How can I manage Occurence Rates in FMEA efficiently when I have many processes?

[Jason Volney]

Hello,

Somebody, please can help me with this?

In my company we make FMEA for each products, today about one thousand FMEAs because have been required by some customers. I am moderator of the FMEA meeting and responsable by update this document of the quality.
I need to answer an action plan of certificator audit in fifteen days!

And we analysed whole the dimensional for each feature of complex part, more one hundred features per part number. E.g.: requirement=>>diameter 16,0-0,1 failure mode=>> major or minor diameter=>> Root cause=>> tool wear , Occurence=>>3 (PPM<10)

But Is it correct to do this? In this case the focus is the product and not process.


Instead in my opinion PFMEA for machining the failure mode could be...
E.g.: requirement=>>Assure the finishing and form of the diameter external =>>failure mode=>> decrease of contact area of tool =>>effect =>> diameter out of specified, roughness out specified =>>Root cause=>> breaking cutting edges, insufficient coolant or intermitent, tool life exceeded, tool wear premature Ocurrence=>>2 (PPM<10)

I pretend to implement this concept of PFMEA like a more generic analysis in only one PFMEA for all process in company, but keeping well robust and possible of management, where I could analyse only the critical features, I think.

How I do to manage automatically the occurrences in the FMEA when I have many process?

I accept sugestions for PFMEA like spreadsheets in machining process or other way to keep the PFMEA updated with the informations of manufacturing.

Thanks and take care,
Jason

 

[Chennaiite]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Hello,

Somebody, please can help me with this?

In my company we make FMEA for each products, today about one thousand FMEAs because have been required by some customers. I am moderator of the FMEA meeting and responsable by update this document of the quality.
I need to answer an action plan of certificator audit in fifteen days!

First and foremost, it is not appropriate for Moderators to be held responsible for updating FMEA. Moderators can streamline FMEA discussion to keep it focused on the subject and can ensure consistency in application of FMEA guidelines and so on. But Moderators updating FMEA will strip ownership out of the guy who actually designs. FMEA in essence helps this guy to do his job, which is to Design, and make sure that his Design is capable of meeting Customer requirements. Moderators job is to bring out this perspective in Designers rather than substituting them. Sometimes Top Management places this responsibility on Moderators to off-load Designers, but I believe this will not work wonders, beyond perhaps answering the Auditors. If my interpretation of your situation is right, I would suggest you to take it up with the Top.

And we analysed whole the dimensional for each feature of complex part, more one hundred features per part number. E.g.: requirement=>>diameter 16,0-0,1 failure mode=>> major or minor diameter=>> Root cause=>> tool wear , Occurence=>>3 (PPM<10)

But Is it correct to do this? In this case the focus is the product and not process.


Instead in my opinion PFMEA for machining the failure mode could be...
E.g.: requirement=>>Assure the finishing and form of the diameter external =>>failure mode=>> decrease of contact area of tool =>>effect =>> diameter out of specified, roughness out specified =>>Root cause=>> breaking cutting edges, insufficient coolant or intermitent, tool life exceeded, tool wear premature Ocurrence=>>2 (PPM<10)

I pretend to implement this concept of PFMEA like a more generic analysis in only one PFMEA for all process in company, but keeping well robust and possible of management, where I could analyse only the critical features, I think.

To me, both the approaches are acceptable if only they serve the purpose of identifying, prioritizing and mitigating risk in the process. Also, it will be worth understanding your Manufacturing process to comment on which approach is generally adopted and therefore might work best. If it is Machining process, as it appears, I would personally choose to go with the former i.e identifying failure mode in terms of Product characteristics, and cause in terms of Process Characteristics. But then, you also have an option of drilling down further in the same FMEA to identify failure at Process characteristics level and identify the next level of cause, if deemed appropriate. In any case, it is uncalled for to consider all the print dimensions. Past problems, customer specific requirements and Critical characteristics provides excellent help in doing a simple but effective FMEA. And one more point, definitely it appears Family FMEA is a good option in your case.

How I do to manage automatically the occurrences in the FMEA when I have many process?

I accept sugestions for PFMEA like spreadsheets in machining process or other way to keep the PFMEA updated with the informations of manufacturing.

Thanks and take care,
Jason

I am skeptical about automatic occurrence rating, especially because it depends on other factors such as Preventive controls. Microsoft Excel or even a Business intelligence software can certainly help in Data collection and basic analysis, but in-depth analysis and interpretation of data requires Human intelligence. A software cannot separate data from factual data. It can serve at best in keeping the Occurrence column filled.

 

[Jason Volney]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Ok, Chennaiite. First, thank you. I understand your point of view on the moderator not be the same person who updated FMEA, however, the company does not want to invest in more staff to this function because we have a very lean staff and some people still do not give due importance for FMEA, they think it is only complete paper. And today it has been, because we can not keep these updated documents.

Today we use Excel spreadsheets and I have already made family FMEA, but the viewpoint of the auditor should have the occurrence reflecting the reality of the process. And this takes time in meetings, so I would like to use only a single generic PFMEA to get easier information management. I agree that nothing can replace human intelligence. But software of this level is very expensive. In my company we are not responsible for the project only by the process. We project leaders who are people who manage the APQP.

Well, this audit was made ​​last week by the certifying organ which is now demanding an action plan within 15 days to show that we have improved our means of determining the occurrence and if we do not have these actions completed in 60 days we can lose the ISO certificate TS 16949.

Today we estimate the occurrence as follows: 1) statistical Pp & Ppk or information scrap factory 2) We calculate the PPM (parts per million) 3) use the table as FMEA manual 4th edition.

 

[Jen Kirley]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Does your scrap factory know which process produced which scrap?

