Modeling failure data

[bob valerius]

Hi

I have been asked to look at different methods for modeling failure data. On one of our programs, we model the reliability by using predictions from various sources. This program has been in test (sample size = 1) since 9/99. The approach we are taking is that when a component or subassembly fails, we replace its predicted failure rate with the actual. This of course almost always results in a lower reliability estimation. When the FRACAS procedure results in a corrective action being implemented, we then replace the actual failure rate with the predicted failure rate. This is how we report monthly on our reliability metric.

Does anyone know of a reputable alternative approach? It has been suggested that since predicted failure rates are an average, to replace the predicted one with the actual, which is only 1 data point, is way too conservative. Any ideas or suggestions would be appreciated. Thank you very much.

Bob Valerius
410-260-5247

[RS Support]

I am assuming that you have a single unit and you are recording times between failures and repairs. If this is the case an NHPP model can be used, popularized in Reliability Growth analysis.
There are multiple books dealing with this subject. Reference [8] at http://www.reliasoft.com/Books/index.html includes a chapter on this.

Additionally ReliaSoft also offers a reference on this, http://www.reliasoft.com/RSPub.htm (Reliability Growth Reference).

Hope this helps

[DGibson]

I am trying to estimate reliability performance on products with very little failure data, at this point in time.

If I select a 2 parameter Weibull with Beta and Eta values that are based on past history and use the Monte Carlo tool to generate times-to-failure data that fit the selected distribution, how much confidence should I have in reliability estimates based on this generated data?

[srikanthjoshi]

I have failure times(in terms of km) of components of a truck. Can I blindly use 2 parameter Weibull distribution for all components?
Also am I correct in using conditional weibull reliability model to estimate the reliability of a truck fleet for an additional usage of 20000km?

Can some body help?

[Pantelis RS]

Doing anything “Blindly” is never a good idea…

In general a 2-P weibull model should work, but it would behoove one to check model accuracy before proceeding.

As for the conditional probability, yes this would be the correct approach.