Machine Learning Approach for Broken Rotor Bar Detection in Induction Motors

Induction motor is one of the most commonly preferred motors used in the industry due to its long life & rugged construction. Faults occurring in an induction motor can significantly affect the performance of the motor and the drive which is running it. This project aims to study & analyze the faults which can occur internally in an induction motor .To implement fault analysis first the machine is modeled using MAGNET software. The motor is studied under healthy conditions. Subsequently faults are introduced & fault analysis is carried out. In the model of induction rotor bars are broken and the study is mostly based on magnetic field analysis of the motor under healthy conditions & under faulty conditions. The variation of starting torque with respect to increasing number of broken rotor bars is studied and a graph was plotted to show the variation. The stator winding currents produce a magnetic field that rotates in a counterclockwise direction. The changing magnetic field of the stator induces electromotive force in the rotor cage winding. The induced emf causes current to flow and magneto motive force in the rotor windings. In turn the rotor mmfs produce a magnetic flux pattern which also rotates in the air gap at the same speed as the stator winding field. Induction motors are the most commonly used prime movers in industrial applications. They are best suited for constant speed applications. Speed control can also be done with the help of power converter circuits. The 3-phase squirrel cage induction motor is the workforce of the industry as it is rugged & reliable. The interaction between the primary field and secondary currents produces torque from zero rotor speed onwards. The rotor speed at which the rotor currents are zero is called ideal no-load or synchronous speed. The rotor windings may be multiphase (wound motors) or made of bars short-circuited by end rings (cage rotors). All primary and secondary windings are placed in the uniform slots stamped into thin silicon steel sheets called laminations. The induction machine has a rather uniform air gap of 0.2 to 3mm.The secondary windings may be short circuited or connected to an external impedance or to a power source of variable voltage and frequency. Induction motors are the most commonly used prime movers in industrial applications. They are best suited for constant speed applications. Speed control can also be done with the help of power converter circuits. The 3-phase squirrel cage induction motor is the workforce of the industry as it is rugged & reliable. Electrical related faults are frequently occurring faults in three-phase induction machine which will produce more heat on both stator and rotor winding. This leads to the reduction the life time of induction machine. Stator winding fault in Induction Motors can be detected using Unknown Input Observer (UIO) & Extended Kalman Filter methods. They are used for speed estimation and fault detection methods respectively(1).Using analysis of permeance and MMF harmonics, frequency of air gap flux density harmonics which occur due to irregularities are calculated. It helps in detecting faulty ball bearing conditions (2). A procedure for electromagnetic design of three phase Induction Motor is discussed. This procedure is based on self-consistent equations. The electrical and magnetic properties are imposed by the user but the geometric dimensions are automatically calculated (3).The choice between Copper rotor bars and fabricated Aluminum rotor bars is discussed and debated. The fundamentals of rotor construction and basic information on how the Induction Motor works are discussed (10). The new IEC standard which defines the stator winding insulation requirement when the Induction Motor operates with Adjustable frequency drives (AFDs) is discussed. The

Early detection and diagnosis of incipient faults is desirable for online condition assessment, product quality assurance, and improved operational efficiency of induction motors. The characteristic frequency component(CFC) of broken rotor bars is very close to the power frequency component in frequency domain but far less in amplitude, which brings about great difficulty in detecting the broken bars in induction motors. A new method based on wavelet ridge is presented in this paper. As a motor accelerates progressively and the CFC of its broken rotor bars approaches the power frequency component gradually during the motor's starting period, the wavelet ridge-based method is adopted to analyze this transient procedure and the CFC is extracted effectively. The influence of power frequency can be eliminated, and the detection accuracy can be greatly improved. Furthermore, experimental results show this is truly a novel but excellent approach for the detection of the broken rotor bars in squirrel-cage induction motors.

