Do I need to use the tables in the FMEA reference manuals to determine risk? What about RPN limits?

We currently use a 1-6 scale for determining risk. Do we need to use a 1-10 scale?

Assuming we use a 1-10 scale, is there a commonly accepted RPN limit or are we able to determine that on our own?

Paragon468 said:

We currently use a 1-6 scale for determining risk. Do we need to use a 1-10 scale?

Assuming we use a 1-10 scale, is there a commonly accepted RPN limit or are we able to determine that on our own?

Click to expand...

Others who are more knowledgeable about automotive requirements can respond to the automotive requirements.
From a scientific and engineering standpoint you can use any scale you want. I used a 1-5 scale as it worked for my organization. I would have preferred a 1-3 scale, but my organization wasn’t quite ready for it. In my experience and knowledge of mathematics the more rating levels one allows the more debate about the finer gradations. Definitions for each level become more nuanced and only provide an allusion of additional precision. After all too many people only guess at occurrence and detection instead of relying on actual values determined thru appropriate development studies and MSAs. The debate only delays action to mitigate the occurrence rate or improve detection.

Since RPN is mathematically flawed and manipulated by guessing and ‘consensus’ for occurence and detection there really is no scientifically valid RPN limit. Indeed the mathematical flaw ( you cannot add, subtract, divide or multiply ordinal data based on subjective assessment) allows for similar RPNs that are quite different in their actual risk profile. (Assuming the occurrence and detection are not just wild a$$ guesses)

Dr. Wheeler explains in detail.....

In the past, I developed my own scale (all 3 features) which was tailored specifically for our manufacturing processes and products. Although it worked really well for US, our customers did not like it, and instructed us to use the AIAG tables.

As for a 'target' RPN value, that would defeat the whole purpose. I have had customers in the past who have stated 'RPN cannot exceed 100'. If I have a safety characteristic that has a SEV of 10, that severely limits me in my OCC and DET scoring. This can (will) result in downgrading the SEV score, so that the RPN is less than 100. Its basically just data manipulation to meet a stated requirement, without the person stating the requirement really understanding what the RPN number is for.

If you are using the risk analysis info internally, follow the sdvice above.

If you are providing this for a customer, your OEM customers will dictate what to use to meet their requirements. If you are Tier 2 or lower seek direction from your customer.

*If the customer is automotive customer ;there’s 3 guidance tables for DFMEA & others for PFMEA ;


** RPN is not only sufficient for specifying the risks ; the Severity could be reach to S =10 and occurrence could be =1 , the detection =1 ; --hence the RPN =only 10 with a deadly severity --- hence the specifying also the limits of SOD --- with the RPN is more useful

***for automotive customer' customer specific requirements should be followed with the referenced tables in AIAG & VDA FMEA manual; FMEA & control plan may need informing customer before releasing ,example of FORD CSR is attached

As an auditor within the automotive industry, I only have a short amount of time to verify the existence and implementation of the Failure Mode and Effects Analysis (FMEA) process at supplier sites during audits. This entails checking if they utilize their own grading system (whether it's a 3, 5, or even 10-point scale) rather than following the official standards. My subsequent inquiry always revolves around requesting to see the FMEA correlation table aligned with the AIAG&VDA handbook - never saw proof for that corelation performed. Any deviation from this standard constitutes a nonconformity, specifically the inadequacy of the FMEA process due to a lack of alignment between risk assessment and the prescribed standards tables.

PS why you try to force the door if they are open ? Use standard table!

Pawel Zelazowski said:

As an auditor within the automotive industry, I only have a short amount of time to verify the existence and implementation of the Failure Mode and Effects Analysis (FMEA) process at supplier sites during audits. This entails checking if they utilize their own grading system (whether it's a 3, 5, or even 10-point scale) rather than following the official standards. My subsequent inquiry always revolves around requesting to see the FMEA correlation table aligned with the AIAG&VDA handbook - never saw proof for that corelation performed. Any deviation from this standard constitutes a nonconformity, specifically the inadequacy of the FMEA process due to a lack of alignment between risk assessment and the prescribed standards tables.

PS why you try to force the door if they are open ? Use standard table!

Click to expand...

Really? No company should make their system convenient for the auditor. Who cares how long it takes it you? If you really understand how to perform risk assessment/FMEA it won’t take long to see correlation.

Bev D said:

Really? No company should make their system convenient for the auditor. Who cares how long it takes it you? If you really understand how to perform risk assessment/FMEA it won’t take long to see correlation.

Click to expand...

You are correct. However, there is another topic: my SQG and requirements follow either the AIAG FMEA or VDA approaches for projects before 2019, and the VDA & AIAG book for all new projects. Should I still attempt to validate the supplier risk assessment process in accordance with the official book?