Data-driven approach for predicting remaining useful life of belt-conveyor systems using polynomial curve fitting model

What is it about?

A successful data-driven approach has been adopted to predict the future Health Indicator (HI) and Remaining Useful Life (RUL) of the conveyor belt system, aiming to avoid potential problems and reduce the occurrence of future failures. It is necessary to record and analyze each operational profile of the conveyor system to achieve extrapolation using the data-driven model. By following the sensor’s signal at any point in the operation scenario, the RUL of the conveyor can be estimated. The regression algorithm, namely the polynomial curve fitting (PCF) model, was employed for HI and RUL prediction. Experimentally collected data from tests conducted on the conveyor belt were utilized to train the PCF model. The experimental results obtained from four categories, linear and nonlinear deceleration and linear and nonlinear acceleration deterioration, were reported and discussed. The effectiveness of the proposed PCF model in predicting HI was demonstrated using the appropriate performance metric (MAE). Additionally, the results revealed that the proposed predictive strategy provided a reliable estimation of the RUL.

Featured Image