Ensemble machine learning for intelligent condition monitoring

Traditionally, the maintenance process in engineering has been a costly and time-consuming process that requires large amounts of resources and capital. Condition-based maintenance is a maintenance strategy that aims to reduce maintenance-related expenses and disruptions in industrial processes by performing maintenance only when necessary. Condition-based maintenance seeks to reduce equipment downtime by monitoring the actual condition of a machine, so that maintenance is only conducted right before the point of failure so that organizations get the maximum value possible from their machines. Because condition-based maintenance is very data-driven, it is a prime contender for a process that can be enhanced by the application of machine learning. In this paper we propose an ensemble machine learning model that combines the random forest, support vector machine, and artificial neural network classifiers to classify time-series data from sensors monitoring the condition of bearings into normal and faulty categories. If a data point is classified as faulty, the fault is diagnosed into one of the following categories: outer race, inner race, or roller element. By combining previous strengths with new ideas, this paper presents an intelligent condition monitoring model that is capable of accurately identifying and classifying faults in time-series condition data. The efficacy of the model in adequately detecting and diagnosing faults in machines is discussed, and an illustrative example is given to show the application of the model.