0, 𝑚 > 0 𝑎𝑛𝑑 𝑥0 ≥ 0.
∞
𝐸(𝑋) = ∫ 𝑥𝑓𝑁 (𝑥)𝑑𝑥
−∞
∞ 1 (𝑥−𝜇)2
= ∫−∞ 𝑥 𝑒𝑥𝑝 (− ) 𝑑𝑥
√2𝜋𝜎 2 2𝜎 2
∞ 1 (𝑥−𝜇)2
= ∫−∞[(𝑥 − 𝜇) + 𝜇] 𝑒𝑥𝑝 (− ) 𝑑𝑥
√2𝜋𝜎 2 2𝜎 2
∞ 1 (𝑥−𝜇)2 ∞ 1 (𝑥−𝜇)2
= ∫−∞(𝑥 − 𝜇) 𝑒𝑥𝑝 (− ) 𝑑𝑥 + 𝜇 ∫−∞ 𝑒𝑥𝑝 (− ) 𝑑𝑥
√2𝜋𝜎 2 2𝜎 2 √2𝜋𝜎 2 2𝜎 2
We now see that the right integral is the complete integral over a normal pmf qe get
∞
1 (𝑥−𝜇)2
E(X) = (−𝜎 2 ) [𝑒𝑥𝑝 (− )] +𝜇
√2𝜋𝜎 2 2𝜎 2 −∞
1
= (−𝜎 2 )(0 − 0) + 𝜇
√2𝜋𝜎 2
=𝜇
2 ∞
V(X) = 𝐸(𝑋 − 𝐸(𝑋)) = ∫−∞(𝑥 − 𝜇)2 𝑓𝑁 (𝑥) 𝑑𝑥
∞ 1 (𝑥−𝜇)2 𝑥−𝜇
= ∫−∞(𝑥 − 𝜇)2 𝑒𝑥𝑝 (− ) 𝑑𝑥 by letting 𝑡 =
√2𝜋𝜎 2 2𝜎 2 𝜎√2
∞ 1 2 ∞ 2
Then𝑉(𝑋) = ∫−∞ 𝜎 2 2𝑡 2 𝑒 −𝑡 𝜎√2𝑑𝑡 = 2𝜎 2 /√𝜋 ∫−∞ 𝑡. 𝑡𝑒 −𝑡 𝑑𝑡
√2𝜋𝜎 2
2
We now use integration by parts with 𝑢 = 𝑡 𝑎𝑛𝑑 𝑑𝑣 = 𝑡𝑒 −𝑡 𝑑𝑡 it follow that,
−1 −𝑡 2 ∞ ∞
−1 −𝑡 2
𝑉(𝑋) = 2𝜎 2 /√𝜋 ([𝑡 𝑒 ] −∫ 𝑒 𝑑𝑡
2 −∞ −∞ 2
∞
2
1 −𝑡 2
𝑉(𝑋) = 2𝜎 /√𝜋 ∫ 𝑒 𝑑𝑡
−∞ 2
2018-2019
Institute of Technology of Cambodia Statistics
Therefore, mean and variance of Gaussian’s distribution are 𝐸(𝑋) = 𝜇 𝑎𝑛𝑑 𝑉(𝑋) = 𝜎 2 .
