FMEA – Vol. XI: Commentary, Critique, and Caveats

     Standards and guidelines published by industry groups or standards organizations typically undergo an extensive review process prior to acceptance.  A number of drafts may be required to refine the content and format into a structure approved by a committee of decision-makers.
     As one might expect, the draft review and approval process is not consistent for every publication.  The number of drafts, time to review, and types of changes requested will vary.  Though each review is intended to be rigorous, errors often remain in the approved publication.  The content may also require interpretation to employ effectively.
     This is certainly true of the aligned AIAG/VDA FMEA Handbook.  In this installment of the “FMEA” series, the Handbook’s errors and omissions, opacities and ambiguities will be discussed.  Where possible, mistakes will be corrected, blanks filled in, and clarity provided in pursuit of greater utility of the Handbook for all FMEA practitioners.

     AIAG compiles errata documents to inform users of known errors within its publications.  The most recent version of the AIAG & VDA FMEA Handbook 1 Errata Sheet (English – June 2020) is 10 pages long.  It contains many corrections of terminology and formatting, many of which are simply “copy/paste” or “find/replace” errors.
     With an ostensibly rigorous review process in place, why did so many simple, blatant errors go unnoticed until after publication?  Many readers will not have the Errata Sheet at hand when learning the aligned method; some may not even be aware of its existence.  This puts those readers at risk of confusion, misapplication, and ineffective analysis, negating the expected benefits of the aligned approach.  It is disappointing that a highly-anticipated release of an update of this magnitude be so seemingly carelessly done.
     This discussion is not intended to be a line-by-line analysis of the Handbook; that would be redundant and futile.  Rather, it aims to highlight areas where clarification is needed or other improvement opportunities exist.
 
     The installments of the “FMEA” series that discuss the AIAG FMEA Handbook are written to parallel the presentation of information in the Handbook (mostly – more on that later).  This is done to simplify referencing both the original (the Handbook) and the companion resource (“The Third Degree”) concurrently, whether to learn the aligned method or to aid subsequent FMEA activities.
     The “FMEA” series also uses the copy/paste shortcut to maintain consistency, for example, between DFMEA and PFMEA discussions.  Therefore, it exhibits the same potential for error as the Handbook.  However, “The Third Degree” has no committee of reviewers whose responsibility it is to ensure the accuracy of the initial publication; it relies on the author’s proofreading and editing.  Reader feedback is appreciated to facilitate further correction and improvement.
     In addition to offering recommendations to improve the efficiency of analysis, the installments of this series that present visual AP tables (Vol. IX) and recommended formats (Vol. X) also contain some critique and caveats that would be appropriate to include here.  Readers are encouraged to review those remarks in the context of the full discussion rather than reiterating them here.  Critical remarks are intentionally limited in other installments of the series to maintain focus on the 7-Step Approach as presented in the Handbook.
     The Handbook discusses the use of a “foundation” or “baseline” FMEA as a starting point for analysis of a new product or process. This practice is similar to that described in “P is for Process,” where existing product and process information is stored for use across many programs.  The key difference is that the Handbook explicitly states that the intention is to modify the foundation FMEA to suit the new product or process, whereas “P is for Process” focuses on direct reuse.  Stated another way, the Handbook prescribes using the foundation FMEA to look for differences, while “P is for Process” looks for similarities between the new product or process and the existing.
     In this series, “FMEA” is expanded to Failure Modes [plural] and Effects Analysis.  Many sources, including the current and previous AIAG and VDA Handbooks, use the singular form of mode.  It may seem trivial, but it is an important reminder that a Focus Element may have multiple Failure Modes and that thorough, effective analysis requires remaining cognizant of this simple fact.
     Other terminology used in the series also may differ slightly from other sources.  Examples of interchangeable terms include:

• Failure Mode, FM, failure mechanism.
• Cause of Failure, FC, failure cause.
• Effect of Failure, FE, failure effect.
• Structure tree, function tree (structure tree with function information), failure tree (structure tree with failure information; not used in this series to avoid confusion with “fault tree” or FTA).

 
     A series of comments, points of concern, and suggestions are presented below.  Though there is some overlap, or applicability to more than one FMEA type, the remarks, again, approximately parallel the Handbook’s presentation order (i.e. Design FMEA, Process FMEA, FMEA-MSR).  Given that, the remarks are still somewhat disjointed, precluding a coherent, flowing discussion; therefore, a bullet-point format is used to reflect this.

