Dear Mr Wagner,
I really appreciate your help! I think your method is very clear and simple but I think it has some disadvantages.

Example 1
I would like to quote FMD-2013 introduction:
“The RIAC has collected failure mode and mechanism information from many different sources, with each source generally having its own unique way of reporting this information. In some cases, failure modes of assemblies were presented as constituent part failures of the assembly that failed. In other cases, the actual failure mode/mechanism of the assembly was presented. There are cases in which the failure mode/mechanism classification may appear to be inconsistent, since there may be two sources of data for a particular part, one presenting constituent part failure modes/mechanisms and the other actual failure modes/mechanisms of the assembly. However, in cases where the failure modes/mechanisms listed are a combination of consequences of failure and of constituent part failure, the user can tailor this data to his particular needs by converting one to the other. An example of this approach is “Transformers”, in which one case a failure mode listed may be “Shorted”, whereas another may be “Insulation Failure”. “Short” is an observable mode of failure and the “Insulation Failure” is the site or constituent component of failure. In this case, the user can discard the “Insulation Failure” mode and re-normalize the distribution or, if there is enough confidence that the “Insulation Failure” resulted in a “Short”, the two percentages can be combined under “Short”. In any case, this would have to be accomplished based on a knowledge of the physical properties of the part/assembly and its related reliability issues. The RIAC has attempted to make these listings consistent, where possible. However, in some cases, these two different types of failure modes are presented to allow the user flexibility in tailoring the distributions to his particular needs.”
“Additionally, some of the presented failure modes/mechanisms may be redundant. For example, one failure mode/mechanism may lead to another failure mode/mechanism (e.g., “Corrosion” can lead to “Sticking/Binding”). Also, there may be failure modes/mechanisms with the same meaning (e.g., Shorted/No Operation). In general, the RIAC has reported data as it was reported by the original data source. If the user wishes to merge these failure modes/mechanisms, it can be accomplished by combining the quantities associated with the failure modes/mechanisms to be combined, then re-calculating the associated percentages. ”

And now… we have “Heating Element, Electric”
1)….
2) Inoperative 22,6%
3) Failed to operate 9,3%
4)….
5)….
6)….
7)….
I’m not a native English speaker but for me “inoperative” and “failed to operate” is the same (description is different because this two can result from 2 different data sources or 2 different persons) – we should merge this two . Maybe this is not the best example but I’m sure that we can find better one in FMD.
How to merge this two according to your method? This is impossible. If we use your method we will have to failure modes (22,6% , 9,3%) instead of one (31,9%) which actually weak true FAILURE MODE CRITICALITY NUMBER (FMEA-1629A) – I hope you understand what I’m trying to explain.

Example 2.
I performed FMECA according to 1629A. If I put all FM (even that which are not applicable) then according to FMEA tables I must put information (also for that which are not applicable )about the effect, detection methods and severity classification (to be consistent) – this is time consuming and useless (for that which are not applicable).

Example 3.
I don’t know what you tried to say, and sorry if I’m not right, but beta (1629A) is not “probability of occurrence” but probability that particular failure mode results in indicated severity classification. This also may be a little bit confusing.

I would like to ask you one more thing… when you obtain FR( from NPRD for example) then which data you use “summary” or for specific environment and quality? In my opinion summary is better because we have greater sample size. Sometimes number of failures is 0 and FR depend only on duration test. If we have greater sample size the results are more accurate.

Sometimes also, Env. and quality are not so influential and FR (for specific part) for different env. and quality are the same. If we use conversion factor true FR will be much higher than it should be. When we use “summary” we can’t use conversion factors. Am I right?
Thank you for your help again!
Regards,
MM