When to take action if an FMEA RPN number is high?

SQ_Joe

Involved In Discussions
#31
I am not all that passionate about how many categories go into a severity rating. Certainly there isn't a single perfect or optimal number that fits everyone's needs. for me it's not whether or not it's easy to use 10 different categories, it's more about the value that more categories will give us. does it actually help us get to preventing and mitigating Problems? Will a 9 drive a different action than a 10?

Unlike Severity I am passionate about the whole occurrence rating. SXO is still bad math...
S times O is bad math. A S - O, Criticality Matrix is an excellent way to categorize risk.
Sometimes ranking systems with ten Severity rankings may drive different Recommended Actions than a system with nine Severity rankings.

The number of S, O, and D rankings is determined by the higher level customer FMEA. If the rankings are not the same it is more difficult to evaluate the risk of the higher level FMEA be it a simple system, sub-system or system level.
 
Elsmar Forum Sponsor

Bev D

Heretical Statistician
Staff member
Super Moderator
#32
I understand your response, but I would challenge the dogma of FMEA further.
If you really think about the occurence and detection ratings their determination is not just bad math, it isn't math at all but opinion.
 
Last edited:

SQ_Joe

Involved In Discussions
#33
I understand your response, but I would challenge the dogma of FMEA further.
If you really think about the severity and detection ratings their determination is not just bad math, it isn't math at all but opinion.
It is not dogma. It iis a way to look at risk from many views. If you have poor detection, you cannot cannot have low occurrence because of the uncertainty of the measurement system. If you have very low severity, your time can be used on bigger "fish" issues.

Unfortunately, due to the poor listening, discussions and advice on FMEAs, the AIAG and VDA are publishing a no change, much more difficult to understand, set of FMEA requirements for Automotive suppliers. Good luck now. (They advise a master facilitator.)

Non-automotive suppliers may do as they wish. But a true understanding of the QFD Houses, the V-model, and FEMAs would make project management/APQP with its PPAP a walk in the park. Not to mention the technical memory database development.

A blended AIAG Ver. 4 and AIAG-VDA Rel. 1 FMEA forms can be obtained through Systems2win. (See https://www.systems2win.com/solutions/FMEA.htm) It is Excel based, has AP instead of RPN (yeah), data validation, and multi-language capability. Data/information can easily associated across many forms. Well worth a look and the money.
 

Jim Wynne

Staff member
Admin
#34
If you have poor detection, you cannot cannot have low occurrence because of the uncertainty of the measurement system.
I'm not sure what "poor detection" means. A relatively high detection number isn't necessarily a value judgment. It's an observation of a state of being. It's at least theoretically possible to have a very low occurrence rating and a relatively high detection value because we can develop the occurrence number based on history. If process control has been developed conscientiously, the idea is to move away from detection-based controls.
 

John Broomfield

Staff member
Super Moderator
#35
If you have poor detection, you cannot cannot have low occurrence because of the uncertainty of the measurement system.
Poor detectability is not poor detection.

Detecting leaks from an underground fuel tank would be an example of poor detectability.

When D is used it stands for Detectability not Detection.
 

SQ_Joe

Involved In Discussions
#36
I'm not sure what "poor detection" means. A relatively high detection number isn't necessarily a value judgment. It's an observation of a state of being. It's at least theoretically possible to have a very low occurrence rating and a relatively high detection value because we can develop the occurrence number based on history. If process control has been developed conscientiously, the idea is to move away from detection-based controls.
Jim, How do you know you have a "occurrence number based on history," if you have not measured it via some method. Why are some parts not meeting customer expectation? If it was shipped, why was it not detected. If not shipped, by what method was it detected not to meet customer values. Prevention methods and detection methods are different from Detection methods to detect failure. History is the measurement of detection methods. Prevention is just that and is "measured" via other detection methods. At some point every process is measured via a detection method.
Some just have much longer intervals of measurement.
 
