The use of features selection and nearest neighbors rule for faults diagnostic in induction motors

This paper deals with the diagnosis of induction motors by pattern recognition methods. The objective is to use existing theories to improve the diagnosis procedures in electrical engineering. First of all, a single signature is determined to monitor several different operating modes. For this purpose, features are extracted from the combination of the stator currents and voltages. Then, the sequential backward algorithm is applied in order to select the most relevant features. The classification is performed by the k-nearest neighbors rule with reject options. The methodology is applied on a 5.5 kW motor in normal conditions, then with stator and rotor faults. The experimental results prove the efficiency of pattern recognition methods in condition monitoring of electrical machines.     

A fault detection scheme based on minimum identified uncertainty bounds violation for broken rotor bars in induction motors

In this paper, a novel method for broken bars fault detection in the case of three-phase induction motors and under different payloads will be presented and experimentally evaluated. In the presented approach, the cases of a partially or full broken rotor bars are being also considered, caused by: (a) drilling 4 mm and 8 mm out of the 17 mm thickness of the same rotor bar and (b) fully drilled (17 mm) one, two and three broken bars. The proposed fault detection method is based on the Set Membership Identification (SMI) technique and a novel proposed minimum boundary violation fault detection scheme, applied on the identified motor's parameters. The system identification procedure is being carried out on the simplified equivalent model of the induction motor, during the steady-state operation (non-fault case), while at the same time the proposed scheme is able to calculate on-line the corresponding safety bounds for the identified variables, based on a priori knowledge of the measuring corrupting noise (worst case encountered). The efficiency, the robustness and the overall performance of the established fault detection scheme is being extensively evaluated in multiple experimental studies and under various time instances of faults and load conditions.     

An efficient,simplified multiple-coupled circuit model of the induction motor aimed to simulate different types of stator faults

This paper proposes an original simplified model aimed to simulate, an easy way, inter turns short circuit fault, phase to phase fault and phase to ground fault. In this model, the stator is considered as six magnetically coupled windings and the rotor as three not magnetically coupled R–L circuits. The paper also presents the star- and delta-connected stator configurations of the simplified model. However, the proposed simplified model is suitable only for steady-state operation. The performance of the simplified model is first verified by a comparison between the simulated current of the multiple-coupled model and the simplified model. Then, since the stator faults have an impact on the symmetrical components of the stator current, this paper uses these components to validate the behavior of the simplified model by simulation and experimentally using a 1.1 kW motor. In addition, simulated results of the simplified model for a 110 kW motor are presented in order to generalize the use of the proposed model to larger motors.     

Offline and online fault detection and diagnosis of induction motors using a hybrid soft computing model

In this paper, a hybrid soft computing model comprising the Fuzzy Min-Max (FMM) neural network and the Classification and Regression Tree (CART) for motor fault detection and diagnosis is described. Specifically, the hybrid model, known as FMM-CART, is used to detect and classify fault conditions of induction motors in both offline and online environments. A series of experiments is conducted, whereby the Motor Current Signature Analysis (MCSA) method is applied to form a database containing stator current signatures under different motor conditions. The signal harmonics from the power spectral density (PSD) are extracted, and used as the discriminative input features for fault classification with FMM-CART. Three main induction motor conditions, viz. broken rotor bars, stator winding faults, and unbalanced supply, are used to evaluate the effectiveness of FMM-CART. The results indicate that FMM-CART is able to detect motor faults in the early stage, in order to avoid further damage to the induction motor as well as the overall machine or system that uses the motor in its operations.     

Accurate diagnosis of induction machine faults using optimal time–frequency representations

This paper presents a new diagnosis method of induction motor faults based on time–frequency classification of the current waveforms. This method is composed of two sequential processes: a feature extraction and a rule decision. In the process of feature extraction, the time–frequency representation (TFR) has been designed for maximizing the separability between classes representing different faults. The diagnosis is realised in two levels; the first one allows the detection of different faults—bearing fault, stator fault and rotor fault. The second one refines this detection by the determination of severity degree of faults, which are already identified on the previous level. The diagnosis is independent of the level of load. This method is validated on a 5.5 kW induction motor test bench.     