The Occurrence number cannot be unilateral, that is one number cannot be shared for the entire set of failure modes. The expectation of FMEA is to identify risk, and work to reduce or control risk through one or more actions. When the actions succeed, we can lower the occurrence number but there needs to be some means to confirm the actions taken resulted in lowering scrap in the process, not just the total.

If your organization's problem is lack of resources, there is an element for that too. Lowering scrap rates can help pay for the additional personnel's compensation.

I hope I understood you properly.

 

[Jason Volney]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Hi, Jennifer.

It is not lack of resources. What I am wondering is how to estimate the loss based on indicators that already have occurred, and furthermore how to keep the FMEA always updated with these indicators. Do you have some experience with this? Is there some good practice for keeping the FMEAs update with the process data? Is possible to implement something? How I may improve the systematic of updating? I'll be very appreciate if you can help me.

 

[Jen Kirley]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Hi, Jennifer.

It is not lack of resources. What I am wondering is how to estimate the loss based on indicators that already have occurred, and furthermore how to keep the FMEA always updated with these indicators. Do you have some experience with this? Is there some good practice for keeping the FMEAs update with the process data? Is possible to implement something? How I may improve the systematic of updating? I'll be very appreciate if you can help me.

It is hard for me to comment without seeing just what your indicators are, and how accurately they reflect cause-and-effect of actions or inaction to address a failure mode.

FMEAs are considered living documents and should be reviewed periodically by a interdisciplinary group. It seems sensible to pick on one or two failure modes at one time, design and implement an action, evaluate the effect and then update the FMEA to show the effect of your efforts to control. Running too many of these experiments at once could compromise the reliability of your measurements.

It also makes sense to focus on key critical features, as you have said. When doing FMEA we are cautioned to not throw our focus into high RPN, but should also focus on severity. So your plan for the auditor could include an overall evaluation of your FMEAs, with priority placed on critical/key features, high severity, high RPN and ability to address the failure modes.

While it is not practical to promise a complete rewrite of FMEAs within a month to three months for closure, the plan can and should include a review and/or update of all the FMEAs in some way - maybe put them on a review schedule.

I may be naive but I am not clear on why there is a PFMEA for every part. I would understand that for a DFMEA. But aren't some processes shared? If so, wouldn't they be similar and respond similarly to control treatments? If that is so, maybe we are not considering 1,000 experiments when 20 or 30 would cover the same subjects.

I hope I am being clear enough. Your calculation for Occurrence needs to indicate occurrence for that failure mode in that process. Without seeing your FMEA I'm afraid I can't be more specific.

On the whole though, I want to outline the expectations for your response to the auditor.

1) Root cause analysis. This means man-material-machine-method-mother nature; it won't be enough to say "we did not understand the requirements." If that really is the issue, why didn't anyone understand the requirements?

2) Containment. Describe what you did to stop the problem from happening right now. Identify all affected FMEAs and stop the thing cited in the NC from happening again right now.

3) The plan for addressing #1 long term. This might involve

- establish your interdisciplinary team for the project and make your charter.
- reworking how you count types and causes of scrap, points in the manufacturing stream and which processes where they happened, when and why. This data could be used to define which PFMEAs and which failure modes to prioritize.
- take that data to the FMEA Manual's tables and rework your RPNs, starting with the critical key characteristics and arranging to go down the line to minor.
- make a document as a team effort, explaining what you did in exact enough terms to serve as a desk reference when you're not there. Share it as a training tool among other persons who are involved with FMEAs, to align the methods.
- provide oversight to ensure the FMEAs are being updated as per the improved method, finishing critical processes and failure modes first and then overall as the documents come up for periodic review. If various products and FMEAs share the exact same processes and can be expected to have the exact same outcomes from the same actions to address the exact same failure modes, I would ask myself why there are so many repeats? Can some of these FMEAs be merged?

This is sure to take many months so I would not promise more than #1, #2 and perhaps one or two reworked FMEAs to close the NC. I am not sure if that will be accepted or if you will be expected to stay in the building until Christmas and update everything.

I hope this helps!

 

[Jason Volney]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Hi, Jennifer!

Brilliant explanation! I like your sense of humor at the end!

I will follow their guidance, yet still waiting for new suggestions.

Thank you very much!

 

[Chennaiite]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Hi, Jennifer.

It is not lack of resources. What I am wondering is how to estimate the loss based on indicators that already have occurred, and furthermore how to keep the FMEA always updated with these indicators. Do you have some experience with this? Is there some good practice for keeping the FMEAs update with the process data? Is possible to implement something? How I may improve the systematic of updating? I'll be very appreciate if you can help me.

It sounds you are trying to update Occurrence rating every time when Quality rejections occur. If that is your current practice, perhaps you might be over-doing occurrence rating updation and burdening yourself with no serious result. Depending on your Production volume, in my view it is fine to wait for 6 months or 1 year data in order for the data to genuinely reflect the occurrence level. On top of that events like CAPA, Continual Improvements, Process changes of course trigger FMEA review and update if required.

 

[Jen Kirley]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

It sounds you are trying to update Occurrence rating every time when Quality rejections occur. If that is your current practice, perhaps you might be over-doing occurrence rating updation and burdening yourself with no serious result. Depending on your Production volume, in my view it is fine to wait for 6 months or 1 year data in order for the data to genuinely reflect the occurrence level. On top of that events like CAPA, Continual Improvements, Process changes of course trigger FMEA review and update if required.

It was my intention to describe updating Occurrence after a project or action has been done to reduce the nonconformities. Sorry if I was vague.

 

[Jason Volney]

Re: How I can management occurences rates in FMEA efficiently when I have many proces

Ok, honey I believe your comments are very useful!

Thanks,
Jason