When the severity rating will change?. In my understanding, it will change only there is a change in design. Will the process design change can also change the rating? Example, implementation of error proofing gauge will change the severity? Please help

The Severity Rating can only be changed by a change in the product design.

A Mistake Proof can improve the Detection Rating if it is the type that alarms when a defect occurrs. It may also improve the Occurrence Rating if it is the type that prevents defects from being built.

When the severity rating will change?. In my understanding, it will change only there is a change in design. Will the process design change can also change the rating? Example, implementation of error proofing gauge will change the severity? Please help
An "error-proofing gage" does nothing to change the Severity rating. It may/will affect the Detection rating only. If you have error-proofing designed into the process so that the Failure Mode can't be produced, you will affect only the Occurence rating.
Severity rating can only be changed by a design change that mitigates the Effect of the Failure Mode (such as a redundancy setup for the final system).

Hope that helps.

In general , the severity rating for an item in the PFMEA must minimally be at the same level as that in the design FMEA.

The caveat is that the PFMEA rating CAN be higher for severity. Especially when an operator can be harmed by manufacturing tooling or practices.

This increased level in the PFMEA (process) does not find its way back to the design FMEA (product). It should be addressed thru a process design change,in which case it will revert back to the same level as found in the design FMEA.

The only way to reduce severity in a design FMEA IS thru a design (product) change.

I agree with Bills comments on Error proofing.

Severity rating can only be changed by a design change that mitigates the Effect of the Failure Mode (such as a redundancy setup for the final system).
I thought some about this last night. Even a redundancy in the system would not affect the Severity. It would certainly (hopefully:rolleyes: ) help reduce the Occurence.

Most of the changes we make to processes as a result of FMEA will result in lower detection and/or occurrence scores. It's rare that the changes we make reduce the severity.

I'm sure someone will post examples - I'm not saying you CAN'T reduce severity, just that (by its very nature) it's the most difficult of the S/O/D values to reduce.

A lot of people make the mistake of thinking that because they have reduced the chances of something happening that they have reduced the severity - this is not the case. Occurrence is "how often will this thing happen". Severity is "when this thing happens, how serious is it?".

For an FMEA to be effective in prioritising risk you have to consider the three values independently. The overall combined effect of the three is your RPN number.

All,
Thanks for all your replies. In the manual it is given that "A reduction in severity ranking index can be effected through a design change to system, sub system or component, or a redesign of the process". Am I interpreting wrongly?

All,
Thanks for all your replies. In the manual it is given that "A reduction in severity ranking index can be effected through a design change to system, sub system or component, or a redesign of the process". Am I interpreting wrongly?

That is a very misleading statement, as I tend to agree with everyone above. I read the severity rating in a Process FMEA as:

"How Serious is the Effect". Looking at a standard FMEA format, the effects column comes before any actions you take, so how can the sevrity rating be affected by a change in the process. The effect will always be the same no matter what you do to the process. The only way you can change the severity of an effect is to change the design, so that the effect is mitigated To use the example given in the FMEA manual,

If an effect of a failure is a tyre blow out, in order to reduce the serverity of this effect, you can design in features that allow the tyre to run flat and therefore this reduces the severity of the blow out!

I think you will find that the Severity rating for failures related to operator safety can be reduced using a redesign of the process. So for example, if you have a drill and the failure mode is that the operator could put their hand into the dangerous area, the severity would be high for operator injury. However, if you introduce a device that cuts the drill off if their hand enters the the dangerous area, you are reducing the severity of the effect of putting the hand in the area!!!!! :confused: You follow?

I think this is might be what it means, relating to operator safety and comfort issues rather than customer satisfaction/end product issues!

Also as an addtion to the above, you could also relate this to environmental issues. If you have a process that impacts heavily on the environment, then the severity of the failure will be high. However, take steps to reduce the environmental impact (design of the process), then you can reduce the severity.

This is more in line with EFMEA's (Environmental FMEA's) but the widespread adoption of ISO 14001 would also benefit from Environmental issues being considered in all FMEA's!

Hope this helps give you an idea!

