PFMEA severity score different applications

[AlanC]

Hi All,
how to manage multi customer applications for the same process in terms of severity scoring. E.g The product produced may go into a safety critical application and score a severity of 9 or a non safety application and would score eg 4. The product goes down the same process but multi end customers can use the product in various ways. Do you have one PFMEA and have several lines for the different severity scores depending application, create several PFMEA's depending application ( but this turns it into a product and not process FMEA as the process is the same for both) or do you just enter the most severe case within the FMEA
cheers
Alan

 

[optomist1]

different uses/applications dictate different RPNs, severity...thus different PFMEAs....my guess is that the customer SQE or release engineer would state the same

 

[leftoverture]

I believe the PFMEA is a risk analysis tool of your process and, therefore, I would suggest using the worst case application to determine your rankings. And then take actions based on the risk associated with the higher RPN and/or criticality scores. Doing two because of the different applications makes one a meaningless paperwork exercise so I would avoid this practice.

 

[optomist1]

I disagree, and if is in the automotive field, unless the parts/assemblies are identical in all aspects, the OEM SQE/Release engineer would likely request separate PFMEAs/Control Plans...

 

[leftoverture]

I disagree, and if is in the automotive field, unless the parts/assemblies are identical in all aspects, the OEM SQE/Release engineer would likely request separate PFMEAs/Control Plans...

I interpret that the original poster as indicating the parts/process is identical. With that in mind, I still say we need to push back on our customers when they ask us to do things that don't make sense. We don't want to let the automtive guys push us around. A process FMEA is about OUR process, not theirs. We have control over how it is ranked, and we can see this in the AIAG manual with a careful reading. It doesn't happen too often but I push back when my customers disagree with something in our PFMEAs. I mean this with all sincerity....letting customers badger us into meaningless paperwork exercises is not a lean practice and is a disservice to our own companies! We may not win every argument but I sure believe we need to try.

 

[GRP]

I disagree, and if is in the automotive field, unless the parts/assemblies are identical in all aspects, the OEM SQE/Release engineer would likely request separate PFMEAs/Control Plans...

A few months late, but for whatever it is worth, I agree with your statement. Suppliers should coordinate and establish an effects and severity ranking table with their clients, based on the application.

The process is the same, but the application is different. The PFMEA severity ranking should be tailored to the application. Your organization might have different rules for processing, maintaining and documenting characteristics related with government regulation or safety.

In the interest of the supplier, having separate PFMEAs will avoid burdening with responsibility a project with lower severity.

 

[leftoverture]

Honestly I think you are confusing DFMEA and PFMEA. For the DFMEA, I agree with you. Different applications may have different rankings. But for the PFMEA, it is about the process. The owner of the process should set the rankings based primarily on the process not the application. For example, in my field of plastic molding, a threaded insert of the wrong length could have a low severity in the DFMEA because the part would still assemble and perform well whereas in the PFMEA, it could have a high severity because it could damage the mold and stop production. Always remember that the PFMEA is a tool primarily for the process owner, not the end user.

 

[GRP]

If you have a DFMEA from your own organization or from the client, you should take the effects and severity and incorporate them to the PFMEA rankings. This is an opinion.

You should do this because the effects of a failure mode are not only related to the process owner, but also to the client, the end user, and govt/safety regulations. And you have to account effects both on process and product. This is in well-known FMEA sources.

Think of a trim component for the automotive industry. What severity would you give to a tiny dent or scratch? If you just take the guideline from any of the said sources you could say a 2-4, maybe a 5 or 6. But if you have tuned your rankings and aligned them with the DFMEA and/or the client you could end up with an 8. The client might argue the appearance of a trim component is its primary function, and anything less than none is unacceptable.

I can think of another case of a plastic injected holder. This seemingly inoffensive component held together a set of parts which constituted a device linked to a function, the failure of which implied a govt regulation noncompliance. A "process-owner-only" PFMEA could have ranked a failure mode related to the characteristic with a 7, say scrap < 100% (process effect), but the whole picture yielded a 9.

Imho, practitioners of PFMEA should tune their effects and severity rankings by taking input from an available DFMEA or by coordinating this with their clients, and always consider both process and product effects; generally pooling all effects and sev together and using the highest one.

Back to the original post from AlanC, if you look at SAEJ1739 it is stated that manufacturers of stock products may not know the applications/uses of their products, .e.g. screws. In this case the PFMEA can consider a general severity such as no primary function, degraded primary function, and so on...

 

[Jim Wynne]

If you have a DFMEA from your own organization or from the client, you should take the effects and severity and incorporate them to the PFMEA rankings. This is an opinion.

You should do this because the effects of a failure mode are not only related to the process owner, but also to the client, the end user, and govt/safety regulations. And you have to account effects both on process and product. This is in well-known FMEA sources.

Think of a trim component for the automotive industry. What severity would you give to a tiny dent or scratch? If you just take the guideline from any of the said sources you could say a 2-4, maybe a 5 or 6. But if you have tuned your rankings and aligned them with the DFMEA and/or the client you could end up with an 8. The client might argue the appearance of a trim component is its primary function, and anything less than none is unacceptable.

I can think of another case of a plastic injected holder. This seemingly inoffensive component held together a set of parts which constituted a device linked to a function, the failure of which implied a govt regulation noncompliance. A "process-owner-only" PFMEA could have ranked a failure mode related to the characteristic with a 7, say scrap < 100% (process effect), but the whole picture yielded a 9.

Imho, practitioners of PFMEA should tune their effects and severity rankings by taking input from an available DFMEA or by coordinating this with their clients, and always consider both process and product effects; generally pooling all effects and sev together and using the highest one.

Back to the original post from AlanC, if you look at SAEJ1739 it is stated that manufacturers of stock products may not know the applications/uses of their products, .e.g. screws. In this case the PFMEA can consider a general severity such as no primary function, degraded primary function, and so on...

Here's the way that it should work: The customer does the DFMEA and the results are incorporated in the design output--drawings and specifications, mainly. Thus if the supplier satisfies the specifications given to them, the design requirements will be met. The PFMEA should address the supplier's process(es) and not the customer's. It's the responsibility of the customer to ensure that they adequately communicate their requirements. When developing Risk Priority Numbers (RPNs) for the manufacturing process, the supplier should take into account only the potential failures of their own processes, and the concomitant risks to their own business, not the customer's.

Over the years I've seen many situations where a supplier met all of the specifications, but somehow or another the product didn't work as intended. The operation was a success but the patient died. This is solely the responsibility of the design authority. This is not to say that a supplier shouldn't bring to light things that the customer might have missed.

 

[leftoverture]

Jim Wynne you are a genius!