Ask the Expert

Professor Robertson says: In my experience you can usually trace the root cause of a product disaster, such as a crash, a collapse, a personal injury, or simply a malfunction to a corruption of or deviation from the engineering design process. Many times as an engineer and an expert witness in product liability and personal injury cases, I've traced the trouble back to problems such as a lack of testing or reasonable foresight about the danger a product design could pose. I try to steer young engineers onto the right path in a sophomore seminar I teach called "Masters of Disaster."

The design process is a way of thinking about a product or structure from the initial stages of the concept, all the way through to when it starts being used. After engineers develop an initial concept, they'll build a prototype and enter a "closed loop" where they test and refine that prototype until it does what it is supposed to do and will last as long as it is supposed to last. But that cannot be the end. To avoid a disaster, they must also anticipate how users might misuse or abuse the product. Once engineers have considered that, they should either build in safeguards, or, if that isn't possible, at least package the product with appropriate warnings.

Of course there is a line between responsible anticipation and ridiculous overdesign. No engineer should feel obliged to build a parachute into a washing machine just because it's possible that someone will throw it off a building, but it is standard practice by now that cars have seat belts, table saws have guards, and ladders have warnings.

Sometimes a disaster comes about from carelessness. In one instance, a company's engineers thought they could make a less expensive underfloor home heating system using rubber hoses rather than copper or plastic pipes to carry the hot water beneath the floor. They neglected to test the system so it was not too surprising when the hoses began leaking and damaged thousands of homes. When I worked on that case, I conducted the tests the company didn't and showed that the engineers should have realized they were using an inappropriate material.

On the flip side, sometimes product abuse is properly anticipated and the proper warnings are given, but bad user behavior causes the product to fail anyway. When an automotive intake manifold gasket manufacturer was sued, the company was able to show that the gaskets only failed when the car owners let their coolant levels go well below the explicitly recommended amount.

Oftentimes product makers will claim that they couldn't possibly be expected to have anticipated the problem, especially if it seems subtle. They'll point out that failures are always more obvious once they've occurred. This was the case when some contraceptive makers were accused of negligence because their intrauterine devices (IUDs) were giving women infections in the 1970s and 1980s. It turns out that bacteria were exploiting a flaw in the design of a string attached to the IUDs. I was called to testify in these trials because I had studied bacterial movement and designed an IUD in 1970, with a string that did not have the same flaw. My experience showed that the problem could have been (and was) anticipated and preventable.

The bottom line for engineers–what I teach in my class–is to be sure that your validation and testing is done appropriately, that you think about misuse and abuse, that you think about guarding when you can't design danger out, that you think about warning, and more than that, once you get a product out into the marketplace, be sure you have a surveillance program in place so that you will learn from the consumer things you didn't learn in in-house testing. That way you can make new versions that are better and safer and move farther away from the possibility of disaster.