In a recent white paper discussing profitable reliability, Schneider Electric identifies four levels of maintenance from reactive through to prescriptive maintenance. Prescriptive maintenance uses information from predictive maintenance tools to prescribe a solution. This is done using prescriptive analysis.
The goal of manufacturers is to improve productivity and hence profitability. This starts with controlling operational profitability in real time. Part of this links to controlling and measuring the reliability of the plant’s assets down to equipment level. Advances in process control have enabled manufacturers to increase operational throughput. But that has come with a risk because pushing equipment harder moves it closer to its thresholds of reliability and safety. This puts assets under continuous strain, degrading reliability and affecting operational performance.
Industrial maintenance tools and practices have progressed and evolved over the last two decades. Firstly reactive-maintenance first developed into preventive maintenance. Then it moved to predictive maintenance and finally to prescriptive maintenance. Each advancement has led to a corresponding increase in asset reliability. But manufacturers were soon stuck in a cycle. As techniques improved asset reliability, process controls improvements increased production processes, further challenging reliability.
- – Reactive maintenance: Repairing equipment that has broken down, focusing on restoring it to normal operation.
– Preventative maintenance: Regular maintenance on working equipment to prevent or reduce unexpected failure.
– Predictive Maintenance: Determining the condition of in-service equipment in order to predict the need and timing of maintenance.
- – Prescriptive maintenance: Using analytics can show that a piece of equipment is heading for trouble. It can prescribe prioritised, pre-determined, expert-driven mitigation or repair.
According to the author, it does not require more advanced technology, but a rethink how users address the issue. That begins with how asset reliability is measured in the first place.
Historical performance is not the best indicator
Previously, measuring the reliability of industrial assets revolved around analysing historical performance. The assumption was that past behaviours would indicate future performance. A more effective approach is measure how likely it might be that a reliability incident would occur. The name for this is reliability risk.
Advancements in data science and condition monitoring in industrial operations are making the direct real-time measurement of asset reliability workable. They will make more sophisticated, real-time approaches to controlling asset reliability workable too.
From extensive experience, there is considerable information on reliability at the equipment asset level. For example, accurate reliability curves, coupled with condition and process measurement, enable accurate measurement of asset reliability risk as seen in this pump failure curve.