Scribd
Upload
29 views3 pages

MATLAB Reliability Engineering Lab Exercise1

The document outlines a MATLAB lab exercise focused on time-to-failure analysis of an industrial pump, where students will learn to fit failure data to a Weibull distribution and estimate key reliability metrics. It includes a detailed procedure for data import, Weibull parameter estimation, goodness-of-fit testing, and reliability calculations, along with a design-change scenario to evaluate the impact of a new seal. The final deliverables include MATLAB scripts, figures, and a comprehensive lab report.

Uploaded by

William gichora
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
29 views3 pages

MATLAB Reliability Engineering Lab Exercise1

The document outlines a MATLAB lab exercise focused on time-to-failure analysis of an industrial pump, where students will learn to fit failure data to a Weibull distribution and estimate key reliability metrics. It includes a detailed procedure for data import, Weibull parameter estimation, goodness-of-fit testing, and reliability calculations, along with a design-change scenario to evaluate the impact of a new seal. The final deliverables include MATLAB scripts, figures, and a comprehensive lab report.

Uploaded by

William gichora
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

EES 502 RELIABILITY ENGINEERING

MATLAB Reliability Lab Exercise 1

Title: Time-to-Failure Analysis of an Industrial Pump

Learning objectives
After completing this lab, the student will be able to

1. Fit field-failure data to a Weibull distribution with maximum-likelihood estimation.

2. Estimate reliability, hazard rate, MTTF, and B10 life with confidence bounds.

3. Decide whether a proposed design change meets a required reliability target of R(500 h) ≥ 95
%.

Data set (delivered file [Link])

Vector ttf – time-to-failure (h) for 50 pumps that all ran until failure.

Vector censored – logical array: 1 = failure observed, 0 = suspended (still running).

(The data are simulated but realistic: β ≈ 1.8, η ≈ 3 200 h.)

Lab procedure

1. Import and inspect

```matlab
load [Link]
summary(ttf)

```

2. Weibull parameter estimation

```matlab
pd = fitdist(ttf(censored==1),'Weibull','Censoring',censored);
betaHat = pd.B
etaHat = pd.C

```

3. Goodness-of-fit

```matlab
wblplot(ttf(censored==1))
h = adtest(ttf(censored==1),'Distribution','Weibull');
fprintf('Anderson–Darling p-value = %.3f\n',[Link])

```

4. Reliability & hazard functions

```matlab
t = linspace(0,8000,400);
R = 1 - cdf(pd,t);
haz = pdf(pd,t)./(1-cdf(pd,t));
plot(t,R,'b',t,haz,'r'); grid on
legend('Reliability R(t)','Hazard h(t)')

```

5. Key reliability metrics

```matlab
MTTF = mean(pd) % analytical
B10 = icdf(pd,0.10) % 10-th percentile
R500 = 1 - cdf(pd,500) % reliability at 500 h
fprintf('MTTF = %.0f h, B10 = %.0f h, R(500 h) = %.2f
%%\n',...
MTTF,B10,R500*100)

```

6. 90 % two-sided confidence bounds on R(500 h)


```matlab
[Rlo,Rup] =
mle('wbl',ttf,'Censoring',censored,'Alpha',0.10,...
'CDF',@(t,par) wblcdf(t,par(1),par(2)));
R500lo = 1 - wblcdf(500,Rlo(1),Rlo(2));
R500up = 1 - wblcdf(500,Rup(1),Rup(2));
fprintf('90 %% CI for R(500 h): [%.2f , %.2f]
%%\n',R500lo*100,R500up*100)

```

7. Design-change scenario
Engineers claim a new seal raises the shape parameter to β = 2.2 while keeping the same η.
a) Create the new reliability curve.

b) Compute the new R(500 h).

c) Does it meet the spec R(500 h) ≥ 95 %?

8. Report

One to three-page PDF containing:

– Weibull probability plot with 90 % confidence bands.


– Table of β, η, MTTF, B10, R(500 h) with CIs.

– Short recommendation (≤ 100 words) on the seal change.

Deliverables to submit

1. MATLAB script pumpReliability.m (commented).

2. Figure file [Link] (or .png).

3. Repeat the analysis assuming a 2-parameter log-normal distribution and compare AIC values
to decide which model is preferable.

4. PDF lab report as described above.

You might also like