As with all general "rules of thumb," it should not be taken as a absolute. Perhaps there are cases where a process design change would reduce Severity, but not likely. The Severity rating is generally looking at the product which results from the process. If anything, a change in the process is likely to reduce the Occurance, or perhaps Detection.

I agree with Bill's comments on error proofing, and I also agree with Michael's comments on severity rating. The severity rating should not be any lower than the severity rating coming from the Design FMEA. The severity rating can be higher if the Process endangers the operator, or has a worse effect on a down stream process. In this specific case it appears that the Sowmya is attempting to reduce the severity by error proofing a gage. My take on this is that the severity should not be reduced in this case. Just because I have a better gage doesn't mean that the effect of a failure has changed. The word gage leads me to believe the cause or failure mode has all ready occurred. So a change in the detection rating is more appropriate.

Bill, I am interested in hearing you thoughts on why you think redundancy only affects the occurrence rating. It is my experience that redundancy would reduce the severity because we only consider one failure mode at a time. If we consider both the primary and redundant system/component failures occurring then we are considering multiple failure modes. When people start considering multiple failure modes then even trivial failures can have severity rankings driven to 10.

Severity Rating can be reduced in most of the cases with a Product design change. Not Restricted to Product Design but also a Process Design.

Going by the definition of Severity 'Which can endanger an operator' during the process if some sort of process change is done to eliminate the danger to the operator then the severity rating can be reduced.

For Example

Where welding operation is carried out , there are possibilities of Flames and fumes coming out , If an automation is done by using Robots which is remote controlled then the severity to the operator is out of the issue. in this case the change to the process reduces the severity.

Thanks for putting a query for which i know the answer

Balaji

Tymer and Bill,

Redundancy affects the occurance in a negative fashion (increases the liklihood of failure). Ref the "warranty" aspects of Reliability.

It affects the "severity" in a positive fashion (reduces it). Ref the notion of "mission" reliability.

Thanks to all. Process change confused me. Now I am clear.

Tymer and Bill,

Redundancy affects the occurance in a negative fashion (increases the liklihood of failure). Ref the "warranty" aspects of Reliability.

It affects the "severity" in a positive fashion (reduces it). Ref the notion of "mission" reliability.

Michael, Great comment. :agree1: However I would add that it depends on what cause you are considering. If the cause of the system failure is a part like a sensor and you add a redundant sensor then yes the occurrence doubles as well. But if the cause is a root cause like contamination and you add a redundant fuel filter and path, then the severity of the primary path being plugged is reduced. In this case the occurrence rating would not increase because it is associated with the particle count of the contaminant not how often it triggers the failure.

Bill, I am interested in hearing you thoughts on why you think redundancy only affects the occurrence rating. It is my experience that redundancy would reduce the severity because we only consider one failure mode at a time. If we consider both the primary and redundant system/component failures occurring then we are considering multiple failure modes. When people start considering multiple failure modes then even trivial failures can have severity rankings driven to 10.
I've thought about this and, coupled with Michael's comments, I think I've decided that I need to think about it some more :bonk:

Let's see if I understand somewhat. On the shuttle (with its back up/redundant computers), the Severity of the main computer would not be as high as the Severity of the 4th, correct? I do understand Michael's "Mission Statement" concept but don't I think I see why the Severity isn't the same for each computer at the PFMEA level. I do see at the SFMEA level where the failure of the system would be addressed with redundancy to address a high severity, but does that mean the DFMEA would have a reduced severity (which might be OK) and then that ranking cascades into the PFMEA? If that's how it works, why wouldn't each computer failing have an equal severity ranking? (Am I making any sense:confused: )

"FMEA Severity Rating - Will a Process Design Change also Change the Severity Rating?"

If you add redundancy here yes.

