FMEA RPN Reduction Process

I recently received this as a PM.
I am posting this here for discussion

hello Howard, my main interest is how to address this RPN process from a bad initial FMEA, shall I returned to the drawing board and re do the fmea or try to get it reduce with this approach,

Click to expand...

First why was the FMEA bad?

Was it not through enough? Did it not define the failure modes well? Did it use the wrong rating system?
I would ask these questions first.

I would use this as a training excercsie and using the basis of the current FMEA train everyone to prevent another "bad" FMEA.

In the design phase a FMEA is as good as the wild guess of the participants. Hazards can be missed, the actual occurence rate might not be equal to the occurance rate in the FMEA, etc. Therefore the FMEA must be maintained throughout the lifetime of the product/process and be amended to meet factual data. The factual data might either in- or decrease the RPN, but any change in the RPN must be justified.

First of all, Do we need to really look in to the RPN number while doing FMEA?
OR to work on reducing S X O.

jkuil said:

In the design phase a FMEA is as good as the wild guess of the participants. Hazards can be missed, the actual occurence rate might not be equal to the occurance rate in the FMEA, etc. Therefore the FMEA must be maintained throughout the lifetime of the product/process and be amended to meet factual data. The factual data might either in- or decrease the RPN, but any change in the RPN must be justified.

Click to expand...

Design FMEA's are often developed considering historical data. Past failures associated with similar products are often included. So, I do not quite agree with the "wild guess" comment.

Stijloor.

Stijloor said:

Design FMEA's are often developed considering historical data. Past failures associated with similar products are often included. So, I do not quite agree with the "wild guess" comment.

Stijloor.

Click to expand...

Yes--part of the process should be to separate the wheat from the chaff, and the signals from the noise. This is why the FMEA process (design or process) should include at its outset a brainstorming process that includes wild guesses, but progresses to eliminate them as far as possible in terms of experience and reasonable probability.

This is not to say that a we should expect a FMEA team to be omniscient--Murphy is always lurking in the shadows. In that sense jkuil's advice to maintain the FMEA document in light of new discoveries is well taken.