Advanced predictive maintenance

There is considerable excitement around the potential of advanced predictive maintenance approaches.

The promise of these new techniques is tantalizing. Using machine learning technologies to analyze historical failure and performance data, we aim to forecast when and how a component is likely to fail in the future with a high level of predictability.

Several companies have started down the path by establishing automation and instrumentation, which combined with a rigorous maintenance record, will create the rich data that machine learning systems require.

Although it appears to be an easy project, there are things asset managers should consider:

Data. Unplanned downtime can be concentrated in a small number of large events. This means you will have very few data points for predictive maintenance systems to learn from.

Time. Predictive models operate over time horizons that may be too short to be useful. Predicting that a part will fail in two days or two weeks is useful on a truck or a tool, but may not be useful in a plant where shutdowns take several days and maintenance teams require months to plan interventions and obtain spare parts.

Impact. The facilities operate critical assets with a high degree of redundancy and few individual points of failure. If a pump stops unexpectedly, operators can often switch to a backup unit with little impact on production, reducing the impact of predictive maintenance.

Request a session with our experts to learn how Maximo can help facilitate your path to predictive maintenance.