Dismiss Notice
You must be a registered member in order to post messages and view/download attached files in this forum.
Click here to register.

Using the correct Severity value when calculating RPN or Action Priority

Discussion in 'FMEA - Failure Modes and Effects Analysis' started by karl w, Jan 3, 2019.

  1. karl w

    karl w Member

    Joined:
    Jan 2, 2019
    Messages:
    5
    Likes Received:
    1
    Trophy Points:
    2
    As I understand correctly, there are differing opinions around the ranking scale of 1-10 vs 1-5 vs no numbers, whether to multiple the values or use the face value (9x3x5=135 vs 3x9x5=135, or 935 vs 395), and that there is (or is in the works) a change coming removing RPN and replacing it with Action Priority H-M-L.

    My question is specific to the Severity value. I had read a while back (sorry, I don't recall where) that ALL Possible/Potential Effects on Failure should use the Highest Severity score when calculating the RPN for ALL Effects, Causes & Controls within a single Potential Failure Mode. Assuming I want to keep my current 1-10 scale and then use the new Action Priority H-M-L process (specifically the PFMEA Action Priority Logic - thank you Andy Nichols for this!), is it correct to determine the total SOD value by using the HIGHEST Severity Value within the failure group? Or should each effect, cause & control be individually determined? I've attached a screen shot of my Excel PFMEA template that I've been using which I'm in the middle of modifying to remove RPN and replace with Action Priority. Using the dummy info shown, 10 is my max Severity value for ALL potential effects on failure (for the first Potential Failure Mode) and is therefore used for all 10 RPN (now AP) calculations to determine the H-M-L Action Priority. (I hope I'm making sense here - sorry, this is somewhat difficult for me to explain since I'm fairly new to this).

    So currently my SOD scores are 10-9-3=HIGH, 10-8-4=HIGH, 10-7-5=HIGH, 10-6-3=HIGH, 10-5-4=MEDIUM and so on (always using 10 for the first Potential Failure Mode). The point being that Severity is 10 for ALL Potential Effects on Failure for the first listed Potential Failure Mode. Severity could be different for the next Potential Failure Mode (which I have shown as 2-9-2=HIGH in the screen shot).

    I'm trying to keep this simple and don't want to open up a debate - I would just like to get clarification around whether you're supposed to use 10x9x3, 10x8x4, 10x7x5 and so on OR if 9x9x3, 7x8x4, 5x7x5 and so on. Which method is correct? For RPN? For AP? Or Both?
     

    Attached File(s): 1. Scan for viruses before using. 2. Report any 'bad' files by reporting this post. 3. Use at your own Risk.:

    Last edited: Jan 3, 2019
  2. Andy Nichols

    Andy Nichols Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    3,671
    Likes Received:
    1,856
    Trophy Points:
    112
    Location:
    In the "Rust Belt"
    Karl: Happy to help. Firstly, do you have any established, descriptive criteria for the various numerical values? I think that's going to help us understand.
     
  3. karl w

    karl w Member

    Joined:
    Jan 2, 2019
    Messages:
    5
    Likes Received:
    1
    Trophy Points:
    2
    I'm currently in the process of building everything from scratch and do not have descriptions for the 1-10 rating scales. My question for using only the highest Severity value came from asq.org (https://asq.org/quality-resources/fmea), see line 7. Their Figure 1 example, which is hard to read, shows only 1 Severity score for 3 Potential Failure Effects for the 1 Potential Failure Mode and line 7 specifically states to use only the highest Severity level per Potential Failure Mode if a Potential Failure Mode has more than 1 effect. Is this normal? Is this standard or do some companies/people input individual Severity scores for every effect? And what is the rationale for using only the highest Severity? I can certainly follow what ASQ says to do without knowing why, but I'd prefer to understand the reasoning behind why they say to do this.
     
  4. John C. Abnet

    John C. Abnet Well-Known Member

    Joined:
    May 23, 2017
    Messages:
    432
    Likes Received:
    292
    Trophy Points:
    62
    Location:
    Upper Midwest- USA
    Good day @karl w ;
    Good questions...

    In regards to your methodology I would council that there are two considerations...
    1- Any specific requirements contractually (i.e. supplier handbook etc....) specified by your customer.
    2- Whatever best serves your organization to ensure...
    "An analytical methodology…to ensure…potential problems have been considered and addressed throughout…process development… an effective tool to quantify risk so it can be analyzed prioritized, mitigated, or eliminated.


    Regarding your question specific to Severity level, consider this example...
    Operation/Requirement= Securely attach the wheel
    Process failure mode= Unsecure wheel attachment
    Potential effect of failure=
    * Fender damage
    * Dead driver


    As you can see in this example, there are frequently multiple "Potential effect of failure". However, this clearly shows that some "...effects..." are much more severe than others.

    It is assumed/likely that Occurrence and Detection factors (values) will often be the same from one failure mode to another. When that occurs (assuming you apply the same "Severity" to all), how will you manage Unsecure Wheel Attachment from Rattling Dome Light ?
    (i.e. how will the required emphasis be communicated to the teams using the PFMEA?)

    It is good to identify all potential effects (so that the teams using the PFMEA in the future are aware of what all has been considered) but identify and USE only the highest severity as the multiplication factor .

    Reminder, that "severity" never changes unless the product design or application is changed (for example, if the bolt applied to hold on the wheel is changed to only hold on the tailpipe, etc..etc..)

    Hope this helps. Be well.
     
  5. tony s

    tony s Well-Known Member

    Joined:
    Sep 10, 2015
    Messages:
    1,171
    Likes Received:
    867
    Trophy Points:
    112
    Location:
    Laguna Philippines
    Using the current 4th ed of AIAG's FMEA manual, a Note in page 84 mentioned this statement: "If more than one potential effect is identified... all may be listed, but for purposes
    of the analysis, only consider the worst case when documenting the resulting Severity ranking".
     
    John C. Abnet and Miner like this.
  6. fabio

    fabio New Member

    Joined:
    Oct 15, 2019
    Messages:
    1
    Likes Received:
    0
    Trophy Points:
    1
    Please excuse the noob here but I can't find an answer for a simple question. Can RPN value be zero after control action implementation? Risk eliminated? It doesn't seam so.
     
  7. John C. Abnet

    John C. Abnet Well-Known Member

    Joined:
    May 23, 2017
    Messages:
    432
    Likes Received:
    292
    Trophy Points:
    62
    Location:
    Upper Midwest- USA
    Good afternoon @fabio ;
    No. The RPN is a product of the AIAG manual (current 4th edition), wherein, the suggested SOD scales are all based on factors of 1~10. Therefore, the "math" will never equal 0.

    Hope this helps.

    Be well.
     
  8. tony s

    tony s Well-Known Member

    Joined:
    Sep 10, 2015
    Messages:
    1,171
    Likes Received:
    867
    Trophy Points:
    112
    Location:
    Laguna Philippines
    Based on the scale of 1-10 for each criteria Severity (S), Occurrence (O) and Detection (D), the highest RPN value can reach as high as 1000 and as low as 1. Although the value of RPN starts at 1 and ends at 1000 this can often lead to incorrect assumptions as illustrated below:
    upload_2019-10-16_6-52-3.png
     
    Bev D likes this.
  9. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    471
    Likes Received:
    512
    Trophy Points:
    92
    Location:
    Maine
    The RPN. Well some people still believe that the earth is flat and that the moon landing was a government trick...<sigh>
     
    Andy Nichols and tony s like this.

Share This Page