In my previous blog post, I explained why APM-centric maintenance should replace reliability-centered maintenance (RCM) as the pinnacle of the Asset Performance Management (APM) Maturity Model Pyramid. As organizations rethink best practices in APM, the time to redefine predictive maintenance has also arrived.

 

 

As the maturity model evolves, I assert that the definition of predictive maintenance must also be updated. This label is associated with product claims such as lube oil strategy that suggest additional care might lead to better performance; aspirational but not predictive at all. “Predictive” must include high-quality prognosis: not inspirational guesses. 

True predictive analytics use unobtrusive, near real-time monitoring to assess equipment health, with industrial AI analytics providing warnings of progressive deterioration and likelihood of failure -- without physical inspections. Additionally, applications using data from the maintenance management system detect specific failure conditions and root causes earlier, with greater accuracy, further enhancing maintenance activities. Only such advanced solutions provide alerts that give companies a clear means to avoid degradation and breakdowns, by prescribing specific repairs or process adjustments.

Prescriptive capability is paramount in APM 4.0 solutions. Do not let vendors dumb down “prescriptive.” It originates from the word “prescribe:” to advise a course of action as a remedy. In other words, prescriptive maintenance tells you what to do when an event is detected. The advice could suggest an operator make changes, recommend maintenance service and repair, and/or automate actions to examine multiple risk and reward scenarios that may apply within weeks and months of impending failure warnings. 

Fundamental to APM 4.0, prescriptive maintenance is a cognitive process involving symptoms analysis, health diagnosis, consideration of alternatives for treatment, and then a specific recommendation for action. Prescriptive maintenance assembles, aggregates and validates data streams from machines and manufacturing processes; then uses diverse scientific, computational and mathematical disciplines and business rules to predict outcomes, and lastly recommends and executes actions to deliver the preferred end result.

 

Monitoring Processes AND Equipment to Prescribe Maintenance

A major omission in contemporary maintenance must be corrected. Equipment degradation and distress often occurs due to poor operation when equipment strays outside of safety and design limits. Such conditions include dry feed, pump cavitation, liquid carry-over and incorrect setpoint entries. Predictive technologies must detect these conditions with process analytics and warn operators so they can take action to alleviate harmful conditions. Companies may then completely avoid degradation and failure or minimize risks before a major catastrophe. 

Early predictive forecasts permit service and repair to be planned and scheduled with minimal impact on safety, the environment and plant operation. Correcting minor problems before catastrophe occurs, particularly in the case of process-induced stress, allows organizations to avoid deterioration and incumbent maintenance altogether. Consequently, new APM 4.0 process monitoring capabilities assure the process operator has a much greater impact on process/equipment reliability, where prescriptive maintenance-driven changes to the manufacturing process eliminate deterioration and damage, resulting in more reliable and longer lasting equipment. 

 

To learn more, read Prescriptive Maintenance: Transforming Asset Performance Management.