• Document all FE, FM, and FC before evaluating Severity (S); document all Controls before evaluating Occurrence (O) and Detection (D); document all Monitoring Controls and System Responses before evaluating Frequency (F), Monitoring (M), and S after MSR.  This prevents distraction by preoccupation with high ratings that could lead analysts to move to next steps before a thorough analysis is complete.
• The logic behind the priority assigned to any combination of S, O, and D ratings in the AP Table is not provided in the Handbook.  This gives the appearance of being somewhat arbitrary and an attempt to remove judgment from FMEA.  This is unwise, as it is judgment that ensures the highest priority actions are taken first!
• To prioritize actions with equal Action Priority (AP), return to the method used in classical FMEA, where S, O, D, and RPN (Risk Priority Number) are used to evaluate the relative urgency of actions to be taken.  See “Vol. III:  ‘Classical’ DFMEA” or “Vol. IV:  ‘Classical’ PFMEA” to review this process.
• Columns are combined on the aligned form that were recommended to be split on the classical FMEA form (e.g. Functions and Requirements, Step 3, DFMEA).  This is driven, simply, by the space available.  The additional column entries of an aligned FMEA require such tradeoffs in formatting.
• There are diagrams presented in the Handbook without reference or explanation in the text.  The value of these range from near zero to highly negative, as more complex diagrams may be subject to misinterpretation or cause confusion for the reader.  Either would result in a negative impact to learning and/or analysis.  For an example, see “Figure 3.1-1:  Demonstration of the process for narrowing Preparation.”  Preparation can be narrowed??  That sounds very helpful.  Tell me more!  Please?
• Evaluation of the Severity of Effects of Failure is discussed in Step 5 in the Handbook, though the S column is in the Step 4 section of the FMEA form.  “The Third Degree” discusses Severity evaluation in Step 4 to maintain the flow of analysis and to discuss the topic in context.
• Learning the aligned FMEA method and form is iterative – completion of later steps may be needed to fully understand earlier steps of this approach.  Like most endeavors, repetition is required to develop competency.  However, the “double-back” required by postponing Severity evaluations until Step 5 is a different thing altogether.  A new user could easily be distracted by the incomplete step, fruitlessly searching for information to complete Step 4 before continuing to Step 5.  Successful students often display this type of diligence; simply mentioning that it will be addressed in the next step would be sufficient to avoid derailing these readers.
• Recommended changes to action status choices include: 

     1) Leave space blank instead of using “Open” status; it has no meaning.  How can the status of an action be “no action defined?”  A blank space conveys that it has not been defined; if it had been defined, it would be recorded here! 
     2) When status is “Not Implemented,” remove the action information from the form, place a “*” in the Remarks column, and provide justification for rejecting the recommended action in the FMEA Report. 
     3) Use “On Hold” status to identify actions that have been approved, but are not in progress.  Provide the justification for postponing the action in the Remarks (FMEA Report).  Possible reasons for placing an action “On Hold” include resource limitations, pending regulatory update, pending design change, etc.

• The Handbook advocates the use of “4M” work elements – Machine, Man, Material, and EnvironMent.  The other two Ms typically found on an Ishikawa (Fishbone) diagram – Method and Measurement – are also mentioned as optional categories to consider, but “4M Type” remains on the forms.  All potential sources of failure should be considered, whether or not they fit the “M” pattern!
  
• If a performance requirement is adopted from an external specification or standard, and is required to follow any changes in it, do not include the quantitative value on the FMEA form.  Doing so would require the FMEA to be revised each time the value is changed in the standard; simply referencing the “latest revision” would prevent this while maintaining compliance.  If the requirement will not automatically follow any specification updates, reference the applicable revision and/or quantitative value on the FMEA form.  Periodic reviews of the FMEA should be conducted irrespective of this recommendation.
  
• The Handbook lists “unnecessary activity” as an example of process failure.  A waste, perhaps, but a failure?  If a machine operator whistles while s/he works, it is not a process failure; it is not even a waste!  The positive vibe may even improve performance and well-being among the whistler’s coworkers!  There is also mention of Prevention Controls being used to prevent “possible layout deficiencies of the production facility.”  These attempts to inject lean concepts are clumsy and inappropriate.  Beware the inclination to expand “analysis” and intervention into areas where it has no relevance or authority.
  
