Success with prescriptive and autonomous maintenance

Peter Bosmans, Technology Manager at Sitech, sees an increase in demand for prescriptive and autonomous maintenance. And he doesn't think there's anything particularly odd about that either. Peter: "We can show more and more real-world examples that demonstrate the enormous benefits of prescriptive maintenance, such as a reduction in operating and maintenance costs and improvements to plant performance."

The Asset Performance 4.0 Award presented to Sitech by BEMAS (Belgian Maintenance Association) last October is a symbol of recognition in the field of Data & AI-driven Plant Optimization. The jury concluded that the following two cases produced impactful results.

Case 1) Predictive maintenance: Electrical motor fault

Using our real-time monitoring dashboard, we were able to predict in advance that windings in an electric motor would break down within a short time. Because we had placed measuring equipment on these assets, we saw that the temperature of one winding suddenly skyrocketed – an increase that was not normal (anomaly). We set about on the same day to investigate where this rise had come from. Our data center sent a notification to the engineer, who then went to have a look at the motor. Grains of salt were found to be present in the motor. These grains generated electrical sparks that touched the windings and caused the rise in temperature.

Thanks to online monitoring of the asset, we were able to detect this in good timely and managed to prevent a fire in the entire motor.

Predictive maintenance means always looking for the optimum point. We do not want either reactive costs nor preventive costs to be too high for the customer.

Case 2) Prescriptive maintenance: Contamination in a heat exchanger

Another case where we were successful was early detection of contamination in a heat exchanger. A heat exchanger becoming contaminated over time is a normal process. This heat exchanger is usually cleaned every three years. But thanks to our monitoring dashboard, we can now see exactly when the heat exchanger needs to be cleaned. Only then will an alert be sent to the operator and/or technician at the plant. 

But when we analyzed the data on this occasion, we discovered more. In fact, we saw that the contamination disappeared due to changed behavior by two common process parameters. From this we concluded that there was a potential that the heat exchanger could clean itself. After a number of experiments (prescriptive maintenance, meaning that we indicate what the operators should do) we know for sure now. By modifying the process, it is indeed possible to have a heat exchanger clean itself (autonomous maintenance – we have taught the machine to solve the problem the way the operator always did), without removing the plant from service. A huge gain for the whole process industry, and one we are extremely proud of!