KPC & KCC verification and updates

#1
Hello everyone,

I'm new to the forum and excited to be posting on here.

I know there have been threads on how to generate KPC and KCCs, but I would like some suggestions on how KPC's are updated based on the PFMEA.

For example, if I establish during the DFMEA that a specific feature, lets say the sheet metal thickness is a KPC due to the criticality (severity x occurrence) being high, however during the PFMEA analysis, we establish that the resultant KCC has a occurrence rating that is low, which lowers the criticality. After this PFMEA assessment of the potential KPC identified during DFMEA, could I go back and "remove" the KPC identifier in the drawing and DFMEA since it's established that the KCC associated is well controlled?

The goal is not to have too many KPCs on the drawing, since a lot of our severity levels are high.

Any suggestions are much appreciated.
 

Coury Ferguson

Moderator here to help
Staff member
Super Moderator
#2
As long as you are the design authority, I wouldn't see a problem. As long as you can substantiate your decision(s) based upon statistics or similar identifiers, for the reduction. Prove to your customer, management and such, the reason why the decision was made.

Just my opinion.
 
#3
As long as you are the design authority, I wouldn't see a problem. As long as you can substantiate your decision(s) based upon statistics or similar identifiers, for the reduction. Prove to your customer, management and such, the reason why the decision was made.

Just my opinion.
Thanks for the reply.

A design engineer may argue, that even though the specific feature they identified as a KPC may be justifiably well controlled, it does not take away from the severity of the effect of the failure mode, which may require it being identified as a KPC on the drawing.

What are your thoughts on that?

Thanks
 

Coury Ferguson

Moderator here to help
Staff member
Super Moderator
#4
Thanks for the reply.

A design engineer may argue, that even though the specific feature they identified as a KPC may be justifiably well controlled, it does not take away from the severity of the effect of the failure mode, which may require it being identified as a KPC on the drawing.

What are your thoughts on that?

Thanks
I have had debates many times with the Design Engineers. Design Engineers think about the theory not how it will affect manufacturing. I have nothing against Design Engineers. But, sometimes they are only thinking about concept and theory and not how the part is made/manufactured.

If the evidence through statistical analysis reflects that the occurrence is a low risk, why place that type of critically on that feature? I am not saying that in the event there was a fallout of that call-out that it couldn't return back to the original risk assessment.

When all of your supporting data does support that decision, of reduction, it is a waste of time and money to keep that identified as a high risk.

This is just my opinion.
 
#5
I have had debates many times with the Design Engineers. Design Engineers think about the theory not how it will affect manufacturing. I have nothing against Design Engineers. But, sometimes they are only thinking about concept and theory and not how the part is made/manufactured.

If the evidence through statistical analysis reflects that the occurrence is a low risk, why place that type of critically on that feature? I am not saying that in the event there was a fallout of that call-out that it couldn't return back to the original risk assessment.

When all of your supporting data does support that decision, of reduction, it is a waste of time and money to keep that identified as a high risk.

This is just my opinion.
I would have to agree with you regarding Design Engineers (generally speaking)

I guess it comes down to KPC criteria during DFMEA

Should a severity of 8 and above ALWAYS denote a KPC? That's industry standards, but I believe that will lead to way to many KPCs (or potential KPCs depending on how you look at it) being defined.
 

John Predmore

Involved In Discussions
#6
If the evidence through statistical analysis reflects that the occurrence is a low risk, why place that type of critically on that feature?
The severity of a failure mode remains high even though the occurrence may be low. To a nation of consumers, a failure rate of 1 ppm has low criticality. To the one consumer out of a million who is injured by a malfunction, the criticality is high. Process controls and error-proofing measures may reduce the risk of occurrence, but over time, changes occur which were never anticipated or countermeasures have a way of becoming removed or defeated. Someone once said, it is tough to make any design foolproof because they keep making better fools. Marking a KPC on a design document ensures that the potential risk is communicated and not overlooked, now and long into the future.
 

GRP

Involved In Discussions
#7
Are you in automotive? Ford deals with this with the SCCAF - special characteristics confirmation and agreement form. Since the DFMEA identifies potential special characteristics, the pfmea sets out to confirm or not the inclusion of special controls for the characteristic. In this way you can have many KPCs in your drawing, but you do not have to necessarily do anything about it.

The lines above presuppose there is a client-customer relationship. Are you responsible both for design and production?

I´d be interested in knowing if anyone has seen KPCs removed from the drawing based on the PFMEA (and/or SCCAF). Which brings us back to your original question, I realize.
 

Bev D

Heretical Statistician
Staff member
Super Moderator
#8
Since severity never changes and design engineers and process engineers and manufacturing supervisors move on and product / process changes are inevitable, Any KPC should remain on the drawing as critical until the function itself is removed. The point of identifying KPCs is to ensure that they are well controlled so that they do not fail. Removing them form the drawing simply because they are well controlled today, obviates that intent.
 
#9
Since severity never changes and design engineers and process engineers and manufacturing supervisors move on and product / process changes are inevitable, Any KPC should remain on the drawing as critical until the function itself is removed. The point of identifying KPCs is to ensure that they are well controlled so that they do not fail. Removing them form the drawing simply because they are well controlled today, obviates that intent.
So in essence, when the Design Engineer states that if they found the product feature as a KPC due to a severity of 8 or greater, even if there is an occurrence rating of 1, manufacturing has no say in whether it remains a KPC or not.

I am not the design responsible engineer, but I am finding that way too many KPCs are being called out and I feel as though it is going to lead to concern with our customer and more scrutiny. I don't necessarily agree that a Severity rating of 8 or above should single handedly trigger a KPC. Occurance should always be taken in to account.

Would anyone agree with this?
 
#10
Are you in automotive? Ford deals with this with the SCCAF - special characteristics confirmation and agreement form. Since the DFMEA identifies potential special characteristics, the pfmea sets out to confirm or not the inclusion of special controls for the characteristic. In this way you can have many KPCs in your drawing, but you do not have to necessarily do anything about it.

The lines above presuppose there is a client-customer relationship. Are you responsible both for design and production?

I´d be interested in knowing if anyone has seen KPCs removed from the drawing based on the PFMEA (and/or SCCAF). Which brings us back to your original question, I realize.
I have spent many years in automotive and am familiar with the YCs and YSs from reviewing Fords FMEA handbook thoroughly.

I will have to check out the SCCAF form.

I just still feel as though there is no clear methodology of KPC revision based on PFMEA analysis.
 

Top Bottom