Predictive maintenance is a model-driven monitoring approach that uses both maintenance and operational observations to improve reliability and overall plant operations. This strategy requires that an operational model exists for the asset, system, or process, for a given set of ambient and input values. The model provides anticipated values for process parameters. It is
Expected decreases in outlays when using proactive maintenance strategies.
extremely effective in answering the question "When is the asset or process not operating as it has in the past?"—also referred to as anomaly detection. Some industries call it advanced pattern recognition (APR).
Detection and diagnosis of abnormal process conditions or critical conditions is essential for the effective operation of Proclndustries. Fault detection from measured data is typically done by identifying a region of normal process operations and evaluating in real time if it falls outside this operating region. Many algorithms define abnormal conditions using statistical process control (SPC) techniques, multivariate analysis, and machine learning (ML), to name a few. The key is to predict abnormal process conditions using advanced analysis for a certain cause-and-effect situation.
With the availability of faster, more intelligent ML algorithms, predictive maintenance is now becoming feasible on a division- or enterprise-wide scale. Unprecedented opportunities afforded by the industrial internet of things (IIoT) have further changed the playing field, and there are potential benefits yet to be realized. For example, predictive maintenance, originally based on selected asset condition data, has grown to accommodate online, real-time streams of multiple types of condition data received via sensors and even drones. Companies are applying ML algorithms to further refine their predictive analytics and prognostics. Edge computing is enabling process sensor data very close to process units for fault detection and vibration monitoring (Dewald and Santhebennur 2019).
Paul Morgan, the plant maintenance manager, realizes the value of his team's easy access to process and equipment real-time data. He explains, "Predictive asset management is quite a different approach from what we have been doing reactively in the past. A proactive enterprise asset management strategy demands partnership across the enterprise to ensure that all operations act in concert to keep assets working at their optimal level of performance."
Paul is also looking at the ISO 55000 standard by which corporations establish long-term business goals to manage assets. Bill Roberts, vice president of operations, has asked the team to find a way to reduce implementation time by utilizing modular asset templates for similar assets across the refinery: "We can implement in one place and replicate many times for all our sites."
The newest opportunity, prescriptive maintenance (RXM), is a multivariate approach that merges asset condition data with any combination of operating, environmental, process safety, engineering, supplier, or other related data to better diagnose conditions and prescribe specific options for corrective action. Advanced analytics, pattern recognition, modeling, ML, and artificial intelligence (AI) enables RXM so that companies can minimize the need for reactive maintenance on critical equipment. In addition, as older, experienced maintenance workers leave or retire, their knowledge will remain through development of these algorithms.