Model-based surface roughness estimation using acoustic emission signals

This paper proposes a model-based surface roughness estimation method using acoustic emission (AE) signals. Two powerful theoretical AE generation models are modified in this method to provide a comparative study. The proposed method extracts for the first time the theoretical relationship between surface roughness and the mechanism of AE generation in mechanical systems, and estimates roughness values using this relationship. AE signals collected from a pin-on-disc tribometer sliding contact are used to calibrate and validate the proposed method. The estimation results show that the proposed method gives results with an average error of 14% compared to experimental surface roughness measurements under various lubrication regimes and operating conditions. This AE-generation knowledge-based method will be advantageous for the field of surface condition monitoring because of its robustness and computational efficiency.