Examining equipment condition monitoring for predictive maintenance, a case of typical process industry

Maintenance of the production machine plays an important role to keep the production process properly run. Improper maintenance may cause problems in production. Corrective Maintenance strategy, a response to equipment failure, is needed to rectify the failures. However, this strategy may cause higher downtime and higher costs due to loss of production. Predictive Maintenance (refer to Condition-Based Maintenance) utilizing equipment monitoring data can be used for optimizing the maintenance strategy by predicting the future machine condition. This research attempts to examine the condition-based equipment data using the data analytics approach to developing a Predictive Maintenance program. Several methods are applied. K-means for clustering the failure characteristic, Support Vector Regression (SVR) model used for predicting equipment failure. The result and discussion represent that SVR and K-means model suitable for developing equipment failure prediction, useful support for managing the maintenance activities.