Study of Worn Surface Characterization Based on Singular Entropy

A new method based on singular entropy is suggested for estimating the worn surface using wavelet analysis, singular value decomposition and information entropy. Low frequency feature parameter and high frequency feature parameter of worn surface based on statistical feature were extracted. The running-in wear test for materials of piston ring against cylinder liner was conducted on a pin-on-disc tester under CD40 lubrication. It was shown that low frequency feature parameter and high frequency feature parameter can be used to reveal the topography feature of running-in surface. Low frequency feature parameters reflect contact area of the running-in surface, the larger low frequency feature parameters, the larger contact area. High frequency feature parameters reflect roughness of the running-in surface, the larger high frequency feature parameters, the smaller roughness. Therefore low frequency feature parameter and high frequency feature parameter can be considered as the feature parameter for the running-in surface evaluation.