In this paper, a new approach is proposed to perform broken rotor bar fault detection in induction motors using of support vector machine (SVM) classifier. New features such as harmonic curve area, harmonic crest angle and harmonic amplitude have been extracted from power spectral density (PSD) of stator current in steady state condition using of Fast Fourier Transform (FFT). It is shown that combination of the first couple of these features had very better results compare with the harmonic amplitude feature in fault detection of motor. The proposed method was applied to a 1.5kW standard three phase induction motor using of different rotors that had various types of broken rotor bars. Experimental results confirmed the high efficiency of the proposed method for broken rotor fault detection in induction motors.

This paper offers a reliable solution to the detection of broken rotor bars in induction machines with a novel methodology, which is based on the fact that the fault-related harmonics will have oscillating amplitudes due to the speed ripple effect. The method consists of two main steps: Initially, a time-frequency transformation is used and the focus is given on the steady-state regime; thereupon, the fault-related frequencies are handled as periodical signals over time and the classical fast Fourier transform is used for the evaluation of their own spectral content. This leads to the discrimination of subcomponents related to the fault and to the evaluation of their amplitudes. The versatility of the proposed method relies on the fact that it reveals the aforementioned signatures to detect the fault, regardless of the spatial location of the broken rotor bars. Extensive finite element simulations on a 1.1 MW induction motor and experimental testing on a 1.1 kW induction motor lead to the conclusion that the method can be generalized on any type of induction motor independently from the size, power, number of poles, and rotor slot numbers.

Induction motor failures are mainly due to stator and rotor faults. In this paper, a novel method based on Park's vector approach for fault detection of induction motor is presented. Using suggested method it is possible to detect the various types of faults in squirrel induction motors. Further, paper focuses on the detection and discrimination of the simultaneous stator inter turn and broken rotor bar faults in induction motor. In order to evaluate the ability of the proposed method several experimental results are presented. The result demonstrates that the proposed method is effective and accurate and can be widely used in the induction motor fault detection.

Induction motors are subjected to thermal, electrical and mechanical stresses during continuous operation. If any of these stresses become excess than normal condition, it is a symptom of commencement of fault in the induction motor. If these faults are not detected at an incipient stage, it may result in the failure of the motor. Turn-to-turn short-circuit faults are the major causes for the stator winding insulation failure. In the proposed work, a novel approach is suggested to detect these faults in the induction motor. In this approach, the discrete wavelet transform (DWT) is performed on the Park’s vector modulus of current signals. The accuracy of DWT can be enhanced by selecting the proper wavelet type and its order along with levels of decomposition. With this context, an attempt is made to investigate the best-suited mother wavelet by testing various orthogonal wavelet functions on simulated signal and justified that the db44 is to be the best suitable wavelet for the detection of an inter-turn fault in the induction motor. The current signals are obtained experimentally under operating conditions with balanced supply voltages, unbalanced supply voltages, load slip variation and gradual variation in load condition and analyzed. The obtained results show the sensitivity and robustness of the proposed approach. The fault severity factor is also suggested for evaluating the severity of fault in the induction motor.

Fault diagnosis of anomalies in induction motors is essential to ensure industry safety. This paper presents a new hybrid Invasive Weed Optimization and Machine Learning approach for fault diagnosis in an induction motor. The vibration signal provides a lot of information about the motor's operating conditions. Therefore, the vibration signal of the motor was chosen to investigate the fault diagnosis. Two identical 400-V, 50-Hz, 4-pole 0.75 HP induction motors were under healthy, mechanical, and electrical faults tested in a laboratory with different loading. A hybrid model was developed using the vibration signal, the Invasive Weed Optimization algorithm (IWO), and machine learning classifiers. Some statistical features were extracted from the signal using Discrete Wavelet Transform (DWT). The invasive weed optimization algorithm (IWO) was utilized to reduce the number of the extracted features and select the most suitable ones. Then, three classification algorithms namely k-Nearest Neighbor neural network (KNN), Support Vector Machine (SVM), and Random Forest (RF), were trained using k-fold cross-validation and tested to predict the true class. The advantage of combining these techniques is to reduce the training time and increase the average accuracy of the model. The performance of the proposed fault diagnosis model was evaluated by measuring the Specificity, Accuracy, Precision, Recall, and F1_score. The experimental results prove that the proposed model has achieved more than 99.90% of accuracy. Furthermore, the other evaluation parameters also show the same representation of performance. The hybrid model has proved successfully its robust for diagnosing the faults under different load conditions.