Last edited:

Jim Wynne

Staff member
Admin
#37
Jim, How do you know you have a "occurrence number based on history," if you have not measured it via some method. Why are some parts not meeting customer expectation? If it was shipped, why was it not detected. If not shipped, by what method was it detected not to meet customer values.
The statement that I responded to was, If you have poor detection, you cannot cannot have low occurrence because of the uncertainty of the measurement system. By this I assume that you mean that if I have an issue with detecting a defective unit, the occurrence RPN value can't be low because of the difficulty of detection. My response to that was that it's possible to have a relatively high detection factor (relatively little chance of detection) and low occurrence rate if I can base the occurrence rate on history of occurrence. Imagine a high-speed process that's producing a very large number of parts in a relatively short period of time. My measurement system is capable, but the sheer number of parts (and perhaps customer need) prohibits sampling to the extent that would be necessary to have a statistically valid characterization of a particular lot. Process validation has been done, however, and the relevant process controls have been invoked. I know from history that there have been few or even no escapes reported by the customer. I can't depend on detection, thus the detection RPN factor is high, but I know from experience that there have been very few (if any) defects, so the occurrence RPN factor is low.

Prevention methods and detection methods are different from Detection methods to detect failure. History is the measurement of detection methods.
I'm sorry, but this makes no sense to me.
 

SQ_Joe

Involved In Discussions
#38
The statement that I responded to was, If you have poor detection, you cannot cannot have low occurrence because of the uncertainty of the measurement system. By this I assume that you mean that if I have an issue with detecting a defective unit, the occurrence RPN value can't be low because of the difficulty of detection. My response to that was that it's possible to have a relatively high detection factor (relatively little chance of detection) and low occurrence rate if I can base the occurrence rate on history of occurrence. Imagine a high-speed process that's producing a very large number of parts in a relatively short period of time. My measurement system is capable, but the sheer number of parts (and perhaps customer need) prohibits sampling to the extent that would be necessary to have a statistically valid characterization of a particular lot. Process validation has been done, however, and the relevant process controls have been invoked. I know from history that there have been few or even no escapes reported by the customer. I can't depend on detection, thus the detection RPN factor is high, but I know from experience that there have been very few (if any) defects, so the occurrence RPN factor is low.


I'm sorry, but this makes no sense to me.
I agree with your statements as described. However, my position would be that you have a detection level would be whatever the design validation and process verification values would suggest. This value would support a good DETECTION value on the FMEA. (Now, see below.) The Frequency in this senerio should be documented on the PFMEA for customer concurrence as documented in the Control Plan..

The Control Plan would indicate how many and when samples for verification of expected process shift variation will occur. If the level of DETECTABILITY is below what the customer is comfortable with the issue is addressed via the control plan and the new frequency number feed upwardly to the PFMEA. The ability to measure which drives detection has not changed. Detectability is not satisfactory. This difference should be documented in the action plans independent of AP level. (Just another reason for AP and not RPN.)

This is why a submission always has a FMEA and control plan. The comany's "secret" to fast producction can be addressed via a black box, but the senerio you list above will not become an issue if a large spill occurres. The customer will not be happy, but they will understand. It is best for every product to have submitted FMEAs in their PPAP submissions as stated in another thread..
 

SQ_Joe

Involved In Discussions
#40
Are you referring to a PPAP submission to an automotive customer? Please clarify.
Any company with a quality department.

Would you purchase from a company who told you nothing about how they do things? (correct quantity, receiving material, line rate control, inspection intervals, shipping, material identification, ....

Only a company wishing to go out of business would sign a "no knowledge" contract. There are methods to keep confidentiality within desired documentation. If materials are secret, then refer to material specification which customers do not get, same with process specs, test specs, etc. They show due diligence and forethought. No documentation is not wise.

Ask Deming, anyone on the Material Review Board during WWII, Founders of ASQ, any Japanese company, ...
 