• Product Special Characteristics are not recorded on the aligned DFMEA form and only during Optimization (Step 6) of the PFMEA.  Waiting until the final stage of analysis to bring attention to critical aspects of the product’s design seems terribly misguided and counterproductive.
• Example form segments are shown throughout the Handbook; Appendix B is dedicated to “Step by Step Hints” using these form segments.  Whenever consulting these, also review the corresponding Standard Form and Errata Sheet to ensure use of the most current information.
• Design and Process FMEAs are developed with an assumption that “all inputs are good;” that is, analysis focuses on that which is within control of the design or process team.  The Handbook states that FMEA-MSR, in contrast, must account for erroneous signals from external components.  The design team can define supplier performance requirements (e.g. 99.5% signal accuracy), but it cannot directly take action to improve the performance of supplied components.  The design team could define a mitigation strategy for the effects of erroneous signals, but this would be a potentially boundless expansion of scope!  Therefore, practitioners must exercise extreme caution when analysis enters this realm.  (See note above on expanding analysis beyond its relevance or authority.)
• Some terminology used to describe FMEA-MSR failure scenarios could be confusing for readers; terms are not consistent with previous usage.  In Section 4.4, on Failure Analysis, timeline diagrams depict the occurrence of a “Fault,” followed by the occurrence of a “Failure Effect,” then a “hazardous event” that may or may not occur.  Throughout the series, only “failures” have been referenced; the use of “fault” is new.  Other terms may be clearer if aligned with usage in “Safety First!  Or is It?” where the Handbook’s “hazardous event” would be labeled an incident.  In the time between the occurrence of the Effect of Failure and the incident, a “hazardous condition,” or simply a hazard, exists.  This term better reflects the risk inherent in the situation and the potential for mitigation.  Also, it seems incongruous to refer to an incident as subsequent to the occurrence of an Effect of Failure.  The incident, or “hazardous event,” is likely to be the most salient Effect of Failure conceivable!
• The Handbook’s presentation of AP for FMEA-MSR (Section 4.5.8) refers readers to previous chapters for discussion of “reducing risk first by S, then F, then, M.”  However, F (Frequency) and M (Monitoring) are not used in the previous chapters and no attempt is made to correlate them to O and D, respectively.  The subsequent descriptions of High, Medium, and Low priorities make no mention of lowering S, though it should always be considered.  MSR even offers two opportunities to lower S – one for the unmitigated Effect of Failure and one for S after MSR.  The latter is even neglected in the Optimization step of FMEA-MSR, though it directly impacts customer experience!
• The Handbook allows DFMEA and PFMEA revision histories to serve as the historical records of product and process development, but requires the S, F, and M ratings of FMEA-MSR to remain unchanged after actions have been taken to reduce them.  This is inconsistent with established practice and could result in an illegible FMEA form as multiple rounds of development are executed with no maintenance of the Optimization section.  Inconsistent, unnecessary, and ill-advised.

 
     The preceding collection of remarks is certainly not comprehensive.  Errors that have not been identified in the Errata Sheet and other confusing or misleading elements that have not been discussed here remain in the Handbook.  These, and other deficiencies, must be navigated at least until the next edition is published.  The more cynical among us might suggest that AIAG’s catalog of onsite and virtual training sessions limits the organization’s incentive to ever produce a clear and coherent document.  But only the cynical.  Experienced practitioners will interpret, adapt, and share their FMEA wisdom as they have done for decades.
     Outside the automotive industry, there is little reason to adopt the new FMEA methodology.  Many of the supporting tools, such as structure trees, can be applied to classical FMEA, if desired.  The investment required to transition to the new approach may be unwarranted, unless the products of interest contain diagnostic systems that would benefit from the additional scrutiny of FMEA-MSR.  An online search yields presentations by both advocates and opponents of the aligned approach.  Readers are encouraged to consider both positions, whether to inform a decision on transition or to best utilize the new method when required.
 
     For additional guidance or assistance with Operations challenges, feel free to leave a comment, contact JayWink Solutions, or schedule an appointment.
 
     For a directory of “FMEA” volumes on “The Third Degree,” see Vol. I:  Introduction to Failure Modes and Effects Analysis.
 
References
[Link] “FMEA Handbook.”  Automotive Industry Action Group and VDA QMC, 2019.

Jody W. Phelps, MSc, PMP®, MBA
Principal Consultant
JayWink Solutions, LLC
[email protected]