For Weibull’s distribution
∞ ∞ 𝑚 𝑥−𝑥0 𝑚−1 𝑥−𝑥 𝑚
E(X)= ∫−∞ 𝑥𝑓𝑊 (𝑥)𝑑𝑥 = ∫𝑥 𝑥 ( ) 𝑒𝑥𝑝[− ( 0 ) ] 𝑑𝑥
0 𝛼 𝛼 𝛼
𝑥−𝑥0 𝑚 1
𝛼 1
Let 𝑡 = ( ) => 𝑥 = 𝛼𝑡 𝑚 + 𝑥0 𝑠𝑜 𝑑𝑥 = 𝑡 𝑚−1 𝑑𝑡 𝑤ℎ𝑒𝑛 𝑥 → 𝑥0 𝑡ℎ𝑒𝑛 𝑡 → 0 𝑎𝑛𝑑 𝑥 → ∞, 𝑡 → ∞
𝛼 𝑚
1
∞ 𝑚 1− 1 𝛼 1 −1 −𝑡
Hence E(X)= ∫0 (𝛼𝑡 𝑚 + 𝑥0 ) 𝑡 𝑚 𝑡𝑚 𝑒 𝑑𝑡
𝛼 𝑚
1
∞ ∞
= ∫0 𝛼𝑡 𝑚 𝑒 −𝑡 𝑑𝑡 + 𝑥0 ∫0 𝑒 −𝑡 𝑑𝑡
1 ∞
= 𝛼𝛤 ( + 1) + 𝑥0 (𝐺𝑎𝑚𝑚𝑎 𝐹𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝛤(𝛼) = ∫0 𝑡 𝛼−1 𝑒 −𝑡 𝑑𝑡)
𝑚
𝑥−𝑥0 𝑚 1
𝛼 1
Let 𝑡 = ( ) => 𝑥 = 𝛼𝑡 𝑚 + 𝑥0 𝑠𝑜 𝑑𝑥 = 𝑡 𝑚−1 𝑑𝑡 𝑤ℎ𝑒𝑛 𝑥 → 𝑥0 𝑡ℎ𝑒𝑛 𝑡 → 0 𝑎𝑛𝑑 𝑥 → ∞, 𝑡 → ∞
𝛼 𝑚
1 2
∞ 𝑚 1− 1 𝛼 1 −1 −𝑡
So, 𝐸(𝑊 2 ) = ∫0 (𝛼𝑡 𝑚 + 𝑥0 ) 𝑡 𝑚 𝑡𝑚 𝑒 𝑑𝑡
𝛼 𝑚
2 1
∞ ∞ ∞
= ∫0 𝛼 2 𝑡 𝑚 𝑒 −𝑡 𝑑𝑡 + ∫0 2𝛼𝑥0 𝑡 𝑚 𝑒 −𝑡 𝑑𝑡 + ∫0 𝑥02 𝑒 −𝑡 𝑑𝑡
2 1
= 𝛼 2 𝛤 ( + 1) + 2𝛼𝑥0 𝛤 ( + 1) + 𝑥02
𝑚 𝑚
2 1 1 2 1
= 𝛼 2 𝛤 ( + 1) + 2𝛼𝑥0 𝛤 ( + 1) + 𝑥02 − 𝛼 2 [𝛤 ( + 1)] − 2𝛼𝑥0 𝛤 ( + 1) − 𝑥02
𝑚 𝑚 𝑚 𝑚
2 1 2
=𝛼 2 (𝛤 ( + 1) − [𝛤 ( + 1)] ).
𝑚 𝑚
𝑍−𝜇𝐽 0−𝜇𝐽
Since 𝑅𝐽 = 𝑃(𝐽 > 0) = 1 − 𝑃(𝐽 < 0) = 1 − 𝑃( < )
𝜎𝐽 𝜎𝐽
2018-2019
Institute of Technology of Cambodia Statistics
𝜇
= 1 − 𝛷(− 𝐽 )
𝜎𝐽
𝜇 𝜇 𝜇𝐽
= 1 − (1 − 𝛷( 𝐽 ) = 𝛷( 𝐽 ) = 𝛷(𝛽) where 𝛽 =
𝜎𝐽 𝜎𝐽 𝛿𝐽
𝜇𝐽
Therefore the reliability 𝑅𝐽 = 𝛷(𝛽) where 𝛽 = is the reliability index (or safety index)
𝛿𝐽
Since 𝑃𝑓 = 𝛷(−𝛽)
2018-2019
Institute of Technology of Cambodia Statistics
Reliability index for a structure element
20𝑘𝑁
We have, 𝑞̅ = , 𝑉𝑞 = 8%, 𝑄̅ = 60𝑘𝑁, 𝑉𝑞 = 12%, ̅̅̅̅
𝑀𝑟 = 290𝑘𝑁. 𝑚, 𝑉𝑀𝑟 = 10%
𝑚
The standard deviation is equal to the product of the mean value and coefficient of variation, that is,
Limit state function is defined by 𝐽 = 𝑆 − 𝐿 know that, the structural strength 𝑆 = 𝑀𝑟 and the load effect L is
equal to the sum of maximum at exposed distribution load q and the concentrated load Q, that is
𝑙2 𝑙
𝐿 = ( ) 𝑞 + ( )Q.
