When Failure is the Only Option

When Failure is the Only Option

Failure is not always a bad thing. Sometimes it’s exactly what you want. You might start with one component that’s been well tested with the existing system. It can be a simple clamp or a complicated swing valve (like the one pictured from DHV Industries). You probably tested it out when it was originally implemented and proved you had plenty of safety margin. But now your system has changed. Maybe you’re sending your system into freezing winter temperatures or hot, arid deserts. Maybe you’ve got a more powerful compressor that generates much higher pressures than you’d been dealing with before.

So how do you test that component? Most standards would suggest you take your max operating conditions and increase the magnitude by 50% and test at those conditions. That proof test would verify your equipment can operate safely at your max conditions giving you a 1.5 safety margin. But in some cases, you may want to revise that depending upon your other factors.

Testing for your safety factor is one thing, but you also want to know when your item is going to fail. I’ve seen these margins be anywhere from 2x the max conditions to even more depending on how critical your component is or how repeatable the manufacturing.

It’s also important to keep in mind the other contraints in your system. There’s no reason to test a hose clamp to a higher level of pressure if your hose will fail much sooner. Or if the whole system will break under vibrations that exceed a certain maximum movement then there’s no reason to expect your component to ever exceed that offset. Though you also might have more give in your system that allows the sub-components to undergo more strain than the crucial part of the system would ever. All important things to keep in mind when you’re figuring out how to test and qualify you’re new component.

Just a few things I’ve been thinking about this week as I’m trying to transfer one assembly to another one that undergoes much more severe conditions in the same place. How have you handled this? Do you try to design a new component? Isn’t there also something really fun about burst testing or testing your components to failure?

1 comment

I agree, but in chemical engineering we don’t normally try to break things. We have enough trouble with the “Teardrop of Death.”

Does sound like fun, though.

From a chemical engineering viewpoint, your issue can be stated, “How much fat should I include in the design?” Fifty percent is usually more than I would include, but once, I overdesigned a system by over three hundred percent. I had consulted ‘experts’ who were convinced that one hundred percent of the calculated number was fine. I honestly don’t know why I made it so big. Three times what the experts said. I just had a feeling that I should put in all I could get away with.

I received rave reviews for my design that, when tested in the field, had a thirty percent over design. It turned out that there were factors nobody, even me, took into account.

You can speculate about female intuition all you want, but engineering intuition is a fact.

Comments are closed.