Application of morphological wavelet and permutation entropy in gear fault recognition

  • PDF / 802,064 Bytes
  • 10 Pages / 595.276 x 790.866 pts Page_size
  • 5 Downloads / 155 Views

DOWNLOAD

REPORT


SPECIAL ISSUE

Application of morphological wavelet and permutation entropy in gear fault recognition Wenbin Zhang1 · Yasong Pu1 · Dewei Guo1 · Jie Jiang1 · Libin Yu1 · Jie Min1 Received: 22 April 2020 / Revised: 17 July 2020 / Accepted: 8 September 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract In this paper, a new gear fault recognition method was proposed by using morphological wavelet and permutation entropy. Firstly, the morphological Haar wavelet was proposed based on morphological wavelet, and it was used to pre-process the measured gear vibration signal. Then, the permutation entropy was used as the eigenvalue of gear fault to be extracted from the vibration signal, which included four working conditions: normal, mild wear, moderate wear, and broken teeth. Finally, according to different faults corresponding to different permutation entropy distributions, the various fault states were classified, and the permutation entropy distributions of non-denoised signals were compared. It could be seen that the morphological Haar wavelet had good de-noising effectiveness, and permutation entropy could express the feature of different gear conditions. The example of gear fault recognition proved that the combination of morphological wavelet and permutation entropy could effectively improve the ability of gear fault classification. Keywords  Permutation entropy · Morphological wavelet · Fault recognition · Gear

1 Introduction The gearbox is the main part of mechanical equipment for movement and power transmission. In gear transmission, the failure of the gear often induces the failure of the machine, and then causes the equipment to stop or even damage. Therefore, the research of gear fault feature parameter extraction technology under a strong noise environment has always been a hot spot in fault diagnosis of rotating machinery [1–6]. Researches show that the signal characteristics of the gear after failure are typical non-stationarity and nonlinearity. There, it is particularly important to study the method of how to effectively extract the fault state characteristics from gear fault signals, which reflect the different working conditions [7–9]. Permutation entropy is a kind of information entropy, which is often used to study the complexity of nonlinear time series. Compared with the Lyapunov index, fractal dimension, sample entropy and other commonly * Wenbin Zhang [email protected] 1



College of Engineering, Key Laboratory of Mechanical Performance Analysis and Optimization of Plateau in Yunnan Province, Honghe University, Mengzi 661199, Yunnan, China

used nonlinear dynamics methods, permutation entropy has the advantages of simple calculation, fast operation speed, good anti-noise and anti-interference ability, and short time series needed to calculate the permutation entropy [10–13]. However, before feature parameter extraction, the original sampled signal will be interfered by various noises. Therefore, the selection of appropriate noise reduction methods for noise reduction prep