Question on PFMEA Failure Effects

[Mandi]

Hi everyone,

I'm implementing a Process FMEA using the AIAG & VDA methodology. I'm having trouble defining Failure Effects related to Failure Modes identified at the Process Step/Operation level. In this instance I'm only interested in internal Failure Effects, those that take place in my plant.
I will write the question first and then clarify it with a practical example:

"In a PFMEA, how should I define the Failure Effects for an operation where the immediate consequence of the Failure Mode is not directly observable, but manifests itself later in the process?"

Example:
Operation: Treatment applied to a component, essential for resin adhesion at a subsequent operation.
Failure Mode: The treatment is not applied effectively.
Direct Effect in production: Poor resin adhesion at a subsequent operation (not directly visible).
Observed Effect in production: Product failure in the end of line test, indicating the lack of resin adhesion => Scrap

Should the Failure Effect for the treatment operation be:
The immediate consequence (poor resin adhesion potential)?
The observed consequence (product failure in final testing/scrap)?
Or both?

Hope this clarifies my doubt. Thank to anyone that will take the time to answer!

 

[Johnny Quality]

Mandi,

I would include both as it will highlight you have a high risk failure (building the part to final inspection only to scrap it) but the immediate effect has a very poor detection (not directly visible). This will highlight the risk and encourage process improvements to minimize this risk from happening and improve inspection at earlier production stages to save costs.

 

[Miner]

Example:
Operation: Treatment applied to a component, essential for resin adhesion at a subsequent operation.
Failure Mode: The treatment is not applied effectively.
Direct Effect in production: Poor resin adhesion at a subsequent operation (not directly visible).
Observed Effect in production: Product failure in the end of line test, indicating the lack of resin adhesion => Scrap

Your example is missing information that is critical to identifying potential failure modes.
Operation: What are the function(s) and the requirements of this operation (i.e., type of material, min/max thickness, area to be coated/uncoated, etc.)?
Failure Mode: This is way too vague. Break out all the potential failure modes based on the function and requirements (e.g., wrong material, too thin, too thick, missing, coated areas uncoated, uncoated areas coated, substrate not clean, substrate too cold/too hot, etc.)

 

[Mandi]

Dear Johnny and Miner,
Thank you for the quick answers!

I understand Miner that my example was way too superficial. I didn't give more details since the focus of the question is on the Failure Effects of the operation.
The operation I'm analyzing is a plasma treatment: amongst other functions and requirements, it increases wettability of surface. This is done to improve resin adhesion at a subsequent operation. Now let's imagine that something goes wrong in the plasma treatment op. so that the wettability of the surface remains low (at the moment I'm not interested into what or how it went wrong).

The Effects I identified in my plant are the following:
- The resin does not adhere to the surface at the casting operation
- The product is scrapped at the end of line performance test (since the resin did not adhere, the final product is not able to reach target performance)

While the first Effect is the direct consequence of the Failure Mode, the second one seems more practical and an actual description of the action taken at the plant (i.e. scrapping) due to the Failure Mode. I'd say considering both as Johnny suggested makes sense (though I'm not sure about the different Detection rating given that we are talking about an effect, to which we associate a Severity).

Hope this clarifies my issue / exposes inconsistencies in my reasoning that we can discuss

 

[Miner]

I would show the direct effect as the resin does not adhere... The secondary effect of scrap could be listed as well and used to inform your severity rating. Is there a scenario where the wettability is just good enough to make it to the customer then fail early (like the cars where the paint peels off in sheets)?

 

[SeanN]

I am also new to FEMA, but I may not agree with you and Johnny (i.e., to include both...). A ‘failure mode’ in FMEA refers to the way in which a component or process could potentially fail to meet the expected performance standards. In your case is the expected outcome is "to pass QC" then the failure could be "not pass QC = scrap", for instance. All previous steps (e.g., poor resin adhesion...) contribute to p1, p2... in RPN cal.

 

[Mandi]

Unfortunately that happens quite frequently actually
The potential Effects on the End User are taken into account in the DFMEA and a linking between the two documents is made through the dedicated software in use so that they also emerge in the PFMEA.

Again thanks for your suggestions (and Jhonny's!). I will probably end up listing both effects

 

[Bev D]

As someone who is very experienced in FMEA - over 30 years - and gas used it successfully to improve products and processes by preventing failures (not just checking a box for some Customer’s filing cabinet) I am in total agreement with Miner. It is all about recording and understanding what the physical effect of the failure mode. It is NOT scrap; it’s the adhesion failure. Simply saying scrap isn’t actionable nor does it increase the knowledge of your organization.

 

[Mandi]

Thank you Bev D for your experienced point of view! May I ask you to expand on the actionable term when you refer to scrapping?

My brief experience in FMEA comes (unfortunately?) only from the AIAG & VDA methodology, which I've been studying, adopting and doubting for approximately 1 year. In the PFMEA section of the Handbook (page 94) it states:
"Failure Effects are given a Severity rating according to:
1. You Plant: the effect of the failure mode assuming the defect is detected in the plant (what action will the plant take, e.g., scrap)"
From this, it emerges that Scrapping is indeed an action taken by the plant, through which the organization is able to determine a Severity rating. Then, in the subsequent pages, it lists examples of Failure Effects: these include both the "inability to perform a certain operation" (that would be the adhesion of resin in my example i think) and "100% of product scrapped". Moreover, the Severity Table provided in the Handbook, gives a different S rating depending on the percentage of product scrapped.

All this to say that it seems to me that a reasonable amount of attention is put on Scrapping as an Effect too.

Again, thanks to everyone who is sharing knowledge. It is really appreciated!

 

[Bev D]

Actionable means that you can take action to prevent a nonconformance, defect, functional failure etc.. It does not mean taking an action to disposition a nonconformance.

As for whether or not you list the adherence or the scrapping: IT IS BOTH. First comes the nonconformance and then the negative effect of the nonconformance (scrap, rework, death of an employee what ever…) that drives the severity rating. You must list the nonconformity(s) that results from a failure mode.

A senior manager asked me why I was so successful in solving and preventing problems while others could not. My answer was that the first thing I accepted (knew) was that causes don’t lie on the same plane. There is a sequence to them. A causes b and b causes c and c causes d but only when x is at the wrong level…and then d causes “the problem” (nonconformance, defect, functional failure, etc.). Then teh Problem ahs some negative effect on the organization. Usually cost, and/or liability…. FMEA is an attempt to link these things in their logical sequence of occurence.