8 4
The mean value and load variation on the structure (in this case the bending moment)
𝑙 2 𝑙
𝐿̅ = ( ) 𝑞̅ + ( )𝑄̅
8 4
52 5
= ( ) 20 + ( ) 60
8 4
= 137.5 .
𝑙2 𝑙
And the mean value 𝜌𝐿2 = ( ) 𝜎𝑞2 + ( )𝜎𝑄2
8 4
52 5
= ( ) 2.56 + ( ) 51.84
8 4
= 106 .
The mean value and the variation of the structure capacity (in this case the bending capacity)
𝑆̅−𝐿̅
𝑆̅ = ̅̅̅̅
𝑀𝑟 = 290, 𝜎𝑆2 = 𝜎𝑀
2
𝑟
= 841. The reliability index is equal to 𝛽 =
√𝜎𝑆2 +𝜎𝐿2
290−137.5
Then 𝛽 = = 4.95. And the probability of failure 𝑃𝑓 = 𝛷(−𝛽) = 𝛷(−4.95) = 3.12 × 10−7 .
√841+106
Therefore the reliability index 𝛽 = 4.95 and the probability of failure 𝑃𝑓 =3.12 × 10−7 .
𝑍̅ = 0.8 × 10−3 𝑚3 , 𝑉𝑍 = 7%
̅̅̅̅
𝑀𝑢 = 140𝑘𝑁. 𝑚, 𝑉𝑀𝑢 = 10%
2018-2019
Institute of Technology of Cambodia Statistics
Let 𝐽 = 𝑀𝑝 − 𝑀𝑢 is the state function
= 𝑃(𝑀𝑝 − 𝑀𝑢 < 0)
𝑀𝑝
= 𝑃( < 1)
𝑀𝑢
𝑓𝑦 𝑍
Since 𝑀𝑝 = 𝑓𝑦 𝑍 then 𝑃𝑓 = 𝑃( < 1) . Variances of 𝑓𝑦 , 𝑍 𝑎𝑛𝑑 𝑊𝑢 are given respectively by
𝑀𝑢
Since, 𝑓𝑦 , 𝑍 𝑎𝑛𝑑 𝑀𝑢 𝑎𝑟𝑒 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 𝑙𝑜𝑔𝑎𝑟𝑖𝑡ℎ𝑚𝑖𝑐𝑎𝑙𝑙𝑦 𝑛𝑜𝑟𝑚𝑎𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 𝑠o,
variances and means for 𝑌𝑓 , 𝑌𝑍 𝑎𝑛𝑑 𝑌𝑀 𝑎𝑟𝑒
= 𝑌𝑓 + 𝑌𝑍 − 𝑌𝑀
= 21.22 × 10−3
𝑓𝑦 𝑍 (𝑙𝑛1−𝑌̅) 0−0.47
Then the probability of failure is 𝑃𝑓 = 𝑃 ( < 1) = 𝛷 [ ] = 𝛷 [( )] = 𝛷(−3.08)
𝑀𝑢 𝜎𝑌 √21.22×10−3
= 1.24 × 10−3 .
2018-2019
Institute of Technology of Cambodia Statistics
Therefore,
The reliability index 𝛽 = 3.08 and the probability of failure 𝑃𝑓 = 1.24 × 10−3 .
2018-2019