Process FMEA (PFMEA) Same Multiple Causes for Different Effects

[BeachBoy]

In my pFMEA, I have different effects related to the same set of process causes.

A rectangle extruded part can be too thick, thin, narrow or wide...
Each of these failures has a different severity.... so you need one line for each of the effects because they have different scores.

But each of these modes are associated to the same set of root causes.. temperature, pression, cycle, etc.

How can I prevent from writing 6 lines for eahc of the 4 effects? Do I really hve to repeat the same 4 lines every time or thereis a way to simplify?

Thanks

 

[adamsjm]

I do not know your processes, but I will uses your information for example purposes. We'll assume that if the product is too wide then the severity is an 8. But if the product is too thick then the severity is a 4. We will also assume that temperature effects both dimensions. The question to ask at this point is: At a temperature higher than the set point temperature, how many occurrences of too thick do I produce and how many occurrences of too wide do I produce & at a temperature lower than the set point temperature, how many occurrences of too thick do I produce and how many occurrences of too wide do I produce (2 causes for 2 failure modes).

Since the first objective in a PFMEA is too reduce risk (lower occurrence rate of the cause for the highest severity), then the preventive control(s) that should be placed in effect is the one(s) which lowers the occurrence rate of the temperature cause for the product being too wide. With these controls in place what is the effect of that control on the temperature caused occurrence rate of the product being too thick?

If all (4 failure mode) occurrence rates are effected the (exact) same and the preventive & detective controls are the same, then just place something like: "Temperature set too high; See Temperature set too high, Failure mode 'Product too wide" from above." in the cause cell.

If the occurrence rates are different, then each cause must be treated separately.

Either way it sounds like a good time for a DOE.

As a professional FMEA facilitator, I prefer that each line is completed (even if copied) because in addition to Risk Management, FMEA's can also generate much more knowledge and management information when completed in that format. You just have to look at the FMEA's from different angles.

 

[Chennaiite]

In my pFMEA, I have different effects related to the same set of process causes.

A rectangle extruded part can be too thick, thin, narrow or wide...
Each of these failures has a different severity.... so you need one line for each of the effects because they have different scores.

But each of these modes are associated to the same set of root causes.. temperature, pression, cycle, etc.

How can I prevent from writing 6 lines for eahc of the 4 effects? Do I really hve to repeat the same 4 lines every time or thereis a way to simplify?

Thanks

'thick, thin, narrow, wide' are considered as potential failure modes or potential effects in your FMEA? (you are using them interchangeably in your post). They are two different approach.

I do not know your processes, but I will uses your information for example purposes. We'll assume that if the product is too wide then the severity is an 8. But if the product is too thick then the severity is a 4. We will also assume that temperature effects both dimensions. The question to ask at this point is: At a temperature higher than the set point temperature, how many occurrences of too thick do I produce and how many occurrences of too wide do I produce & at a temperature lower than the set point temperature, how many occurrences of too thick do I produce and how many occurrences of too wide do I produce (2 causes for 2 failure modes).

My interpretation of AIAG FMEA Manual which is being predominantly used across auto industry suggests to evaluate the Occurrence rating for a stand-alone cause i.e. without linking to likelihood of potential failure mode occurring if the cause occurs. For instance, in 100 process cycles if 10 times temperature overshoots, but only 5 times the process delivers a defective part, the occurrence according to this approach is 5%, where as 10% is the actual occurrence of the cause. While the approach would remain effective in risk evaluation, I will personally opt out of it given the level of subjectivity it is bound to induce. Also, I am interested in hearing how AIAG FMEA is being interpreted and applied by fellow covers in this context.

 

[Jim Wynne]

In my pFMEA, I have different effects related to the same set of process causes.

A rectangle extruded part can be too thick, thin, narrow or wide...
Each of these failures has a different severity.... so you need one line for each of the effects because they have different scores.

But each of these modes are associated to the same set of root causes.. temperature, pression, cycle, etc.

How can I prevent from writing 6 lines for eahc of the 4 effects? Do I really hve to repeat the same 4 lines every time or thereis a way to simplify?

Thanks

This sort of confusion is part of the reason I decided a long time ago that when doing a PFMEA, it's almost always better to consider process failures as the failure modes, and product nonconformities as the effects. The idea is, after all, to control the process.

In your case, I would consider (to use one example) something like "Failure to maintain temperature within required limits" as a process failure mode (it's how the process fails) and then consider what effects the failure will have on the product. While the parameters for temperature, cycle time, pressure, etc. might have similar influence on product conformity, they might have different causes, and it's causes that you want to identify and act upon.

There might be customer who will insist on the orthodox method of identifying product nonconformities as failure modes, but probably not many, especially when you explain how the alternate method does a better job of focusing on process controls.

 

[adamsjm]

Trying to understand and learn.

Jim, I like your concept, but what would the function be for the Failure Mode "Failure to maintain temperature within required limits"? How does that function tie to the Process Step?

Chennaiite, if 5 of 10 temperature overshoots produce defective parts and the other 5 of 10 temperature overshoots do not, then why / how does this cause interact with the failure mode? I would (as a facilitator) skip over the controls section of the FMEA and go directly to the action plans. The behavior of temperature 'control' needs to understood through modeling or DOE.

BeachBoy stated the effects had different severities. FMEA's are implemented to reduce severity first, occurrence second, and detection as a last resort (third).

 

[Jim Wynne]

Trying to understand and learn.

Jim, I like your concept, but what would the function be for the Failure Mode "Failure to maintain temperature within required limits"? How does that function tie to the Process Step?

It's a function of the process step, whatever the process step might be. All process controls are tied to process steps or operations, no? Simply describe the operation and consider the possibilities of process failures.

 

[Chennaiite]

Chennaiite, if 5 of 10 temperature overshoots produce defective parts and the other 5 of 10 temperature overshoots do not, then why / how does this cause interact with the failure mode? I would (as a facilitator) skip over the controls section of the FMEA and go directly to the action plans. The behavior of temperature 'control' needs to understood through modeling or DOE.

Let us say 'too thick' a potential failure mode and 'temperature overshooting' a potential cause. For Occurrence ranking affairs, I will preferably make use of historical data of similar process, if available. Under the circumstance, more often than not, I get to know the data of 'how many times the temperature had overshoot' rather than 'how many times the temperature had overshoot and also resulted in wrong thickness'. I had practically experienced this while performing PFMEA for forging and foundry process. In this case, there are good chances that we would take subjective route to rank occurrence on account of data constraints. I agree, in FMEA, we cannot do away with subjective route always. But in this case, I would prefer to assess the risk based on overall occurrence of the cause 'objectively' albeit exaggerated and would try to reduce or eliminate the cause itself. I am open for more discussion.

BeachBoy stated the effects had different severities. FMEA's are implemented to reduce severity first, occurrence second, and detection as a last resort (third).

Reduction of Severity is much more debatable as well. I don't see any situation in PFMEA that you can mitigate risk by reducing the Severity of the failure, except eliminating the failure itself. I understand Severity of a failure is the Severity of the most serious effect of the failure. Even in DFMEA, product design changes will help in eliminating the failures rather than reducing the severity. Probably, in areas where we incorporate safety warning in the design, there is a room for bringing down severity.