How to Calculate RPN in Design FMEA (DFMEA) vs. Process FMEA (PFMEA)

[nsraj]

Hi,

In FMEA normally we'll calcualte the RPN as Sev x Occ x Det,

but recently I came to know that in DFMEA the RPN calculation method is slightly difference. i.e.

First Calulate RPN for Failure Mode (RPN FM = Sev (Max) x Occ X Det(FM))

and then RPN for causes,

but in calculating RPN for causes, I was told the formula as

RPN C = Sev (max) x Occ x Det (Min (FM, Causes)

Example

Assuming
for Failure mode RPN FM = 5 X 5X 3 = 125

for cause 1 suppose if i get Sev = 5, Occ=5 and Det = 4,

I normally calculate as 5 x 5 x 4,

but i was told to calculate as 5 x 5 x 3(Det of FM)

is there any reference to this method???


nsr

 

[Valeri]

Re: How to Calculate RPN in Design FEMA (DFMEA) vs. Process FMEA (PFMEA)

Good morning,

I'm not quite sure where you found/heard that calculation for RPN but it should be the same (S)x(O)x(D)=RPN (severity x occurrence x detection = risk priority number) no matter if it's a DFMEA or PFMEA. Then again, maybe I'm not understanding the question

 

[Jim Wynne]

Re: How to Calculate RPN in Design FEMA (DFMEA) vs. Process FMEA (PFMEA)

Hi,

In FMEA normally we'll calcualte the RPN as Sev x Occ x Det,

but recently I came to know that in DFMEA the RPN calculation method is slightly difference. i.e.

First Calulate RPN for Failure Mode (RPN FM = Sev (Max) x Occ X Det(FM))

and then RPN for causes,

but in calculating RPN for causes, I was told the formula as

RPN C = Sev (max) x Occ x Det (Min (FM, Causes)

Example

Assuming
for Failure mode RPN FM = 5 X 5X 3 = 125

for cause 1 suppose if i get Sev = 5, Occ=5 and Det = 4,

I normally calculate as 5 x 5 x 4,

but i was told to calculate as 5 x 5 x 3(Det of FM)

is there any reference to this method???


nsr

I've never seen anything like it, and it seems to be a needless overcomplication of an already dubious practice.

 

[nsraj]

Hi Jim

Thanx for the reply, me too thought the same, but wanna know if realy anything is there like that or not. If anybody come across such info pls pass it on

and Valeri,

let me put it in other way, do we need to calculate RPN for "Failure mode" Separately and RPN for its "causes" separately???

 

[Jim Wynne]

Hi Jim

Thanx for the reply, me too thought the same, but wanna know if realy anything is there like that or not. If anybody come across such info pls pass it on

and Valeri,

let me put it in other way, do we need to calculate RPN for "Failure mode" Separately and RPN for its "causes" separately???

You said that you "came to know" this method--from where? Unless you have a customer that requires it or you can find some utility in it in your own application, I would just stick to the SxOxD method and be done with it.

 

[parallel]

Hi

I have a slightly different doubt about RPN:

RPN=S x O x D

If S is more which is bad, is given maximum rating.
If O is more which is bad, is given maximum rating.
If D is more, which is good (prevents failure from reaching the customer), should we not give minimum rating?

This way priority will have balanced scale.

I know my thought does not tally with our conventional method of calculating RPN.

Please advice.

 

[Chennaiite]

Hi

I have a slightly different doubt about RPN:

RPN=S x O x D

If S is more which is bad, is given maximum rating.
If O is more which is bad, is given maximum rating.
If D is more, which is good (prevents failure from reaching the customer), should we not give minimum rating?

This way priority will have balanced scale.

I know my thought does not tally with our conventional method of calculating RPN.

Please advice.

Without getting into the semantics, please refer whichever rating criteria you have. I am sure it says 'lower the better' for Detection rating as well.

 

[parallel]

Thank you. However my doubt still remains.

 

[Chennaiite]

If D is more, which is good (prevents failure from reaching the customer), should we not give minimum rating?

AIAG FMEA rating criteria does provide minimum rating if the detection control is good. In other words, if D is more, it is not good i.e. the defective part will be undetected and will in all probability reach the Customer. If possible, please share the rating criteria you are referring to, this will help us clarify better.

 

[amit_rd]

Hi

I have a slightly different doubt about RPN:

RPN=S x O x D

If S is more which is bad, is given maximum rating.
If O is more which is bad, is given maximum rating.
If D is more, which is good (prevents failure from reaching the customer), should we not give minimum rating?

This way priority will have balanced scale.

I know my thought does not tally with our conventional method of calculating RPN.

Please advice.

There is no scope for this doubt as what you think is how the rating is to be done. The ratings for severity and occurrence go 1 to 10 whereas for Detection, it is 10 to 1 which mean better the detection control lower the rating. I hope it is clear for you now.