Example of a P-Diagram for Process FMEA - Uncontrollable noises

[Sleepy Cat]

If I create PFMEA with 6M, I don't know how to write the following items for "Method".
Step2 Process Step - Process Work Element
Step3 Function of the Process Step / Product Characteristic - Function of the Process Work Element / Process Characteristic
Step4 Failure Mode (FM) of the Process Step - Failure Cause (FC) of the Work Element
For example, is the following correct?
Step2 Mixing - Stirring speed according to the work instruction
Step3 Uniform agitation - the work instruction instructs the procedure to achieve uniform stirring
Step4 Uneven mixing - Incorrect or insufficient instructions, damaged or dirty, not near the operator

 

[Bev D]

Sleepy Cat: Have you had any formal training on PFMEA? Are you working on this alone or with a trained team? I ask because your example is a “hot mess” and I don’t say this to be mean but to be clear that this example is not how a PFMEA is intended to be.

The basic method of FMEA is to:
1. List the functions of the process (not ‘steps’ in a process)
2. List the failure modes of the functions (full failure, partial failure, intermittent failure and/or unexpected failure as appropriate)
3. List the Effects of each failure mode.
4. List the potential causes of each failure mode.
The ‘6Ms’ (Man, Machine, Method, Measurement, etc) is helpful in thinking about causes of failures but it is not required or even appropriate to always list causes in all categories

So, “uneven mixing” is not a step, it is a failure mode of mixing. Now think about what the effect of uneven mixing is. How severe is uneven mixing? Next think about how uneven mixing can occur. This is often a multi step thing (like dominoes in a line falling). For example uneven mixing can occur because the mixing time is too short; how can the mixing time be too short? Then you need to think about how you might prevent, detect or mitigate uneven mixing…

FMEA is not a form you fill out. It is a guide to developing robust processes (and designs).
FMEA is complex and requires substantial training, coaching and experience.

 

[Sleepy Cat]

Bev D: Yes, we have had PFMEA training. And Process Work Elements learned to think in terms of 4M.
However, as noted in Handbook 3.2.3, the number of categories varies from company to company, and additional categories such as Method and Measurement should be acceptable.
Sure, it's not necessary to list in every category all the time, but we decided at certain Process Step we should mention Method.
In the example above, I called it "Mixed" because it contains confidential information. You are right, "uneven mixing" is a failure mode. I wrote it in the example above, but I'll write it again.
Step2 Process Step; Mixing
- Process Work Element; Stirring speed according to the work instruction
Step3 Function of the Process Step; Aagitate powder uniformly
- Function of the Process Work Element; the work instruction instructs the procedure to achieve uniform mixing
Step4 Failure Mode (FM) of the Process Step; Uneven mixing
- Failure Cause (FC) of the Work Element; Incorrect or insufficient instructions, damaged or dirty, not near the operator
If I am wrong, please let me know specifically which step is wrong and how to fix it.

 

[Jimmy123]

Now, after 5 years of the new AIAG&VDA FMEA Handbook, have anyone experience to use the P-Diagram ?
We use it not in this style. We follow the Taguchi style with internal and external noise, or accord.Davis from FORD.
5M makes no sense. This diagram should removed or revised into the book.

 

[Bev D]

The 5-6Ms is a futile attempt to hold on to the fishbone diagram. Ishikawa also used this construct in the process diagrams. It is basically a good thought reminder to think of all causes. In a fishbone diagram it breaks the cause and effect string as most causal mechanisms have a string of different Ms: Mother a true causes a Material weakness and a poor Measurement system cannot detect it and the Machine fractures it and…

 

[JohnsonCM]

Now, after 5 years of the new AIAG&VDA FMEA Handbook, have anyone experience to use the P-Diagram ?
We use it not in this style. We follow the Taguchi style with internal and external noise, or accord.Davis from FORD.
5M makes no sense. This diagram should removed or revised into the book.

Yes. I haved used the P-Diagram successfully many times in product development. The Taguchi factors you refer to for DOE are different "noise" factors. They are controllable factors (you must be able control them for your DOE right?). The P-Diagram Noise Factors are uncontrollable (or at least extremely difficult to control). Think like, for an automobile, the weight of the passenger sitting in the seat. As a designer, it's difficult to "control" the weight of the passenger (large male, small female, baby). Somehow, you need to know what to do with that situation. The P-Diagram helps you identify those factors and consider your options of what you might do to reduce your sensitivity to the design (e.g., change technology, bulk-up the design, etc.).

You referenced FORD. Ford was the source of where I originally learned about the P-Diagram as it was a significant part of their REDPEPr process back in the day.

 

[Miner]

I am no longer in the automotive industry, but do keep abreast of developments. I do use the P-diagram at my current company using the following elements:

• Signal/Inputs (necessary to provide the desired output)
• Control factors (factors that engineers can specify to achieve the desired output)
• Noise factors (factors that are difficult/costly to control)
  • Strength variation noise
    • Piece-to-piece variation (design tolerances and manufacturing variation)
    • Changes over life variation (component deterioration)
    • Internal environment variation (system interactions, component clearances)
  • Stress variation noise
    • Customer usage variation (applications, misuse, etc.)
    • External environment variation (application conditions)

 

[Jimmy123]

I am no longer in the automotive industry, but do keep abreast of developments. I do use the P-diagram at my current company using the following elements:

• Signal/Inputs (necessary to provide the desired output)
• Control factors (factors that engineers can specify to achieve the desired output)
• Noise factors (factors that are difficult/costly to control)
  • Strength variation noise
    • Piece-to-piece variation (design tolerances and manufacturing variation)
    • Changes over life variation (component deterioration)
    • Internal environment variation (system interactions, component clearances)
  • Stress variation noise
    • Customer usage variation (applications, misuse, etc.)
    • External environment variation (application conditions)

100% agree, and as Taguchi, Phadke or Davis recommend, this categories can also be used for design development, process development and production. No need to invent the wheel new with 5+M as new noices as the new AIAG&VDA 2019 suggested.