I have taught and implemented FMEAs for over 30 years.  I have applied them to well over a thousand products from a wide variety of industries.  I have seen their proper use improve the design and manufacture of a wide variety of products including molecular assays for the detection of various diseases including HIV, late stage cancer treatments, hybrid vehicles, flat screen TVs and just about every automotive component you can think of.  Standard AIAG Design and Process forms were used that included a class column.  Let me begin by saying that the forms are fine, it is what we are putting in the forms that is the problem.  If you know what goes in each column and properly calculate and use the Class Column to determine the issues you must work on instead of RPN, you will be well on your way to improving your products and processes.

The paper spends considerable time talking about an FMEA.  There is no such thing as an “FMEA”.  There are multiple types of FMEAs all with very different purposes.  Design FMEAs are used to manage risk due to the design’s current design specifications.   Improper design specifications are one of the major sources of risk that companies who design products face.

The Process FMEA is used to manage risk due to the current process.  There are many sources of risks that Process FMEAs can be used to manage.  Most Process FMEAs are done on manufacturing processes due to the demands of industry for their use.  Exposure to harm occurs in these processes when an out of spec condition such as scratches is created by the process.

Process FMEAs can be used to manage risk due to non-manufacturing processes as well.  Any time a process fails to create a desired output or creates an undesirable output, exposure to harm occurs.

Machinery FMEAs, when used correctly (99.9% are not), is an excellent to tool for managing risk due to improper manufacturing process design.  Since the focus of the Machinery FMEA is the design of a process, the majority of people create a Machinery FMEA that looks more like a Process FMEA which negates the value of the tool.

The following example is provided by the author as an improper FMEA:

Potential failure mode: Part scratched Potential effect(s) of failure: Customer rejects part Severity: 5 Cause(s)/mechanism(s) of failure: Operator did not detect the scratch Recommended actions: None (because the occurrence was rated at 5 and detections at a 4, which makes the total risk priority number = 100). Typically, organizations do not apply a risk treatment to RPN values at or under 100.

Contrary to what the article states, the problem with the example is not the ”Part Scratched” failure mode.  This is actually a legitimate failure mode assuming scratches are an out of spec condition.  The three major problems with the example is the cause listed and the fact that the author left out two critical columns of the Process FMEA.  “Operator did not detect the scratch” is not a failure cause.  There are 10 different types of root causes that all processes should be examined for.  One of those is mishandling of parts.  In the example provided by the author, operators can create scratches by stacking parts on the conveyor.    Consequently, our failure cause is “Operator stacks parts on conveyor”.   The proper definition of the root causes of process failure is the most important step of the Process FMEA process since it allows us to define the required “Prevention Controls”.  The Prevention control column describes how the failure causes (sources of incidents that lead to exposure to harm) will be prevented.  This column contains the primary risk controls for the process.

The next step is to fill out the “Prevention Controls” column.  When the Failure Cause column is properly filled out, the majority of the required Prevention Controls become obvious.  One simply must ask what control must be implemented to prevent the cause.  In the example provided, one may decide to implement procedures on product handling.  While implementing Process FMEAs as a Plant Manager for Ford Motor Company which resulted in average weekly cost savings of $144,000 after nine months of implementation, we implemented over 15,000 controls many of which were simple and cost very little money.

The last column that must be filled out that the author left out of the paper is the “Detection Controls” column.  This column is used to capture the checks that will be made to contain unacceptable process outputs or failure modes.  In a properly constructed Process FMEA your Prevention Controls often outnumber the quantity of Detection controls by a factor of 10 to 1.

The author proposes the following as a proper FMEA:

Potential failure mode/Potential process function failure: Operators are unable to utilize a one-piece flow to package parts Potential effect(s) of failure/Potential effects of the process function failure: Parts are removed and stacked on a table, which causes appearance defects such as scratches, scuffs, dents, and chips. Customer rejects Severity: 5 Cause(s)/mechanism(s) of failure/Root cause (why would the process function failure occur?): One-piece flow not designed into or utilized at the process work cell.

It is difficult to tell what kind of FMEA the author intends to do.  If one were creating a Machinery FMEA, the requirement that the “Operator must be able to utilize a one-piece flow to package parts” would be a legitimate entry for the Requirements column of the Machinery FMEA.  The failure mode and effects provided by the author would then be legitimate.  The example falls apart at the cause entry.  In a properly completed Machinery FMEA a detailed description of the equipment specified to accomplish the one-piece flow would be listed.  A Design Controls column would also be included to define the methods used to ensure that the one-piece flow could be supported.

It is also surprising the Prevention and Detection Controls column are left out of the example of a "good" FMEA provided by the author.  The Recommended Actions column is used to capture actions that will be attempted to reduce risk.  In real life, some will be work and be implmented and some will not.  This column is not intended to contain your risk prevention and mitigation controls.  That is the job of the Prevention and Detection columns.

It is imporant to know how the various FMEA types are used to manage risk.  As an example, let us look at the Process FMEA.  The Failure Mode Column of the Process FMEA is the objectionable incident that leads to exposure to harm.  The Effects Column contains a description of the harm.  The Severity Column is a measure of the level of harm.  The Failure Cause Column is the source of the objectionable incident.  The Occurrence Column is a measure of the probability of the objectionable incident and subsequent exposure to harm due to the source capture in the Failure Cause column.  The Class Column, determined from the Severity and Occurrence Columns defines the level of risk for the Process FMEA line being evaluated.  It tells you what the lines that have objectionable risk and what must be worked on.  The Prevention and Detection columns are used to capture the risk prevention and detection controls.  The Recommended Actions area is used to capture actions that will be taken in an attempt to reduce risk.  Some will be implemented and some will not.  If implemented, the associated columns of the Process FMEA must be updated.

In closing, Machinery, Design and Process FMEAs are powerful examples of the “Risk Based Thinking” required by ISO 9001:2015 when understood and implemented correctly.  New Machinery, Design and Process FMEA forms are not needed.  Knowledge on how to properly use them is. Tremendous benefits are available to you when these tools are used correclty.

If you would like to learn more about FMEAs and/or how to use FMEAs to help you comply with ISO 9001:2015 visit www.harpcosystems.com or contact me at Richard.harpster@harpcosystems.com.