Success-Run test when a unit fails...

[Steve]

Say I have a success-run test on 100 units. I plan to run them for 6 months. I want to demonstrate 97%Reliability at 95%LCL.

After 4 months I have a failure.

I say that I have demonstrated 97%Reliability at 95%LCL at 4 months since up until that time, I had no failures (though the original goal is out the window). Is this correct?

[Tarik El-Azzouzi]

First, I just wanted to say that there are different reliability demonstration methods you can use depending on the assumptions you can make, for more details, please visit http://www.weibull.com/LifeDataWeb/demonstration_test_design.htm.

I’ll assume that you are using the Non-Parametrical Binomial approach. Yes, you can say that you demonstrated that you have R=97% at CL=95% at 4months by testing 99 units and seeing no failures up until 4months. If you allow one failure, then the method shows that you can demonstrate a R=97% at CL=95% at 6months if you had tested 157units and only experienced 1 failure. So you can test 58 additional units for 6months and hope that you have no more failures.

Another way to look at this is to determine the confidence level that you can use to say that you have demonstrated your original goal. R=97% at 6months, 100 test units and 1 failure correspond to a 80.5% confidence level.

Or you can also say that, with 1 failure, you demonstrated a reliability of 95.3% not 97% at 6months.

Using the Parametric Binomial method, you can see that if you were to extend testing the remaining units that didn’t fail to 7months (1 additional month for each unit) and see no more failures then you can still demonstrate your original goal.

The above analysis was done using the (Design of Reliability Test) DRT tool in Weibull++

good luck

[Pravin Kadekodi]

Steve,
What is the units use life in field? In other words the 6 months of test represent how many cycles/miles/years of use in field?
The test you are running, what failure mode/mechanism is it simulating?
for example : spring test would look for spring fatigue, gear tooth bending fatigue?

Knowing the ratio of the test life to actual life and the failure mechanism (estimate of weibull slope) you can use weibayse to conclude the reliability and confidence demonstrated.

Weibayse eqn
R = exp [(ln(1-CL)/(n(test time/actual time)^slope)]

Hope it helps....