When to take action if an FMEA RPN number is high?

M

malasuerte

Not all industries are required to follow the AIAG criteria. FMEA's are broad based and can be applied to many different processes.

Also regarding the actions taken on Severity of 9 & 10. Since the severity is the one ranking that can't be reduced without a design change, any actions will only have the effect of reducing the occurance or detection. If those are already at very low rankings, than realistically there is probably little more that can be done.

?? Design Change is the ONLY way? Sorry, don't agree.
 

SQ_Joe

Involved In Discussions
1) Severity can only be lowered by a design change.
Reason - Requirements are driven by the design choices of the team. Failure modes are anti-functions of the requirements. Potential Effects of the Failure Mode(s) drive Severity.
Therefore, if there is no change in the requirement, there can not be a reduction in severity. Period.
2) Actions MUST be taken to reduce Requirements of 9 or 10, unless the Occurrence is a 1 (Never occurs) or a 2 with management signoff.
Detection must be capable of measuring an Occurrence of a 1 or 2.
3) RPN is a useless number. See the many articles by Dr. Donald Wheeler.
AIAG teams want to drop RPN as a risk indicator, but because of history, they are reluctant to pull the plug.
 

Bev D

Heretical Statistician
Leader
Super Moderator
3) RPN is a useless number. See the many articles by Dr. Donald Wheeler.

you can also read my paper on the RPN number. It exposes all of the weaknesses of the rating system and it's invalid math as well as the consequences of blindly using any single statistic or mathematically calculated number...it also references Wheeler's article with a direct link
 

Attachments

  • FMEA and RPN Numbers.pdf
    138.9 KB · Views: 590

Stijloor

Leader
Super Moderator
The (automotive supplier) industry has two severe addictions:
  • Risk Priority Numbers (RPN)
  • Capability Analysis (Cp-Cpk and Pp-Ppk)
Both should be removed.

Focus on robust Process Controls and Continuous Improvement.
 

Ajit Basrur

Leader
Admin
you can also read my paper on the RPN number. It exposes all of the weaknesses of the rating system and it's invalid math as well as the consequences of blindly using any single statistic or mathematically calculated number...it also references Wheeler's article with a direct link

Thanks Bev
 

SQ_Joe

Involved In Discussions
Stijloor, Many industries use RPN and Cp-Cpk and Pp-Ppk values. They are easy for SQE's to report and for competitive -"benchmarking" comparisons.
Why engage your brains, when a computer can spit out a number.
 

dhakadmilind

Starting to get Involved
Some of the customer are added some other criteria like multiplication of severity and occurrence ,Severity and detection should not be more than 36 else it is necessary to take a action . But as FMEA is live document so instead on considering some figure for action ,it is better to always decide actions to reduce the RPN . IN our PFMEA CFT, we follow the same rule.
 

SQ_Joe

Involved In Discussions
Hi Bev D,

I have a slightly different perception of Severity rankings. In your paper you have the following:

Recommendations:
Severity Scales: Keep it simple
Create a simple scale that everyone uses. Experience has shown that a 1-5 scale is usually sufficient. 1-3 scales can be perceived as lacking in resolution and often leads to distress as teams struggle with large categories. 1-10 scales tend to have too fine a resolution and teams struggle with the small differences. 1-5 seems to be ‘just right’.

I have facilitated hundreds and hundreds of D & P FMEA’s, and never have I had a problem with a FMEA Team not comprehending nor constructing a 1 – 10 ranking method for their industry or service organization. In rare cases, some industries desire more resolution. The simplification of categories comes by defining regions on the S x O, Criticality Matrix.

On the Criticality Matrix as little as 3 categories/classifications, to a more appropriate selection of 5 categories/classifications can be defined. Industries, customers, or companies can define the definition/selection of the S x O intersections as part of the FMEA pre-work. It is these definitions which lead to the categories of:
“5) Critical”,
“4) Critical with Failure eliminated through preventive control” (if desired),
“3) Significant”,
“2) Annoyance” (if desired), and
“1) Characteristics with Appropriate Actions / Controls Already in Place”.

“5) Critical”, “3) Significant”, and “2) Annoyance (if defined)” Classifications always require Recommended Actions.
The “4) Critical with Failure eliminated through preventive control” Classification requires a special Recommended Action note, such as, “None at this time based on low Occurrence and Detection rankings.” or “Based upon historic data, no action is required at this time.” Have your company’s lawyer(s) to advise your company and teams for specific language.
The “1) Characteristics with Appropriate Actions / Controls Already in Place” Classification should have the Recommended Action of “None” as suggested in the AIAG FMEA Reference Manual.

The Criticality Matrix, a pre-FMEA planning activity, takes all of the guess-work out of defining Product/Process Characteristics and one dimension of when and type of Recommended Actions are required. Other pre-FMEA planning activities make FMEA development quite an easy task after a company or team creates them. These become business game changing strategies.
 

Bev D

Heretical Statistician
Leader
Super Moderator
I am not all that passionate about how many categories go into a severity rating. Certainly there isn't a single perfect or optimal number that fits everyone's needs. for me it's not whether or not it's easy to use 10 different categories, it's more about the value that more categories will give us. does it actually help us get to preventing and mitigating Problems? Will a 9 drive a different action than a 10?

Unlike Severity I am passionate about the whole occurrence rating. SXO is still bad math...
 
Top Bottom