Remaining Useful Life Prediction Using an Ensemble Learning-Based Network for a Belt Conveyor System

Abstract

The belt conveyor system is widely used in production and distribution industries because it is more cost-effective than manpower and can be used in a variety of ways. Prognostics of the belt conveyor system is the main activity to maintain efficiency. Lack of performance of the system is most often an error in which the system is no longer available to meet the desired performance which arises the entire system can be damaged and fatal industrial accidents may occur. In this paper, we present a model that predicts the remaining useful life of the head pulley, a key part of the belt conveyor system. The ensemble learning-based model to predict is composed of a deep learning-based representation model and boosting model. The model is trained using a combination of classification and regression rather than simple regression to predict the remaining useful life. The data used to train the model was collected by directly building a test bed with an environment similar to a belt conveyor system.