We have already laid out the idea of AI Powered Maintenance, where Technology, Artificial Intelligence and humans work hand on hand. But the dominant maintenance strategies are still the ones where parts are changed based on time intervals. The management and bureaucratic effort for this strategy compared to just"run things until they fail" are much higher, but the benefits of increased uptime justify this effort. But the increased complexity is something which sneaked in over decades, as well as did the costs related to it. This is a much overseen effect, but the single maintenance operator is reminded of this fact every day. So here are some examples what maintenance crew needs to manage with time-based strategies.
Examples what maintenance crew needs to manage with time-based strategies
Get the information on "ideal" spare part change intervals from machine supplier
Create, store and maintain all the information
Create and manage IT-systems which provide information on machine run-time
Align maintenance slots with production
...
The list goes on and on. And there is also no doubt about the fact, that building the procedures around the IT-systems (e.g. integrating feedback of the maintenance crew on the real condition of the parts) takes many years, not considering the improvement process behind it, which takes decades.
The war of talent does not stop before the maintenance department
Managing the maintenance system for “time-based” strategies also requires a very capable maintenance crew. Every adaption to the recommended spare part change cycle by the supplier is just as good as the maintenance operators judgement. And this capable maintenance people are just harder and harder to find. But after centuries of increasing complexity, technology needs to reduce complexity again to make life easier for the single maintenance operator.
Condition Monitoring technologies with AI reduce complexity
Condition monitoring technologies done in this sense can just deliver this long deserved complexity reduction. Imagine a way, where you don't need to "interpolate" to the health status of assets via "usage" or "time", but based on real physics and mechanical wear-down. If you are interested in monitoring the health status of a certain pump,
Why would you write down the last time you changed it?
Why would you write down the operating hours?
Why would you store spare parts "just in case" something happens?
You would not do that, if you have a reliable condition monitoring procedure in place, which indicates a failure in a reasonable enough time-frame. Condition monitoring technologies have been around for quite some time, but do also need highly skilled operators and a lot of experience. Artificial Intelligence in combination with Ultrasound keeps that complexity away from humans, coming up with solid and data driven health status assessments. We also believe, that ultrasound benefits most from Artificial Intelligence and will become the dominant technology. (Why ultraosund benefits most from modern AI Technologies)
Ultrasound for Predictive Maintenance
Recent technological advancements (e.g. Sensors, Electronics, Artificial Intelligence) have made a brand new approach possible for Predictive Maintenance. It is the combination of a well known approach for condition monitoring (ultrasound) with Artificial Intelligence. Currently, analysing ultrasound data is a long and time-consuming task done by humans. But instead of humans sitting for hours in front of computer screens watching charts and data, AI is doing the hard job of coming up with a health status judgement - this video shows you how this works. There are solid advantages of ultrasound compared to other technologies, which is why ultrasound is also used by NASA to detect bearing failures. In factories around the world, ultrasound is also considered your first line of defense for problem detection. Combined with the new technology now available, AI will transform the way maintenance is done forever.
