iSixSigma

Feeding the FMEA

Six Sigma – iSixSigma Forums Old Forums General Feeding the FMEA

This topic contains 4 replies, has 4 voices, and was last updated by  LMH 12 years, 6 months ago.

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #46546

    Scott
    Member

    I have two questions regarding the FMEA. First, I have read that “feeding” the FMEA by either the process map or the most highly ranked variables from the C&E Matrix is acceptable. My question is what do you do and why. Second, I have also seen FMEAs that calculate the RPN by Severity*Occurance*Detection and those that exclude detection, i.e. Severity*Detection. Is this legitamate and if so when is it appropiate to leave out detection? Thanks.

    0
    #154042

    Deep Thought
    Participant

    I suppose you have seen:
     
    RPN = Severity*Occurance instead of RPN = Severity*Occurance*Detection. And “RPN= Severity*Detection” is your oversight.  Pls confirm.
     
    Detection ratings (DET) tell the likelihood that the current system will not detect the cause or failure mode should it occur.
    It rates how well the cause or the failure mode can be detected. Example : 1 = detect every time; ….; 10 = cannot Detect
     
    My opinion is to keep Detection Parameter also if you will not able to change it during the FMEA develop.

    0
    #154044

    Scott
    Member

    What I meant to type was “RPN=Severity*Occurance” not “RPN=Severity*Detection”. Thank you for pointing out the error.

    0
    #154454

    Deanb
    Participant

    Rick,I have seen “detection” effectively substituted with “detection and prevention” in FMEAs. This makes detection relevant only if detection can contribute to prevention.

    0
    #154458

    LMH
    Participant

    A process map or flowchart is ALWAYS a recommended prerequisite for an FMEA, assuming the map is correct and is sufficiently detailed.  A C&E matrix can also be quite helpful, especially for identifying failure modes and the effects of those failures.  It’s immensely important that adequate planning be done PRIOR to your FMEA.  Otherwise, you’ll waste a lot of time, and may not get the results you need.
    Risk is the combination of severity and occurrence.  Quantifying detection capabilities is important too, but reliance solely upon RPN (S*O*D) can sometimes be misleading.  For example, if Severity for 1 failure mode is high, but detection is low (good detection), your RPN may be lower than for another failure mode with lower severity and poor detection.  Which do you work on?  Depends on Occurrence, of course, but also on the thresholds you set. 
     
    One approach I sometimes use is to first calculate Risk level (S x O), and assign risks as low-med-high, or 1-2-3 (using a very simple example in a 3×3 grid).  This prioritizes, in a very basic way, the greatest risks based on the 2 most important criteria.  In another 3×3 grid, compare Risk vs. Detection.  This helps you focus on the higher risks with the poorest detection.  In many cases, results will be similar to using RPN, but not always.  One should always try to reduce severity or occurrence before trying to improve detection.  What people often forget is that an FMEA should ideally be a preventive tool.  How do you prevent a risk from occurring?  Through mistake-proofing, design changes, process changes, process controls, etc. – certainly not by improving detection capabilities.  Sometimes we have no choice (often due to time and $ constraints) and have to rely on good detection, so it’s always a trade-off. 

    0
Viewing 5 posts - 1 through 5 (of 5 total)

The forum ‘General’ is closed to new topics and replies.