Multifractal Detrended Fluctuation Analysis of Frictional Vibration Signals in the Running-in Wear Process

Multifractal detrended fluctuation analysis (MFDFA) provides a valuable tool for extracting nonlinear characteristics of signals, which makes it very powerful for the status recognition of friction pair by analyzing frictional vibration signals. This paper presents an algorithm for denoising the frictional vibration signals by using the ensemble empirical mode decomposition. The denoised signals of frictional vibration were analyzed by utilizing the MFDFA algorithm to derive the q-order Hurst exponent as well as multifractal spectrum. The paper illustrates these issues by analyzing signals taken from the friction and wear experiments on CFT-I testing machine. The results show that the q-order Hurst exponent, as well as multifractal spectrum, presents a certain trend with the running-in wear process of friction pair. The MFDFA algorithm can extract effectively the fractal characteristics of the frictional vibration signals. The frictional vibration signals could be characterized by the q-order Hurst exponent and multifractal spectrum.