Thread starter Similar threads Forum Replies Date
M Requiring action on factors alone of RPN in FMEA APQP and PPAP 7
C How to handle Action Results in a new FMEA Review once we have a new RPN? FMEA and Control Plans 1
E FMEA Action Plan Threshold (RPN) - Auditor says Action Plan for an RPN > 84 IATF 16949 - Automotive Quality Systems Standard 21
T RPN of Action Results per the AIAG FMEA manual FMEA and Control Plans 5
J FMEA RPN Number and When to Recommend to Take Action FMEA and Control Plans 3
S Process FMEA (PFMEA) RPN Corrective Action Plan FMEA and Control Plans 6
S FMEA RPN Corrective Action Plan QS-9000 - American Automotive Manufacturers Standard 1
Ashok sunder Is it possible to reduce FMEA Occurrence and Detection Ranking after corrective action taken for customer complaints? FMEA and Control Plans 6
M FMEA Corrective Action target date changed ISO 14971 - Medical Device Risk Management 1
P FMEA Risk Analysis Recommended Action Priority FMEA and Control Plans 2
V Linking the FMEA Causes & Action Plan with DOE Experiments or Experimentation Design FMEA and Control Plans 1
H Corrective Action: When to update the Process FMEA (PFMEA) FMEA and Control Plans 15
T Recommended Action(s) Column - Process FMEA FMEA and Control Plans 17
R Updated Process FMEA: How much detail? Closure of an External Corrective Action FMEA and Control Plans 3
A Corrective Action vs. FMEA Occurance, Detection, Severance Nonconformance and Corrective Action 5
Raffy Preventive Action - Does generating an FMEA help future containment? Preventive Action and Continuous Improvement 2
R Not every INCIDENT report will lead to a corrective action EU Medical Device Regulations 6
V ISO 10360-5: 2020 Gap analysis and Action plan Excel .xls Spreadsheet Templates and Tools 1
J Action Not defined in procedure - NC - Incoming inspection AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 1
L Separate Corrective Action and Preventive Action Procedures ISO 13485:2016 - Medical Device Quality Management Systems 13
T AS9100D Clause 10.2.1g Supplier Corrective Action for each and every nonconformity? AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 3
R Non conformance (NC) or Corrective & Preventive action (CAPA) CE Marking (Conformité Européene) / CB Scheme 7
I Setting action completion dates Nonconformance and Corrective Action 3
A Medical device Mode Of Action CE Marking (Conformité Européene) / CB Scheme 2
D Corrective Action Timing - IATF CB Deadline IATF 16949 - Automotive Quality Systems Standard 2
R AS 9100D - Containment and Corrective action AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 3
P Customer Corrective Action Requests in OASIS? AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 4
C Corrective action for failure in documents control ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 10
nadhar2 Classification of Action Items Misc. Quality Assurance and Business Systems Related Topics 3
C If it doesn't prevent a non-conformance, is it a preventive action? IATF 16949 - Automotive Quality Systems Standard 13
G Assignable cause/corrective action list for SPC Software Statistical Analysis Tools, Techniques and SPC 3
Q Tips for Action Requests on Communication and Leadership ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 1
S Can we provide training plan as corrective action for IATF 16949 Non conformity? IATF 16949 - Automotive Quality Systems Standard 9
P Corrective Action Response for Missed bumps on brake press Manufacturing and Related Processes 2
R Evaluating the need for preventive action Preventive Action and Continuous Improvement 3
B Stakeholder Initiated Corrective and Preventive Action Misc. Quality Assurance and Business Systems Related Topics 5
B Corrective and Preventive Action (CAPA) A Key Process of the Quality Management System Dec 17... Training - Internal, External, Online and Distance Learning 0
I When exactly can a corrective action be closed? Nonconformance and Corrective Action 3
M Informational TGA – Australian regulatory action on breast implants and breast tissue expanders Medical Device and FDA Regulations and Standards News 0
M Informational Medical Device Safety Action Plan: Protecting Patients, Promoting Public Health Medical Device and FDA Regulations and Standards News 0
Juby Tan Action on Customer Replacement (Nonconforming Product - ISO 9001) Customer Complaints 2
qualprod Add new action plans in CA, while waiting effectiveness - Same problem reappears ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 13
M Informational The FDA Takes Action to Protect Patients from Risk of Certain Textured Breast Implants; Requests Allergan Voluntarily Recall Certain Breast Implants a Medical Device and FDA Regulations and Standards News 0
L What to do if a Corrective Action requires several extensions Nonconformance and Corrective Action 7
qualprod Best criteria to measure Corrective Action effectiveness - Poor Maintenance ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 10
Ashok sunder Is it possible to reduce Risk likelihood and impact Post control Ranking after corrective action taken for risk? FMEA and Control Plans 1
E CAR (Corrective Action Report) with questionable Root Cause ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 14
B How to reply NCR on ineffectiveness of corrective action during IATF external audit? This is repeated issue whereby some mistake was done. IATF 16949 - Automotive Quality Systems Standard 7
C Detection Action for Failure Effects - AIAG 4th Edition Layout FMEA and Control Plans 5
I AS9100D Non-conformance for 10.2.1 b 2 and 10.2.1. b 3 e and f - Corrective Action AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 10

Similar threads

Top Bottom