Prediction of contact temperature rise between rough sliding bodies: An artificial neural network approach

Prediction of contact temperature rise between sliding bodies is difficult due to the large number of parameters that affect the contact phenomenon. It is far more complex if the instantaneous roughness of the sliding pair is taken into consideration since the roughness may change significantly during sliding. It was felt that a multilayer feedforward neural network might be a convenient method to predict the change in surface roughness and the instantaneous maximum temperature rise at the contact between rough sliding bodies. The model accepts initial surface roughness of the mating surfaces, their material properties and the operating parameters as input variables and predicts the final surface roughness and the corresponding maximum contact temperature after a specified sliding time.