Fault diagnosis of EMU rolling bearing based on EEMD and SVM

Multi-Scale Sample Entropy-Based Energy Moment Features Applied to Fault Classification

“…Thus it successfully sidesteps the quadratic programming (QP) required for the training of the standard SVM. In the last decade, LS-SVM has been widely applied to a variety of classification problems [16] Due to the equality constraints, i can either be positive or negative according to the Karush-Kuhn-Tucker (KKT) conditions:…”

“…It has been widely used in bioinformatics and medicine fields to measure repeatability or predictability within a time series [10]- [13]. A few other applications are also observed in defect detection [9], state detection [14], structural health monitoring [15], and mechanical fault diagnosis [16], [17]. Mourad et al [9] detected gears defects by using SampEn to explore vibratory signals' regularity.…”

“…It has not received more attention to monitoring and diagnosis of rotating machines until recent years. Yang et al [16] proposed fault detection and diagnosis method for EMU rolling bearing based on EEMD, sample entropy, and SVM. Guan et al [17] developed a precise diagnosis method of structural faults of rotating machinery based on a combination of EMD, SampEn, and Deep Belief Network (DBN).…”

“…It can describe the distribution of energy on both time and frequency axis synchronously and is more effective than energy to reflect faulty vibration signals' essential characteristics. Therefore, it had been proposed as the eigenvector to capture the fault information hidden in vibration signals [ The current transorm-based entropy or energy moment approaches for mechanical fault feature extraction are based on one kind of time-frequency transformation like wavelet [14], EMD or EEMD [16], [17], [23], [24] or LMD [25], and have been experimentally verified. However, in a practical application, their effectiveness is often reduced by some uncertain factors such as the interference by false components, and the requirement on a prior information, etc.…”

Intelligent fault diagnosis of rolling bearings using variational mode decomposition and self-organizing feature map

“…Bearing fault detection and diagnosis can prevent failure, avoiding loss of life and property (Jiang et al, 2013; Shen et al, 2013). Faults are usually caused by improper processing and defective materials (Yang, 2012). When rolling bearings are damaged by wear and impact, damping, stiffness, and elastic force of the bearings change, resulting in nonlinear and nonstationary vibration signals (Cheng et al, 2013; Zhang et al, 2017).…”