Learning Vector Quantization Neural Network Based External Fault Diagnosis Model for Three Phase Induction Motor Using Current Signature Analysis

Chapter 4 - Intelligent Data Analytics for 3-Phase Induction Motor Fault Diagnosis Using Gene Expression Programming (GEP)

Diabetes Mellitus is a chronic, non-infectious disease that affects people’s health, which has a rapid onset trend with the continuous improvement of China’s economy and material life. At present, there is no practical solution to this condition. Diabetic retinopathy (DR) is the most important manifestation of diabetic microangiopathy, which can be divided into proliferative lesion and non-proliferative lesion. The traditional artificial diagnosis method has strong subjectivity and low accuracy. Because disease diagnosis can be regarded as a two-classification pattern recognition problem, the application of neural network method can provide the possibility for the application of artificial intelligence (AI) in disease-assisted diagnosis and treatment. Furthermore, it has significant meanings to find which kind of neural network has a better efficiency. In this paper, learning vector quantization (LVQ) neural network and back propagation (BP) neural network were used to diagnose diabetes and diabetic retinopathy with MATLAB and their recognition rate were compared. The datasets of diabetes and diabetic retinopathy were available in the UCI database. The experiment results were analyzed to evaluate the efficiency of each neural network classifier. The results demonstrate that both LVQ neural network and BP neural network can classify the two datasets effectively. However, compared with the LVQ neural network, the average accuracy rate and sensitivity of the BP neural network is higher.

This paper presents the implementation of TMS320LF2407 DSP processor Kit to identify the single phasing fault online, based on the stator current. The Induction motors are most widely used motors in industrial, commercial and residential sectors because of enormous merits of these over the types of available electric motors, as the workhorse in industrial applications. The early detection of these deteriorating conditions in incipient phase and its removal is necessary for the prevention of external failure of the induction motors, reducing repairs costs and motor outage time. The various types of external faults in induction motor are single line-to-ground fault, double line-ground fault, line-line fault, single phasing fault, locked rotor, under voltage, overvoltage, phase sequence reversal of supply voltage and mechanical overload. This proposed method is applied to a 3.3KW three phase induction motor using TMS320LF2407 DSP processor Kit to identify the single phasing fault online, based on the stator current and then the measurement results are analyzed.

In this paper, comparisons are made on various parameters of electromagnetic fields like air gap flux density (radial), flux distribution, phase currents, speed and torque of a healthy and stator inter-turn fault induction motor (IM). Using FEM (Finite Element Method), the inter-turn short circuit fault analysis of PWM (Pulse Width Modulated) inverter fed 3 phase IM is analyzed. The inter-turn fault is the initial stage of a short circuit in the stator winding and it progresses to a phase to phase fault or phase to ground fault if left undetected. When the IM is directly fed from the mains, the inter-turn fault can be identified using a popular technique known as MCSA (Motor Current Signature Analysis). But this method fails when the IM is fed from a PWM inverter due to rich noise content observed in the current spectrum due to high frequency inverter switching making it difficult to detect the fault. The electromagnetic field analysis of PWM inverter fed 3 phase IM with 12.5% and 25% stator winding inter-turn fault is carried out on a 7.5kW IM using ANSYS Maxwell FEA (Finite Element Analysis) tool. The spatial FFT spectrum of air gap flux density (radial) indicates the presence of stator winding inter-turn fault with rich noise content compared to healthy motor.

The industrial sector relies heavily on induction motors. Because of the age of the machine in service, fault diagnosis and condition assessment (FDCA) of rotating machines becomes critical. The use of conventional methods generally suffer from inaccurate Induction Motor external faults identification. Condition monitoring of the motor are hence important to identify the faults and take precautions before the operation of induction motor get shut down. The proposed method adopts motor electrical signature analysis (MESA), a non-invasive method for the external fault identification and feature extraction. Multilabel classification of external faults was performed using various machine learning models. Ensemble bagged tree and Support Vector Machine (SVM) are identified as the most accurate classification model for data set with and without noise.

A system has built to monitor conditions of a 3-phase Induction motor. The purpose of the system is to detect fault when the Induction motor is being operated. To build a fault detection system various methods have been used in this work. A prototype setup has built to test and evaluate the results of the developed system. The first method used to develop the system was through current monitoring where sensors used to monitor the current drawn by the 3-phase Induction Motor while operating on non-faulty and faulty conditions. Vibration sensor was used to monitor vibrations from rotating shaft on no-fault and fault conditions and Vibration Energy Harvester setup was made and equipped with the system. Vibration Energy Harvester (VEH) was acquiring vibration from the rotating shaft to generate signals to feed into the system. Results were obtained from the methods used to monitor the 3-phase Induction Motor during both no fault and fault conditions. Appropriate threshold values set to all methods used to monitor condition of the 3-hase induction motor. The results obtained shows that all three methods can be used to monitor condition of 3 -phase induction motor but the stability and reliability of the results obtained differs from one another.