An adaptive artificial immune system for fault classification

Fault diagnosis is very important in ensuring safe and reliable operation in manufacturing systems. This paper presents an adaptive artificial immune classification approach for diagnosis of induction motor faults. The proposed algorithm uses memory cells tuned using the magnitude of the standard deviation obtained with average affinity variation in each generation. The algorithm consists of three steps. First, three-phase induction motor currents are measured with three current sensors and transferred to a computer by means of a data acquisition board. Then feature patterns are obtained to identify the fault using current signals. Second, the fault related features are extracted from three-phase currents. Finally, an adaptive artificial immune system (AAIS) is applied to detect the broken rotor bar and stator faults. The proposed method was experimentally implemented on a 0.37?kW induction motor, and the experimental results show the applicability and effectiveness of the proposed method to the diagnosis of broken bar and stator faults in induction motors.     

Evaluation of stator winding faults severity in inverter-fed induction motors

Three-phase induction motor are one of the most important elements of electromechanical energy conversion in the production process. However, they are subject to inherent faults or failures under operating conditions. The purpose of this paper is to present a comparative study among intelligent tools to classify short-circuit faults in stator windings of induction motors operating with three different models of frequency inverters. This is performed by analyzing the amplitude of the stator current signal in the time domain, using a dynamic acquisition rate according to machine frequency supply. To assess the classification accuracy across the various levels of faults severity, the performance of three different learning machine techniques were compared: (i) fuzzy ARTMAP network; (ii) multilayer perceptron network; and (iii) support vector machine. Results obtained from 2.268 experimental tests are presented to validate the study, which considered a wide range of operating frequencies and load conditions.     

Real-time SVD-based detection of multiple combined faults in induction motors

Early detection of induction motor faults has been a main subject of investigation for many years. Several approaches have been proposed for identifying one or more faults treated in an isolated way. Multiple combined faults on induction motors represent a big challenge since the reliable diagnosis of a faulty condition under the presence of two or more simultaneous faults is really difficult. This work introduces a novel methodology that merges singular value decomposition, statistical analysis, and artificial neural networks for multiple combined fault identification. Obtained results demonstrate its high effectiveness on detecting faulty bearings, unbalance, broken rotor bars, and all their possible combinations. The developed field programmable gate array-based implementation offers a portable low-cost solution for online classification of the rotating machine condition in real time. Thanks to its generalized nature, the introduced approach can be extended for detecting multiple combined faults under different working conditions by a proper calibration.     

定子电阻辨识用于异步电机温度监测的研究

定子绕组温度监测是电机安全运行的重要保障.利用定子电阻值与温度的密切关系,提出了一种定子电阻估计方法以反映绕组温度的变化.在电流定向的坐标系中,异步电机数学模型中与定子电阻相关的部分得到简化.基于异步电机稳态模型,得到了独立于定、转子电阻和转速的稳态磁链估计,然后利用稳态磁链建立了定子电阻在线整定律.该整定律只需量测定子电流电压,简单易行.所提出的定子温度监测方法可以作为电机过热保护的一种辅助措施.     

Multi-class fault diagnosis of induction motor using Hilbert and Wavelet Transform

The information extraction capability of two widely used signal processing tools, Hilbert Transform (HT) and Wavelet Transform (WT), is investigated to develop a multi-class fault diagnosis scheme for induction motor using radial vibration signals. The vibration signals are associated with unique predominant frequency components and instantaneous amplitudes depending on the motor condition. Using good systematic and analytical approach this fault frequencies can be identified. However, some faults either electrical or mechanical in nature are associated with same or similar vibration frequencies leading to erroneous conclusions. Genetic Algorithm (GA) is proposed and used successfully to find the most relevant fault frequencies in radial (vertical) frame vibration signal which can be used to diagnose the induction motor faults very effectively even in the presence of noise. The information obtained by Continuous Wavelet Transform (CWT) was found to be highly redundant compared to HT and thus by selecting the most relevant features using GA, the fault classification accuracy has considerably improved especially for CWT. Almost similar fault frequencies were found using CWT + GA and HT + GA for radial vibration signal.     

