A Review on Data-Driven Condition Monitoring of Industrial Equipment

Abstract

This paper presents an up-to-date review of data-driven condition monitoring of industrial equipment with the focus on three commonly used equipment: motors, pumps, and bearings. Firstly, the general framework of data-driven condition monitoring is discussed and the utilized mathematical and statistical approaches are introduced. The utilized techniques in recent literature are discussed. Then, fault detection, diagnosis, and prognosis on the three types of equipment are highlighted using a variety of popular shallow and deep learning models. Applications of these techniques in recent literature are summarized. Finally, some potential future challenges and research directions are presented.

Charge Pattern Detection Through Electrostatic Array Sensing

“…9. The EMD approach employs an iterative screening process to decompose the original non-stationary signal into a sequence of new and improved signals, known as intrinsic mode functions (IMF), which represent the oscillatory components of the original signal [28]. In the specific example depicted in Fig.…”

Condition Monitoring and Fault Detection in Small Induction Motors Using Machine Learning Algorithms

“…Fourier transforms were studied first, but more modern approaches examine combined time-frequency domain signals. These include wavelet transforms, Empirical Mode Decomposition, short-term Fourier transforms, Hilbert transforms, singular value decomposition, Park's Vector Analysis, and the Wigner-Ville distribution, among others [16,24]. Some faults are easily detected and diagnosed in the frequency domain, but more commonly, the time-frequency band coefficients are passed as features to a classifier algorithm in the usual inferential-sensing approach [16,24].…”

Combining Wavelets and AR Identification for Condition Monitoring of Electric-cam Mechanisms Using PLCopen Readings of Motor Torque

“…In the realm of heavy-duty machinery, the occurrence of component failures is a common concern. Addressing the diagnosis of their health status and predicting their remaining useful life (RUL) has become a focal point of research efforts in the field (Soualhi et al, 2017;Qi, Zhang, & Spencer, 2023). This advancement owes much to the everincreasing computational capabilities of computers, spanning from on-board processing (edge-computing) to external systems like industrial computers and cloud-based computing.…”

“…This work presents a CM application where the combination of wavelets and Model-of-Signals aims at monitoring the state of health of electric motor driven mechanisms. In this context, the major part of the proposed CM procedures relies on vibration signals (Qi et al, 2023). An alternative consists in exploiting current motor signals (Nandi, Toliyat, & Li, 2005).…”