Pilot Process FMEA Occurence Rankings

[Sam1729]

Hello all,


I am a quality engineer currently working in a manufacturing industry. We are working with a pilot process FMEA (We haven't done process FMEA for any part in the past, only the design FMEA). I am the FMEA facilitator and we have so far drafted the possible failure modes, effects and the causes through brainstorming. While working with the ratings, we are stuck with the Occurrence ratings.


We have more than one cause for each failure mode. And the same cause is also responsible for other failure modes as well. (For example, Tool breakage). Now, while giving the occurrence ratings, should we give the ratings based on how frequently the tool breaks or based on how frequently the failure mode occurs due to tool breakage? Effectively, my question is, should the Occurrence ratings be based on how frequently does the failure mode occur or how frequently do the causes occur or how frequently the failure mode occurs due to the particular cause?


And we don't have any previous documented records regarding the occurrences of these causes (not even for the other similar parts ). How should we give the ratings for the occurrences? We haven't had a proper records of rejections in the past but we are trying now to make them proper.


We want to make use of the FMEA to maximum extent to bring in improvements. Your advice and suggestions would be really helpful to us.


Thanks
Sam

 

[Miner]

I recommend basing the Occurrence rating on the probability of the CAUSE occurring, not on the failure mode.

The only way to reduce the Occurrence rating is to reduce the probability of the Cause occurring, so this makes sense. Also, some causes create latent defects that do not result in failure modes until the product has been in use for an extended period of time.

 

[Sam1729]

Thanks a lot @Miner. Could you also tell me how are we supposed to give the occurence Ratings since we don't have any previous records documented for the rejections. Can the Ratings be purely subjective?

Thanks

 

[Miner]

Do you have similar processes that you can use as a surrogate? If you do, use that to estimate your Occurrence rating, if not, then use subject matter experts to (yes, subjectively) estimate it. You might need to rank the causes by probability of occurrence and scale them according to rank order.

 

[Sam1729]

We shall look into some similar occurences in the past or rate them subjectively.

Thanks a lot for the Support.