Exploring the physics of failure

ABB’s approach to reliability is one of the most sophisticated and rigorous in the world. Its 6000 square metre testing facilities in Helsinki, Finland are truly at the forefront of innovation, and worth more than €100 million. On top of that, its collective in-house knowledge and supply-chain resilience is unparalleled.

But building this enormous capacity for quality and reliability assurance was a cumulative process that began with a simple idea; a drive is only as good as its weakest component.

We sometimes take the tools and innovations that power industrial movement for granted. Motors and the drives that control them are the crucial components for much of that movement, and yet most people only ever really consider them when something goes wrong.

Imagine a blackout in hospitals, or the life support on the International Space Station failing, or a ship full of people being stranded in the middle of the ocean. While not every case of motor or drive failure is that dramatic, it does give you an idea of what could be at stake, and just how important it is that those parts are kept in working order.

The problem is it’s incredibly difficult to know exactly how or why those parts will stop working once they leave their safe, sterile factory, and face the unknown pressures of the real world. If we knew precisely what would cause them to go wrong, we’d design that Achilles’ Heel out of them before they left the prototype stage.
That means that, like all evolutionary processes, improving these motors and drives is inherently iterative, using constant feedback to keep adjusting and adapting. The challenge here is that it’s often very tricky to collect that feedback. If something breaks, the people involved (quite understandably) want you to fix it asap, not sit around taking notes – assuming they bring it to your attention in the first place.

In other words, the exact moment when we can glean the most valuable information is also the exact moment when it’s hardest to collect that information. What would be ideal would be to recreate all of those errors and issues under lab conditions. So, at ABB, that’s exactly what we do.

Task failed successfully

With so much focus on success and achievement, it can feel a bit odd to focus on failure instead. But that’s almost inevitable, given the nature of our unique 6,000 square metre testing facilities in Helsinki, Finland.

That’s where we push our products and product components to their breaking points - and beyond. It’s a temple to stress testing, an academy of catastrophe, a CSI lab where CSI stands for “Customer Supplier Investigation” rather than “Crime Scene Investigation.”

Every detail matters, from high-power semiconductors to the smallest resistors on a circuit board, so we try to replicate - and exaggerate - the conditions our components might face.

For example, we run our product components through electrical testing, pushing their voltage and current limits. We do mechanical durability testing, where we apply vibrations, thermal cycling and mechanical stress to simulate years of operation in only a few months. And we run environmental stress testing, subjecting components to extreme heat, cold, humidity, corrosive gases and even salt mist.

Righting the wrongs

As you can imagine, there’s a certain amount of pride that comes from our subjects surviving these gruelling conditions – but also a slight sense of disappointment. Because the real win comes when they don’t survive, and we have the chance to find out why.

It’s at this stage that our people conduct a thorough post-mortem, taking a scalpel (sometimes literally) to our successful failures to see how issues can be avoided in the future.

One of the most exciting features of this process is how counterintuitive our findings can be.

Take elevators in skyscrapers; very much the type of scenario where you want to minimise the chances of something going wrong in situ. Here, you naturally might be more concerned about the elevator that’s running all day every day, as opposed to the VIP lift that only gets used occasionally.

But the first elevator maintains a regular, predictable temperature, while the second one experiences irregular and intense thermal cycling, creating stress on its systems and components. That means the elevator that’s used less often will actually fail faster – not something that VIP getting whisked up to their penthouse will be expecting.

All of this stress testing means safer products for our customers, but also improved sustainability all round. By both measuring and extending the natural lifespans of drives and components, we can also increase the amount of time before these parts make their way to the scrapheap.

For our industrial partners, higher quality and more reliable motors and drives, combined with ongoing monitoring and predictive maintenance, means a clear and compelling competitive advantage.

As Samuel Beckett once said: "Ever tried. Ever failed. No matter. Try again. Fail again. Fail better."  

At ABB, we’ve made studying the physics of failure our speciality. To find out how, we’ve put together this deep dive into our approach – find out more here.