Professional Documents
Culture Documents
Solutions To Exercise 5: Problem 4.10
Solutions To Exercise 5: Problem 4.10
Marvin Rausand
Email: marvin.rausand@ntnu.no
2008-09-30
Problem 4.10
(a) Let X B be the repair time for component B and X C be the repair time for
component C with E (X B ) = τB and E (X C ) = τC , respectively. Also, let X
be the repair time of an arbitrary failure. Since it is assumed that only one
component can be in the failed state at any one time, the expected repair
time can be expressed as:
E (X ) = E (X B ) · P (B failed) + E (X C ) · P (C failed)
λB
Pr(B failed) =
λB + λC
λC
Pr(C failed) =
λB + λC
Inserting this in the expression for the expected repair time, yields:
λB τB + λC τC
MTTR = E (X ) =
λB + λC
1
MTTF =
λB + λC
MTTF
A av =
MTTF + MTTR
1
λB +λC
= λ τ +λ τ
1
λB +λC
+ B λB +λC C
B C
1
= = 0.96852
1 + λB τB + λC τC
1
(c) Assuming independence between B and C , the average availabilities for
components B and C become, respectively:
MTTFB
A I ,B = = 0.99751
MTTFB + MTTRB
MTTFC
A I ,C = = 0.97087
MTTFC + MTTRC
This gives:
A I = A I ,B · A I ,C = 0.96845 < A av
Problem 4.11
A series structure has 10 independent and identical components. Each compo-
nent has MTTF = 5000 hours.
(a) When the components have constant failure rate λ, MTTF = 1/λ, and the
series structure has constant failure rate λS = 10 · λ. The mean time to
failure of the system is:
1 1 1 MTTF
MTTF = = · = = 500 hours
λS 10 λ 10
(b) The life length T of the components is now assumed to be Weibull dis-
tributed with shape parameter α = 2. The MTTF is
1 1
MTTF = Γ +1
λw α
1 1
= Γ(1.50) = · 0.88623 = 5000 hours
λw λw
Solving for λw we get
That is, the system has a Weibull distribution with scale parameter λS =
10·λw ≈ 5.61·10−4 and shape parameter α = 2. The MTTFS of the system
is therefore
1 1
MTTFS = Γ +1
λS α
1
= · 0.88623 = 1581 hours
λS
2
(c) Now consider a parallel structure of 2 independent and identical compo-
nents with MTTF = 5000 hours.
Problem 4.12
(a) The mean time to failure for the vessel is:
1
MTTF = = 4 950 hours
λ
(b) The probability that the vessel does not fail within 4 months can be calcu-
lated by using the survivor function (1 month = 730 hours):
−4
R(4months) = e −2.02·10 ·730·4
= 0.554
(c) The failure rates when the capacities are 40%, 80% and 100% are:
3
0,0008
0,0006
Failure rate
0,0004
0,0002
0,0000
0,0 0,2 0,4 0,6 0,8 1,0
Determining the probability that the burner system survives 2 months can
be found by considering it as a 3-out-of-4 system. Let X (t ) be the num-
ber of burners functioning at the time t . X (t ) will then be binomially dis-
tributed. Dropping the index t we get:
The probability that the system survives 2 months and does not have any
burner element failures is:
The probability that enough burners start and survive 2 months is:
4
The vessel itself has to survive at 80% capacity, which we know from part
c) gives λ0.8 = 2.02 · 10−4 . This gives the following probability for survival
of the vessel itself for 2 months:
−4
Pr(Vessel itself survives) = e −2.02·10 ·24·30·2
= 0.748
(f) The active vessel already has 2 burners active. It therefore needs to turn on
at least one more burner in order to operate at 80% by itself. The solution
to this problem becomes: