Scribd
Upload
6 views

Problem Solving Session 1_Probability and Reliability

The document outlines a problem-solving session focused on probability and reliability, covering three main problems: conditional probability for battery fault detection, system reliability analysis for a series system, and time-dependent system reliability analysis. It includes detailed calculations and methodologies for evaluating the reliability of systems under various conditions. The session emphasizes the importance of understanding failure modes and the implications of alarm systems in detecting faults.

Uploaded by

Amirhossein Shokrani
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
6 views

Problem Solving Session 1_Probability and Reliability

The document outlines a problem-solving session focused on probability and reliability, covering three main problems: conditional probability for battery fault detection, system reliability analysis for a series system, and time-dependent system reliability analysis. It includes detailed calculations and methodologies for evaluating the reliability of systems under various conditions. The session emphasizes the importance of understanding failure modes and the implications of alarm systems in detecting faults.

Uploaded by

Amirhossein Shokrani
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Problem-Solving Session 1: Probability and Reliability

Problem-Solving Session 1
Outline

• Problem 1: Conditional Probability for Detectability Analysis

• Problem 2: System Reliability Analysis for Serial System

• Problem 3: Time-Dependent System Reliability Analysis

2
Problem-Solving Session 1

Problem 1: Conditional Probability for Detectability Analysis


 The fault of a lithium-ion battery can be detected via the use of a
battery management system (BMS) that is capable of real-time
sensing and diagnostics.
 Field data indicate that 1) the probability of battery fault is
0.01%; 2) when the battery has a fault, the BMS activates an
alarm for 98% of the time; and 3) when the battery has no fault,
the BMS activates an alarm 1% of the time.
 Calculate the following: 1) the probability that an alarm is
activated by the BMS and 2) the probability that the battery
actually has a fault, given that an alarm is activated.

3
Problem-Solving Session 1

Problem 1: Conditional Probability for Detectability Analysis


 Let the events be
 A = BMS activating an alarm.
 FB = Faulty battery, P(FB) = 0.0001.
 HB = Healthy battery, P(HB) = 1 – 0.0001 = 0.9999.
 The probability of BMS activating an alarm
 P(A) = P(A|FB) ∙ P(FB) + P(A|HB) ∙ P(HB) Correct detection False alarm (its
= 0.98 0.0001 + 0.01 0.9999 (its probability to probability to be
= 0.0101 be maximized) minimized)
 Given the activation of an alarm, the probability of a fault is
 P(FB | A) = P(FB ∩ A) / P(A)
1 Posterior
1 = [P(A | FB) ∙ P(FB)] / P(A)
2 3 2 Likelihood
= (0.98 0.0001) 0.0101 3 Prior
= 0.0097
 Now let us think about how to improve this conditional probability.

4
Problem-Solving Session 1

Problem 1: Conditional Probability for Detectability Analysis


 Detectability matrix
Fault detection system (part of PHM)

Alarm or fault No alarm (NA)


detected (A)

Missed detection
Faulty Correct detection P(NA | FS) =
(FS) P(A | FS) 1 – P(A | FS)
True
state
False alarm Correct no detection
Healthy P(NA | HS) =
(HS) P(A | HS)
1 – P(A | HS)

5
Problem-Solving Session 1

Problem 2: System Reliability Analysis for Series System


 A steel portal frame structure is subjected to a vertical load V at
the center of the top beam and a horizontal load F at the hinge
joint between the top and bottom-left beams. According to the
yield hinge mechanism theory, the structure has three failure
modes: beam failure, sway failure, and combined failure.

F V F V F V F V

2l

(a) Loading condition (b) Beam failure (c) Sway failure (d) Combined failure

6
Problem-Solving Session 1

Problem 2: System Reliability Analysis for Series System


 Given a fixed, identical yield moment My for all three beams, the
performance (G) functions for the three failure modes are
Beam: G1 = Vl − 4M y
Sway: G2 = Fh − 4M y
Combined: G3 = Vl + Fh − 6M y
 What do the limit-state functions (LSFs) for the failure modes
look like in the V-F space (assume both l and h are constants)?
 How can we draw the failure region for this structural system?

7
Problem-Solving Session 1

Problem 2: System Reliability Analysis for Series System


 What do the LSFs for the failure modes look like in the V-F space?
 How can we draw the failure region for this structural system?
V

P(Esf) = P(E1 ∪ E2 ∪ E3)


Beam: G1 = 0 Ei: occurrence of ith failure
mode, i = 1, 2, 3.
Combined: G3 = 0
Sway: G2 = 0
0 F
V

V
G1 > 0 + G2 > 0
+ G3 > 0

0 F 0 F 0 F
8
Problem-Solving Session 1

Problem 3: Time-Dependent System Reliability Analysis


 Two electronic components are arranged to form a series system.

Component A Component B

 TTF (T1) of component A: = , if t ≥ 0, and 0, otherwise.


 TTF (T2) of component B: = , if t ≥ 0, and 0, otherwise.
 Random variables T1 and T2 are mutually independent: the state of one
component does not affect that of the other.
 Find the joint probability density function (PDF) of T1 and T2.
 Determine the system reliability function RT(t).

9
Problem-Solving Session 1

Problem 3: Time-Dependent System Reliability Analysis


 Find the joint PDF of T1 and T2.
 Solution
 Let us first derive the joint cumulative distribution function (CDF):

FTT1 2 ( t1,t2 ) = P(T1 ≤ t1, T2 ≤ t2 ) = P(T1 ≤ t1 ) ⋅ P(T2 ≤ t2 )


= (1 − e−λ1t1 )(1 − e−λ2t2 )

 Then, the joint PDF can be derived based on the joint CDF as

∂2 FTT1 2 ( t1, t2 ) ∂ (1 − e 1 1 )(1 − e 2 2 )


−λ t −λ t

fTT1 2 ( t1, t2 ) = =
∂t1∂t2 ∂t1∂t2
λ1e−λ t ∂ (1− e−λ t
11 22
) =λe −λ1t1
= 1 ⋅ λ2e−λ2t2
∂t2
= fT1 ( t1 ) fT2 ( t2 )

10
Problem-Solving Session 1

Problem 3: Time-Dependent System Reliability Analysis


 Determine the system reliability function RT(t).
 Solution
 Let us directly derive RT(t):

RT ( t ) = P(T > t ) = P(T1 > t, T2 > t ) = P(T1 > t ) ⋅ P(T2 > t )


= e−λ1t ⋅ e−λ2t = e ( 1
− λ +λ2 )t

 The failure rate of the system is the sum of the failure rates of the two
components (this is applicable to the special case of component TTFs
following exponential distributions).

11

You might also like