AI, agility and accessibility: customer-driven R&D at SKF

Renewed focus on usability and customer needs changes the way SKF conducts its research and development activities

 “Our business in the midst of a massive change: from a product-selling company to a function provider,” says Victoria van Camp, CTO and President, Innovation and Business Development at SKF. “We want to provide our customers with what they actually need. That isn’t lots of bearings, it’s machines that run and run.”

This shift in strategic direction reflects wider changes in society, she explains, with industries from music retail to mobility all moving to models where customers pay only for the services they consume.

The new strategy is encouraging every part of the company to rethink the way it works, says van Camp, and that includes SKF’s core research and development functions. Becoming a service provider rather than a product manufacturer requires “a completely different mind-set in R&D,” she says. “If you are responsible for the performance of a customer’s machine over its lifetime, you need a much deeper understanding of how your products are being used in that machine. Are they continually exposed to hot, corrosive chemicals? Are they being hit with a hammer during installation?”

Usability first
One result of the change is a new focus on making products more accessible and easier to use. That can involve combining commonly used products together, so they are simpler to buy, manage and install. Here, van Camp cites the example of bearing housings supplied with integrated seals. It also includes SKF Bearing Assist, a mobile app that provides guidance for mounting a new bearing. The app walks users step-by-step through the best process for their specific bearing and generates an installation report that can be stored with company maintenance records or held in the SKF cloud, making future maintenance and replacement easier.

The company is also working on new products designed with usability in mind, such as bearing housings that can tell the operator when they are properly aligned.

“Even a small misalignment can significantly reduce the life of a bearing, but these products are often installed in inaccessible places or difficult conditions like underground mines,” says van Camp. “We want to make it easier for our customers to get those tasks right.”

Learning from data
SKF is also making better use of the huge quantities of data generated by modern industrial products. Bearings can be equipped with sensors that measure temperature and vibration. Connected machines and digital management tools can generate detailed records of the way equipment is used and maintained. But turning that data into “machines that run and run” is a different kind of challenge.

“Our engineers are experts at analysing performance data, but SKF produces billions of bearings every year,” says van Camp. “When we started to think about applying those approaches to everything we make, it quickly became obvious that the task was too big for even an army of human analysts. The only way you can do it is with artificial intelligence.

That realisation quickly led to another. “We had people inside the organization working on AI systems, but we recognised that building the capabilities we needed, at the speed we needed to do it, would require outside expertise,” she says.

Adopting a more collaborative approach to R&D was a big change for a 113-year-old company with a strong track-record of independent research but engaging with the wider world began to pay dividends very quickly. The SKF AI group made contact with Presenso, a start-up based in Israel, which was working on the application of automated machine learning in industrial reliability.

“It was immediately clear that there was a good fit between what Presenso was doing, and what we wanted,” says van Camp. “They had a great product and, as importantly, they had a deep understanding of our markets. One of the company’s founders is a mathematician and statistician, the other is a mechanical engineer who specialises in reliability.” The match turned out so well that SKF acquired the Israeli company in October 2019.

SKF industrial customers are already benefitting directly from Presenso technology, which can automatically sift through the data generated by a factory, spotting issues and improvement opportunities. But for van Camp, the real power of the AI-driven approach will come from its application to the aggregated data SKF collects from thousands of customer sites.

“Our engineers spend a lot of time investigating products that fail in the field,” she says. “But the overwhelming majority of our products don’t fail, they outlast the machines in which they are installed.” By analysing big populations of bearings in the field, AI technology can identify the specific issues associated with higher rates of failure, helping SKF to fine tune its research, and making customers’ machines perform better.

“In a paper mill, for example, you may have hundreds of bearings in rollers and conveyors, all working in a hot, wet environment and in the presence of aggressive chemicals,” says van Camp. “Corrosion is a potential problem for those bearings, and we have invested a lot of time developing special coatings and barriers to prolong the life of our products. Now, using machine learning and AI, we can look much more deeply at what is going on in those plants, and differentiate the places where a standard product will work well from the few cases where you need a more specialised solution.”

Innovating through the crisis
While SKF’s new strategy was under way before the coronavirus pandemic, the crisis has proved to be a catalyst for further change. “Lockdowns were extremely challenging for many of our customers,” says van Camp. “They needed to keep their operations running despite personal shortages and access restrictions.”

SKF brought forward the launch of new products and services, such as its Bearing Assist app, to provide better support for those customers. “With Bearing Assist, even a relatively inexperienced member of staff can work like an expert,” she says.

The success of Bearing Assist and other projects during Covid-19 has also accelerated SKF’s transition to a more agile approach to research and product development. “We used to have a rigid roadmap outlining our plans up to five years in the future,” says van Camp. “Now our R&D teams are much more focused on addressing our customers’ most pressing challenges, with the aim of releasing new solutions in weeks or months.”