Simplified detection of broken rotor bars in induction motors controlled in field reference frame

In this paper, a simple method for additional on-line detection of broken rotor bars in a squirrel cage induction motor controlled in rotor field co-ordinates using existing hardware is presented. Based on a previously presented approach, an algorithm for on-line calculation of the variance of stator voltage reference, which depends on the number of broken bars, has been developed. Due to its simplicity, it could run in parallel with a standard control algorithm in field reference frame using contemporary fixed- and floating-point processors, thus requiring minimum processing time. The algorithm uses internal reference values of the stator voltage; therefore no additional dedicated measurements are needed. Results were obtained at different operating points on an induction motor with gradually damaged rotor. Comparison with commonly used diagnostic method confirms the validity of the approach.     

A new flextensional piezoelectric ultrasonic motor—Design,fabrication and characterisation

This paper presents the techniques used for the characterisation of a new type of standing-wave piezoelectric ultrasonic motor. The motor uses a metallic flextensional amplifier, or “cymbal”, to convert the radial mode vibrations of a piezoelectric ceramic disc into flexural oscillations, which are further converted to produce rotary actuation by means of an elastic fin friction drive. The motor operates on a single-phase electrical supply. A beryllium copper rotor design with three-fin configuration was adopted. The best stall torque, no load speed, transient time and efficiency for a 25 mm motor were 2 N mm, 680 rpm, 2 ms and 4.8%, respectively. The operational characteristics of the motor were evaluated by using two methods: one based on the pulley–brake principle and one on high-speed imaging. The results obtained from using these two techniques are contrasted and compared.     

Fault-tolerant cruise control of electric vehicles with induction motors

A fault-tolerant control scheme is proposed for the cruise control of electric vehicles (trains, cars) that make use of induction motors. It relies on a rotor speed reference generator and on a flux observer which is adaptive with respect to the uncertain rotor and stator resistances and to the load torque as well. The closed loop on-line identification of those three critical uncertain parameters allows for: (i) on-line estimating and imposing the motor flux modulus reference value which minimizes power losses at steady-state and improves power efficiency; (ii) the on-line detection of speed sensor faults as well as the fast switching on redundant motor speed sensors. CarSim simulations illustrate the effectiveness of the proposed approach.     

Least squares and genetic algorithms for parameter identification of induction motors

This paper deals with off-line parameter identification of induction motors by means of least square (LS) techniques and genetic algorithms (GA), using stator voltages, stator currents and velocity as input–output data. For analytical identification by LS algorithms, filtering of experimental data is performed by means of anticausal filters. Two models useful for identification are derived in which the products of acceleration and rotor fluxes, usually neglected, are taken into account. The GA-based identification method consists of the determination of the best parameters which match input–output behaviour of the motor. Both methods are investigated and compared by means of experiments carried out on a 1-kW induction motor.     

Automatic detection and classification of rotor cage faults in squirrel cage induction motor

The detection of broken rotor bars and broken end-ring in three-phase squirrel cage induction motors by means of improved decision structure. The structure consists of current signal analysis (CSA), Artificial Neural Network (ANN) and diagnosis algorithm. Effects of broken bars and end-ring on current signal and feature extraction are in the CSA. The rotor cage faults are classified by using ANN. And result matrixes of ANN are considered two different ways for diagnosis. Then the diagnoses are compared with each other. In this study six different rotor faults, which are one, two, three broken bars, bar with high resistance, broken end-ring and healthy rotor, are investigated. The effects of different rotor faults on current spectrum, in comparison with other fault conditions, are investigated by analyzing side-bands in current spectrum. To reduce bad effects of changing of distance between the side-band and main component on the detection and classification of the faults, the spectrum is achieved with low definition. Thus, the improved decision structure diagnoses faulted rotors with 100% accuracy and classified rotor faults 98.33% accuracy.     