Purpose: Developments in Industry 4.0 technologies and Artificial Intelligence (AI) have enabled data-driven manufacturing. Predictive maintenance (PdM) has therefore become the prominent approach for fault detection and diagnosis (FD/D) of induction motors (IMs). The maintenance and early FD/D of IMs are critical processes, considering that they constitute the main power source in the industrial production environment. Machine learning (ML) methods have enhanced the performance and reliability of PdM. Various deep learning (DL) based FD/D methods have emerged in recent years, providing automatic feature engineering and learning and thereby alleviating drawbacks of traditional ML based methods. This paper presents a comprehensive survey of ML and DL based FD/D methods of IMs that have emerged since 2015. An overview of the main DL architectures used for this purpose is also presented. A discussion of the recent trends is given as well as future directions for research.Design/methodology/approach: A comprehensive survey has been carried out through all available publication databases using related keywords. Classification of the reviewed works has been done according to the main ML and DL techniques and algorithmsFindings: DL based PdM methods have been mainly introduced and implemented for IM fault diagnosis in recent years. Novel DL FD/D methods are based on single DL techniques as well as hybrid techniques. DL methods have also been used for signal preprocessing and moreover, have been combined with traditional ML algorithms to enhance the FD/D performance in feature engineering. Publicly available datasets have been mostly used to test the performance of the developed methods, however industrial datasets should become available as well. Multi-agent system (MAS) based PdM employing ML classifiers has been explored. Several methods have investigated multiple IM faults, however, the presence of multiple faults occurring simultaneously has rarely been investigated.Originality/value: The paper presents a comprehensive review of the recent advances in PdM of IMs based on ML and DL methods that have emerged since 2015.

Diagnosis of static eccentricity (SE) fault for induction motors (IMs) is essential for the quality control of the machines, especially during their manufacturing process. Principal slot harmonic (PSH) type IMs have special combinations of rotor bar number and pole pair number, and it has been shown in previous works that conventional methods cannot effectively detect SE fault for these machines. Aiming at finding an effective approach for the SE fault detection for PSH-type IMs, this paper presents an analysis of SE-induced line currents in IMs with the higher-order harmonics of the air gap permeance considered. The analysis reveals that the second-order harmonic in the air gap permeance can induce SE-level-related signals in the line current of PSH-type three-phase IMs. The generation mechanism of the signature current signal is validated by simulations with an analytical IM model and a time-stepping finite element model. The signature signal in the motor's current discussed in this paper provides a new method for quantitative detection of SE fault for PSH-type IMs.

Induction motors, important elements into the industry, are susceptible to faults during its lifetime service; yet, they can keep working without affecting the process, but increasing the production costs as they consume more electrical current. Broken rotor bars (BRB) detection is an important topic due to the fact that this failure is silent and produces a power consumption increasing, vibration, or introduction of spurious frequencies in the electric line, among others. In this regard, a monitoring system that can efficiently diagnose the induction motor condition is highly required. In this work, a new methodology for one and two broken bars detection is presented. First, the compact kernel distribution (CKD) algorithm, a new high resolution time-frequency algorithm, is introduced for the detection of anomalies produced by the BRB failure in the startup current signal by considering that these signals describe changes on its dynamic characteristics due to the fault; then, the variance, a statistical feature, of the signal processed by CKD determines in automatic way the induction motor condition. The obtained results show a high overall efficiency for detecting broken rotor bars as well as healthy condition