I've thought about this and, coupled with Michael's comments, I think I've decided that I need to think about it some more :bonk:

Let's see if I understand somewhat. On the shuttle (with its back up/redundant computers), the Severity of the main computer would not be as high as the Severity of the 4th, correct? I do understand Michael's "Mission Statement" concept but don't I think I see why the Severity isn't the same for each computer at the PFMEA level. I do see at the SFMEA level where the failure of the system would be addressed with redundancy to address a high severity, but does that mean the DFMEA would have a reduced severity (which might be OK) and then that ranking cascades into the PFMEA? If that's how it works, why wouldn't each computer failing have an equal severity ranking? (Am I making any sense:confused: )

Bill & Michael,
When I am working on an FMEA and a confusing scenario like this is presented, I keep two rules in mind. 1) We only consider one failure mode at a time, and 2) We stay at the same level (system vs sub-system vs component.)

In Bill's example above the level we are considering is the computer systems level, where multiple computers exists. We are considering the functions the system needs to meet. If only one computer existed and there was a failure then the Severity may be 10. If there is a redundant computer then the Severity may be 8. The reason we are allowed to reduce the severity is because we are assuming only one failure mode at a time. That is only one computer fails in a computer system. The redundancy means the computers fulfill the same function and are independent of one another.

When we do the sub-system FMEA (the computer in this case) then we are considering the functions the computer is required to meet. If it fails to meet these functions then it is evaluated by the failure to meet this function. In this case it is a severity of 10 because it is the same function as the system but we are evaluating only one computer.

Now when you do the process FMEA on a single computer you will cascade the single computer severity into the process FMEA as was discussed above.

For evaluating multiple Failure Modes at once you should turn to another tool such as Fault Tree Analysis.

I hope this helps.

Regards.

I have reduced the severity of a failure only with design changes. essentially, you must take an action that counteracts what happens when the failure occurs. something in the product must detect the failure and then counteract it.

however, these actions are not always 100% effective so the result is that you reduce the serity of most occurences...some woudl then say you have only reduced the probability of occurence of the severe failure...

a generic example is putting training wheels on your kid's bike. you can't stop them from losing their balance (failure mode), but the training wheels will stop them from hitting the ground (effect), most(?) of the time

Counteract and / or provide warning.

I agree that they may not always be effective. In many cases we forget
the human element of the equation.

If the item is being operated in an unsafe manner / misused / not reasonably operated as specified by the manual or society / ignored, you lose the effectiveness of the system.

The FMEA must also highlight this issue as a counterpoint to the upgrade you have made.

The difference is that in one case you own the liability and in the other,the user does.

some woudl then say you have only reduced the probability of occurence of the severe failure...

a generic example is putting training wheels on your kid's bike. you can't stop them from losing their balance (failure mode), but the training wheels will stop them from hitting the ground (effect), most(?) of the time

I would disagree with saying you have reduced the probability of a severe failure. Your example clearly states that you cannot reduce the occurence of the failure, but your design change has indeed reduced the severity of that failure occuring (the child won't hit the ground)! I am not sure if your post is supporting the argument that these design changes reduce severity of failure or arguing the point that these actions only serve to reduce the occurence rating? So I apologise if I have mis-understood you.

Bill & Michael,
When I am working on an FMEA and a confusing scenario like this is presented, I keep two rules in mind. 1) We only consider one failure mode at a time, and 2) We stay at the same level (system vs sub-system vs component.)

In Bill's example above the level we are considering is the computer systems level, where multiple computers exists. We are considering the functions the system needs to meet. If only one computer existed and there was a failure then the Severity may be 10. If there is a redundant computer then the Severity may be 8. The reason we are allowed to reduce the severity is because we are assuming only one failure mode at a time. That is only one computer fails in a computer system. The redundancy means the computers fulfill the same function and are independent of one another.

When we do the sub-system FMEA (the computer in this case) then we are considering the functions the computer is required to meet. If it fails to meet these functions then it is evaluated by the failure to meet this function. In this case it is a severity of 10 because it is the same function as the system but we are evaluating only one computer.

Now when you do the process FMEA on a single computer you will cascade the single computer severity into the process FMEA as was discussed above.

For evaluating multiple Failure Modes at once you should turn to another tool such as Fault Tree Analysis.

I hope this helps.

Regards.
Thanks Tymer. Your explanation makes some sense but I'm not sure I've fully grasped it yet. I'll think about it a little more and be back.