Plant Maintenance

Predictive maintenance uses sensor data to predict when to schedule plant maintenance. Comparing current sensor data against historical data can point toward conditions that lead to. Traditionally, this data has been integrated by human operators who use rules-of-thumb and intuition to link sensor data to required plant maintenance. Plant operations staff may lack the time and attention needed to detect signs that plant maintenance is required, leading to over- or under-prescribed plant maintenance.

More recently, machine learning and artificial intelligence (AI) have been applied to the problem of plant maintenance. By comparing historical sensor data with asset failure, software can detect failure signatures long before human operators would notice anything outside of normal operation. The Aspen Mtell^® software can notify plant operators months prior to the failure of a valve, seal, heat exchanger or other piece of equipment.

Being able to intelligently schedule downtime gives plant operators more control over plant maintenance. Multiple pieces of equipment can be serviced simultaneously, and supply chain logistics can be adjusted to reduced asset throughput.