This paper introduces a sophisticated approach for identifying and categorizing broken rotor bars in direct torque-controlled (DTC) induction motors. DTC is implemented in industrial drive systems as a suitable control method to preserve torque control performance, which sometimes shows its impact on fault-representing frequencies. This is because of the DTC’s closed-loop control nature, whichtriesto reduce speed and torque ripples by changing the voltage profile. The proposed model utilizes the modified Shapley Additive exPlanations (SHAP) technique in combination with gradient-boosting decision trees (GBDT) to detect and classify the abnormalities in BRBs at diverse (0%, 25%, 50%, 75%, and 100%) loading conditions. To prevent overfitting of the proposed model, we used the adaptive fold cross-validation (AF-CV) technique, which can dynamically adjust the number of folds during the optimization process. By employing extensive feature engineering in the original dataset and then applying Shapely Additive exPlanations(SHAP)-based feature selection, our methodology effectively identifies informative features from signals (three-phase current, three-phase voltage, torque, and speed) and motor characteristics. The gradient-boosting decision tree (GBDT) classifier, trained using the given characteristics, extracts consistent and reliable classification performance under different loading circumstances and enables precise and accurate detection and classification of broken rotor bars. The proposed approach (SHAP-Fusion GBDT with AF-CV) is a major advancement in the field of machine learning in detecting motor anomalies at varying loading conditions and proved to be an effective mechanism for preventative maintenance and preventing faults in DTC-controlled induction motors byattaining an accuracy rate of 99% for all loading conditions.

Detection of broken rotor bars in induction motors using nonlinear Kalman filters

Fault diagnosis of induction motor anomalies is vital for achieving industry safety. This paper proposes a new hybrid Machine Learning methodology for induction-motor fault detection. Some of the motor parameters such as the stator currents and vibration signals provide a great deal of information about the motor’s conditions. Therefore, these signals of the motor were selected to test the proposed model. The induction motor was assessed in a laboratory under healthy, mechanical, and electrical faults with different loadings. In this study a new hybrid model was developed using the collected signals, an optimal features selection mechanism is proposed, and machine learning classifiers were trained for fault classification. The procedure is to extract some statistical features from the raw signal using Matching Pursuit (MP) and Discrete Wavelet Transform (DWT). Then, the Invasive Weed Optimization algorithm (IWO)-based optimal subset was selected to reduce the data dimension and increase the average accuracy of the model. The optimal subset of features was fed into three classification algorithms: k-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Random Forest (RF), which were trained using k-fold cross-validation to distinguish between the induction motor faults. A similar strategy was performed by applying the Genetic Algorithm (GA) to compare with the performance of the proposed method. The suggested fault detection model’s performance was evaluated by calculating the Receiver Operation Characteristic (ROC) curve, Specificity, Accuracy, Precision, Recall, and F1 score. The experimental results have proved the superiority of IWO for selecting the discriminant features, which has achieved more than 99.7% accuracy. The proposed hybrid model has successfully proved its robustness for diagnosing the faults under different load conditions.

This paper deals with the problem of monitoring of induction motors (IM) through the development of fault detection and diagnosis (FDD) approach. The developed FDD technique is addressed such that, the principal component analysis (PCA) technique is used for features extraction purposes and the machine learning (ML) classifiers are applied for fault diagnosis. In the proposed FDD approach the most efficient features are extracted and selected through PCA scheme using induction motor data. Besides, their statistical characteristics (mean and variance) are also included. The ML classifiers are applied using the extracted and selected features to perform the FDD problem. The obtained results indicate that the proposed techniques have a wide application area, fast fault detection and diagnosis, making them more reliable for induction motors monitoring.

Broken rotor bar (BRB) fault is common in cage rotor induction motors. Motor current signature analysis (MCSA) has been a popular method to detect BRB faults. In the MCSA, the fault signature frequencies for BRB are close to the fundamental frequency. This spectral nearness leads to the requirement of larger observation windows, which inherently increases BRB fault detection time. This paper presents an alternative algorithm to detect BRB faults in induction motors from MCSA using two Taylor–Kalman (TK) filters in cascade with a subsampling scheme. The proposed BRB fault detection approach allows to use the TK filter to estimate lower frequencies with less computational burden in comparison to conventional TK analysis. Experimental analysis shows that nearly accurate estimates and competitive detection time can be achieved by the proposed BRB detection method. The performance of the proposed algorithm has been compared with classical spectral techniques using numerical simulations and records acquired from induction motors under real operating conditions.