A nonlinear tracking control for sensorless induction motors

We propose a novel tracking control for induction motors in which only stator currents are used for feedback. Local exponential rotor speed and flux modulus tracking are achieved for any constant reference value and for restricted time-varying reference signals; any known motor parameters values (including constant load torque) and any initial condition, including rotor speed and fluxes, belonging to an explicitly computed domain of attraction are allowed.     

Neural networks for load torque monitoring of an induction motor

In this paper, the possibility to use neural networks for the monitoring of the load torque of induction motors is investigated. In particular, unsupervised neural networks are used to detect possible torque anomalies and supervised neural networks are used to identify the average value of steady-state load torque. These networks are trained and validated on the data gathered from a 1.5 kW three-phase squirrel-cage induction motor. Their generalisation abilities have been tested through the data collected with a 3 kW induction motor.     

Self-organizing map approach for classification of mechanical and rotor faults on induction motors

Two neural network-based schemes for fault diagnosis and identification on induction motors are presented in this paper. Fault identification is performed using self-organizing maps neural networks. The first scheme uses the information of the motor phase current for feeding the network, in order to perform the diagnosis of load unbalance and shaft misalignment faults. The network is trained using data generated through the simulation of a motor-load system model, which allows including the effects of load unbalance and shaft misalignment. The second scheme is based on the motor’s active and reactive instantaneous powers, in order to detect and diagnose faults whose characteristic frequencies are very close each other, such as broken rotor bars and oscillating loads. This network is trained using data obtained through the experimental measurements. Additional experimental data are later applied to both networks in order to validate the proposal. It is demonstrated that the proposed strategies are able to correctly identify, both unbalanced and misaligned load, as well as broken bars and low-frequency oscillating loads, thus avoiding the need for an expert to perform the task.     

A modified direct torque control with fault tolerance

The use of inverters in induction motor control has reduced classical motor faults, such as broken rotor bars or windings short-circuit, besides improving control performance. The control becomes faster and more precise, reducing peaks in current and torque, so that the motor can have a softer operation. On the other hand, new elements are included in the system and it will be necessary to take into account their faults. These elements are sensors and power electronic devices that since a control point of view are the system sensors and actuators. Fault tolerance tries to maintain the system under control in case a fault appears in the system. If this is not possible, it takes the system to a safe operational point. In this paper a fault-tolerant control for induction motors is designed. Based on a direct torque control, new control strategies have been added in case current sensor and power switch faults are detected. The challenge is to overcome these faults without any physical redundancy of sensors or power switches as other authors propose. With the proposed control, it will be possible to guarantee the motor operation in the whole speed-torque range with one or none current sensors instead of the two usually used, though the performance will be slightly worsened. In case of inverter faults, the operation range will be restricted but the performance with respect to the fault situation is improved.     

Diagnosis of broken rotor bars in induction motors based on harmonic analysis of fault components using modified adaptive notch filter and discrete wavelet transform

This paper proposes a method for fast and accurate detection of broken rotor bars (BRBs) in a three-phase squirrel cage induction motor. The fundamental component of the stator current signal is extracted using a linear time-invariant filter. The resultant residual signal which contains the harmonic components of the current is then used to detect the BRBs, by means of discrete wavelet transform (DWT). Since in experiment it is not possible to break the rotor bars while the motor is under load, finite element method and MATLAB/Simulink are employed to accurately demonstrate the behavior of the running machine as the BRB happens. To get more accuracy, differential evolution (DE) optimization algorithm is used to obtain the corresponding fault impedance for the rotor external circuit of the MATLAB model. Detail coefficients (DCs) of the wavelet decomposition are employed as the new fault indicators. Simulation results show that using DCs of the harmonic component signal rather than the actual current signal, leads to more distinctive fault signatures in the wavelet decomposition. The obtained results suggest that the proposed fault detection scheme can be employed as a highly reliable technique for diagnosing rotor bar failures in running machines.