Study on fault diagnosis of broken rotor bars in squirrel cage induction motors: a multi‐agent system approach using intelligent classifiers

Abstract

A critical factor in industrial production maintenance is decision‐making in fault diagnosis. Motor current signature analysis (MCSA) is an established condition‐based maintenance method to make fault diagnosis in induction motors (IMs). The occurrence of multiple interacting faults, as well as emergent behaviour, can be tackled efficiently especially when MCSA is combined with distributed problem‐solving artificial intelligence (AI) techniques. Therefore, in this study, an intelligent multi‐agent system (MAS) is presented to make decisions on the fault conditioning of a three‐phase squirrel cage IM. The incorporated agents represent different health conditions of the same IM, with faults that may occur in the rotor bars. All agents utilise an AI method, and use for training MCSA experimental data. A supervisor agent (SA) initially communicates with agents employing feed‐forward artificial neural networks trained with the back‐propagation algorithm and performs the final fault diagnosis by evaluating their responses. When a decision cannot be made on the fault type, the SA employs another agent that uses the k‐nearest neighbour rule. The proposed method achieves high fault diagnosis accuracy. Its performance is also compared with results previously obtained from the same motor when an adaptive neuro‐fuzzy inference system combined with a subtractive clustering method had been used.

A Normalized Frequency-Domain Energy Operator for Broken Rotor Bar Fault Diagnosis

“…As shown in (24), there are mainly four terms in the energy operator of the faulty motor current signal, including a constant term related to the supply frequency, a fluctuating term related to the fault frequency and two fluctuating sideband terms around twice of the supply frequency. Obviously, the fluctuating term that related to the fault frequency is more significant and can be extracted for the fault detection with higher accuracy.…”

“…Anik et al [23] delivered Rarleigh quotient technique and extended-Kalman based on MCSA can be used to remove the supply frequency and make the BRB related sideband obvious. Maria et al [24] presented an intelligent multi-agent system with an overall accuracy ranges from 81.3% to 89.4% to diagnose the rotor bar status of IM.…”

A Combined Short Time Fourier Transform and Image Classification Transformer Model for Rolling Element Bearings Fault Diagnosis in Electric Motors

“…Traditional methods require feature extraction of the bearing vibration signal, dimension reduction and classification, which lead to complex mathematical models. To automate the process, machine learning (ML) methods have been introduced, such as the k-nearest neighbors (KNNs) [9], the adaptive neuro-fuzzy inference system (ANFIS) [10], fuzzy cognitive networks (FCNs) [11], the multi-agent system (MAS) approach using intelligent classifiers [12] and the support vector machine (SVM) [13]. In recent years, deep learning methods and associated techniques have been achieving dramatically increased popularity among the research areas of neural networks and artificial intelligence.…”

Bearing Fault Diagnosis Approach under Data Quality Issues

“…To expedite the fault diagnosis process and unravel the hidden features, several machine learning (ML) approaches have been proposed successfully. These methods include artificial neural networks (ANN) [15], k-nearest neighbors (KNNs) [16], fuzzy cognitive networks (FCNs) [17], multi-agent system (MAS) approach using intelligent classifiers [18], and support vector machine (SVM) [19]. By analyzing and learning from the bearing data, ML methods can apply what they learned to make informed decisions about bearing health.…”