After graduating from Durham University with a PhD researching the impact of drive train dynamics on the physics of failure of wind turbine components, I joined the rail sector, performing asset management and maintenance support roles. In 2018 I joined Senseye to provide technical support for maintenance Departments on their predictive maintenance journey, and now lead a growing team who guide maintenance practitioners across the full spectrum of industries.
AI and predictive maintenance – a benefit, not a threat
Elon Musk once described AI as an “existential threat”. In our opinion, while it’s true that AI will certainly transform the way we work, his view couldn’t be further from the truth. In fact, not only will it enhance people’s jobs, it will also create new ones – particularly when it comes to maintenance.
When you consider how the development of chatbots and driverless vehicles has disrupted call centers and logistics, it’s perhaps unsurprising that people can be wary of the impact AI will have on their job prospects. When confronted with a new technology that can predict the failure of machines, it’s reasonable to expect that this could encourage the use for lower-paid, less experienced operatives, eradicating the need for highly skilled maintenance technicians.
The reality is entirely the opposite. We’ve seen companies attempt to follow this model, but most of these projects have been doomed to failure. They just haven’t grasped the fundamental concept behind predictive maintenance. Giving highly skilled members of staff a tool that helps focus their efforts, and removes some of the day-to-day tasks that provide no value, will enable them to direct their skills toward more effective work, significantly improving their efficiency.
After all, no matter how good a system is at determining exactly when a particular asset will fail, you still need someone to actually perform the maintenance. The challenge is in making sure their time is spent effectively.
An AI-based predictive maintenance system is, essentially, a decision-support tool that enables maintenance teams to deliver greater value for their employers. But automated systems are only as effective as the people they serve. Human expertise is needed to ensure they’re receiving the data needed to generate meaningful insights.
While AI provides insights into the condition of an organization’s assets, any decision on how to act on those insights must be taken by a human operator. And, of course, no AI-based software is able to physically carry out the maintenance required.
Rather than making maintenance teams redundant, automated predictive maintenance systems support them, allowing them to make better, more informed decisions and achieve far greater efficiency than would be possible using manual processes alone. Senseye PdM typically enables our customers to boost maintenance staff productivity by 55 percent, and improve the accuracy of downtime forecasting by 85 percent. In short, the use of AI-powered predictive maintenance systems has the opposite effect to what many people might have imagined – it is not a threat.
One of our customers in the automotive sector, for example, was performing condition-based maintenance and, while they had data from their machines, the system they used wasn’t automated. The team would get involved and could only look at a maximum of 50 assets (with tens of signals each to attempt to understand), trying to search the data for patterns from which they could carry out maintenance. Automating the first part of the process, however, meant they no longer had to look at the same 50 assets all the time. By constantly looking for the right signals and alerting them to what was of concern - and what wasn’t - the system enabled the team to look at 2,000 assets – the entire balance of the plant in that area. Each person on the team was now much more valuable – rather than reducing headcount, the same people were able to deliver a far greater capability.
A powerful tool
Maintenance managers know their assets well. If they walk past a machine and hear an odd sound or smell something unusual, they’ll typically take it as a sign that something’s not right. An automated predictive maintenance tool isn’t all that different. It’s just using intelligent algorithms to look for changes in data signatures to tell if something’s wrong with an asset. It’s simply improving efficiencies. The human element still matters.
Ultimately, AI is there to support maintenance teams, not replace them. We’ve seen from our customers’ experience that, after just three months, using Senseye PdM has completely changed the way they work. They can make much more enhanced, proactive decisions. By seeing the machine data, and how it’s changing, they’re in a much better position to decide whether or not to replace a particular asset.
The growing adoption of AI is undoubtedly disruptive, managed the right way it is a significant benefit. The job of a predictive maintenance solution is to give targeted insights – it’s still up to the maintenance team to make the important decisions. AI is a powerful tool that, by improving their efficiency and effectiveness, makes maintenance engineers even more valuable to their employers. And that hardly sounds like an existential threat to me.
To learn more about how Senseye PdM could improve the day-to-day operation of your maintenance team, please get in touch.