Proper Orthogonal Decomposition for the prediction of fretting wear characteristics

Fretting happens when there is an oscillatory displacement between two contact parts. In gross slip regime, wear is a dominant failure mode, which has a detrimental effect on the contact parts. To understand the wear behaviour in the fretting process, many experiments and simulations have been done. This phenomenal analysis is time-consuming, especially when fretting wear investigation is required under certain loading cases or during certain number of cycles. Under these conditions, fast prediction tools provide an alternative for this kind of problem. In this paper, a model reduction based on Proper Orthogonal Decomposition with Radial Basis Function (POD-RBF) is used to predict the wear characteristics of specimens under different slip amplitudes and at different fretting wear cycles based on simulation results. To demonstrate the advantages and disadvantages of POD-RBF, Grey Model and Curve Fitting are used to verify POD-RBF for one-parameter prediction, while Artificial Neural Network is applied to verify POD-RBF for three-parameter prediction. The results show that the approach of POD-RBF can predict the wear characteristics under different slip amplitudes and normal loads at different cycles with high accuracy and high computational efficiency. •Predictive model for fretting wear characteristics.•Proper Orthogonal Decomposition with Radial Basis Function (POD-RBF).•Grey Model, Curve Fitting Artificial Neural Network are used to verify the results.•POD-RBF provides high computational efficiency.