For anything complex, the best method is to start back when you design the product by carrying out a review called a Failure Mode and Effects Analysis, universally abbreviated FMEA.

FMEAS - Reducing the risk of failure
FMEAS - Reducing the risk of failure

Michael Mills | isoTracker Solutions Ltd

Have you ever gone to use something for the first time, and right away it breaks? Most of us have. Then you look where the break happened, and it’s so flimsy that you think, “They should have seen that when they made it.”

Yes, they should have; and yes, you generally can. The better approach is to reduce the risk of failure by using a little forethought. For anything complex, the best method is to start back when you design the product by carrying out a review called a Failure Mode and Effects Analysis, universally abbreviated FMEA.

The basic idea behind an FMEA is this:

  • Look at your product design, and systematically list all the ways it can possibly fail.
  • Pick out the ways that matter.
  • Update your design to eliminate them
  • Now update your FMEA with the new design, to see if you have really eliminated the failures (and if you have introduced any new ones!).
  • Repeat as needed.

Let me explain these steps in more detail.

 

List the failures

To list the ways your product could fail, first you have to know what it is supposed to do. This sounds obvious, but sit down with a team and think of your product as a collection of functions. Then consider that every single function can fail in one of five ways:

  • It might do nothing.
  • It might do the wrong thing.
  • It might do the right thing but in the wrong amount (too little or too much).
  • It might do the right thing at the wrong time (too early, too late, or misfiring somehow).
  • There might be unacceptable side-effects.

So for each function of the product, list all the failures that might occur, along with the causes that could bring them about. Maybe you have an ON/OFF switch, and you list “Switch does nothing” as one possible failure mode: how could that happen? Well, there’s a wire connected to the switch: if that wire comes loose, the switch won’t do anything. So “wire comes loose” is a possible cause that you have to prevent. And you work your way through all the others in the exact same way.

 

Pick the ones that matter

Your complete list of possible failure modes will likely be pretty long. Do you really have to address all of them? Normally not. Next you use the basic principles of risk management to trim down the list to the ones that matter.

Concretely, you assign each possible failure mode a Risk Priority Number (RPN). You calculate this number from three other values that you assign first.

  • Evaluate the probability (P) of the failure happening, typically on a scale from 1-5. 1 means that the failure is extremely unlikely, or virtually impossible. 5 means that the failure is frequent or almost inevitable.
  • Evaluate the severity (S) of the damage caused in case the failure does happen, again from 1-5. 1 means that there is no effect on reliability or safety. 5 means catastrophic failure: the product stops working, or someone might get hurt.
  • Evaluate the detectability (D) of the failure, from 1-5, on the grounds that if customers can tell the product is about to fail, they’ll put it down before anything bad happens. 1 means that you are sure to detect the problem in time. 5 means that it will take you totally unawares.
  • Then your RPN = P x S x D.

Now every possible failure in your list has an RPN between 1 and 125 (= 5 x 5 x 5).

 

Update your design

Next you set a threshold value. And you go back to your design to correct every failure with an RPN greater than your threshold.

What does it mean to “correct” a potential failure? It depends on the failure, and on what’s possible. In the best case, you improve the design so that particular failure is no longer possible. Another approach is to mitigate it somehow, so that the effects are less severe if it ever does happen. (In the case of a safety risk, this might involve shielding to protect the user.)

When you have finished redesigning your product, redo the FMEA analysis to see how it has changed. Start by checking all the failure modes that you set out to address: did you eliminate them? If there are any that remain, did you at least reduce their RPN values to a level below your threshold? So far, so good—but that’s not enough. Since you changed things, there’s always a chance that you introduced new failure potentials which didn’t exist before. Check for these, and evaluate them like the others.

Continue this work iteratively until all your possible failures score below your threshold value.

 

How do you know your threshold is in the right place?

This is a judgement call, and it’s one of the reasons that FMEA evaluations have to be done by a team. All I can suggest is to get input from across the organization—from Service and Manufacturing, say, as well as Design—and to use honest common sense. Most of the time, when you list your failure modes by RPN, it will be obvious that the risks at the top are terrible, and the ones at the bottom are trivial. Often it will be equally obvious where to draw the line between them. There might be a few that you have to discuss because they are close to the line on one side or the other, but usually not many.

 

Different kinds of FMEA

If your operation is large enough, or your work is complex enough, you may find it helpful to split your FMEA into two parts.

First is a Design FMEA (DFMEA). This review is led by the design engineers (while taking account of feedback from other parts of the organization), and it evaluates the product design in the way I’ve just described, to look for places it might fail.

But any complex product is more than just a list of parts. There also has to be a manufacturing procedure that tells how to put them together; in fact, the components may have to be bent, or machined, or welded, or treated in order to reach their final state. This procedure introduces risks of its own: some that could cause the product to fail, and others that could pose a working hazard to the manufacturing personnel. So in cases where the manufacturing procedure is either complex or dangerous, it’s a good idea for that team to conduct their own review—a Process FMEA (PFMEA) that evaluates the manufacturing methods in exactly the same way that the DFMEA evaluated the product design. Again, it is a good practice to include some members on the PFMEA team from outside the production organization. At the very least there should be one or more design engineers present, to facilitate communication back and forth between Design and Production.  

 

How often do you have to do an FMEA?

For the same product? Typically more than once. Do at least a partial FMEA every time you introduce a major change, to check whether the change introduces new risks.

Beyond that, keep track of how the product performs in the field. Then re-open the FMEA once in a while, just to see if the actual performance is in line with your estimates. Chances are that, once the product is in the field, some customer will discover a failure mode you never thought of in advance. Add it to the list and rate it. You might also find that certain failures are more likely—or less likely!—than you anticipated when you did your first evaluation. Or maybe your ratings for severity or detectability were off. In any event, redo the calculations and see where you are. If the new calculations are all well below your threshold, that’s great. If one or more of them starts inching above it, you may have some work to do.    

You have to decide how often “once in a while” is. Choose an interval that makes sense.

The whole point is to reduce the risk of failure. If you can do that, you are doing well.

 

Michael Mills, who writes for isoTracker QMS software, has managed quality systems in large and small companies for close to thirty years. He is a member of ASQ and ISO TC 76 and has audited to ISO 9001 since 1996. He blogs regularly at https://pragmatic-quality.blogspot.com/.

 

The content & opinions in this article are the author’s and do not necessarily represent the views of ManufacturingTomorrow

Comments (0)

This post does not have any comments. Be the first to leave a comment below.


Post A Comment

You must be logged in before you can post a comment. Login now.

Featured Product

KEB VFDs for High Speed Motor Control

KEB VFDs for High Speed Motor Control

KEB's VFDs are the key to unlocking peak performance in your high-speed machinery. Engineered for precision control, even without feedback sensors, our F6 and S6 drives utilize advanced software and high-frequency output. This ensures smooth operation, reduced vibration, and optimal efficiency. Whether you're building turbo blowers, power generation systems, or other high-speed applications, KEB VFDs deliver the reliability and performance you need.