Induction motor is the most motor applicable in the world. especially three phase induction motor. Furthermore 95% of induction motor are used as propulsion in industry. However, maintenance costs are large enough more than 50% of operating costs of the industry or the factory. To reduce these costs is to detect early faults in induction motors. One way to detect initial faults is to measure and analyze the temperature of the induction motor. In this research, the authors design temperature monitoring of a three-phase induction motor using the internet of things to make it easier for users to monitor the temperature of an existing three-phase induction motor and to schedule maintenance if temperature reach the operating limit. The field system elements consist of SoC ESP32 to collect data from 2D temperature, to drive lead screw and to send data to cloud. The SoC ESP 32 is connected to the AMG8833 temperature sensor array which collect the frame of temperature of a three-phase induction motor in the form of an 8x8 array. In addition, the SoC ESP32 is also connected to the motor driver L293D to regulate the movement of geared DC motor. Geared DC motor is used to drive the lead screw which already has a mount for placing the AMG8833. From the test on 18.5 kW, 360V induction motor installed in chemical industry, it is obtained the sensor can measure temperature accurately and it travel smoothly above frame. In the storage using MongoDB database, the data transmission fulfilled the QoS requirement.
 

A learning vector quantization(LVQ) neural network classifier is for the first time applied for SAR data classification, of which the training termination rule is modified to make it have both the ability of classification and class signature analysis. A very high land cover classification accuracy is achieved. Especially under the condition that texture is considered, almost all roads can be classified correctly, which cannot be identified by BP MLP neural network and ML classifier.

An attempt has been made to analyze the methods to improve the electromagnetic properties of a solid-rotor induction motor. The slip of the solid rotor is reduced notably if the solid rotor is axially slitted. The slitting patterns of the solid rotor are examined. It is shown how the slitting parameters affect the produced torque. Finite element analysis is used to numerically analyze the effect of axial slitted rotor on the electromagnetic performance characteristics of a 180 kW, 170 Hz 3-phase induction motor. The paper considered the effect of variation of slit depth, variation of no of rotor slits and the variation of air gap length on the performance of 3-phase induction motor using Flux-2D Maxwell software package from ANSYS. It is found that larger and deeper slit of the rotor with minimum valve of air gap gives the better and higher operating torque. It is also found that deeper slit gives the higher efficiency of the slitted rotor, high speed induction motor.

The use of an induction motor requires the process of stopping the motor speed quickly, using mechanically and electrically generated braking torque. Dynamic braking, which is done by making a magnetic field stationary motor. This condition is carried out by injecting a DC current into the three phase induction motor stator coil after the connection of the stator coil is released from the AC supply voltage source. The advantages of dynamic braking include ease of speed regulation of three-phase induction motors and mechanical losses can be reduced. A system for reversing the direction of star-delta three phase induction motor rotation with Programmable Logic Control (PLC) as a controller. requires an exchange of rotating directions from the direction of turning right to turn left. The process of changing the direction of rotation of the motor requires dynamic braking, the braking process is carried out by injecting a DC current into the stator coil of an induction motor, which is related to PLC. By applying dynamic braking on a three phase induction motor it is obtained faster than without dynamic braking. The injection current is given to a 3-phase induction motor of 1 KW for dynamic braking of 2.2 A. The timing of the motor stops for a minimum load of 100 Watt without braking at 6.2 seconds and with braking of 4.4 seconds, for a maximum load of 600 W, without braking.

Standard back propagation (BP) neural network has disadvantages such as slow convergence speed, local minimum and difficulty in definition of network structure. In this paper, an learning vector quantization (LVQ) neural network classifier is established, then it is applied in pattern classification of two-dimensional vectors on a plane. To test its classification ability, the classification results of LVQ neural network and BP neural network are compared with each other. The simulation result shows that compared with classification method based on BP neural network, the one based on LVQ neural network has a shorter learning time. Besides, its requirements for learning samples and the number of competing layers are also lower. Therefore it is an effective classification method which is powerful in classification of two-dimensional vectors on a plane.

The design of pattern classifiers such as multiprototype classifiers and neural network classifiers such as learning vector quantization and radial basis function neural networks requires reducing the size of the training data sets. In addition, memory storage, computation complexity and time, and data redundancy demand many pattern classifiers to use a smaller subset of a training data set. In this paper, we present a data reduction algorithm which automatically selects the subset of training data that faithfully represents the training data set for pattern classification. The applicability of this algorithm is demonstrated through k-nearest neighbor and learning vector quantization neural networks classifiers using both speech and synthetic data sets.

The Raman spectra of tissue of 20 brain tumor patients was recorded using a confocal microlaser Raman spectroscope with 785 nm excitation in vitro. A total of 133 spectra were investigated. Spectra peaks from normal white matter tissue and tumor tissue were analyzed. Algorithms, such as principal component analysis, linear discriminant analysis, and the support vector machine, are commonly used to analyze spectral data. However, in this study, we employed the learning vector quantization (LVQ) neural network, which is typically used for pattern recognition. By applying the proposed method, a normal diagnosis accuracy of 85.7% and a glioma diagnosis accuracy of 89.5% were achieved. The LVQ neural network is a recent approach to excavating Raman spectra information. Moreover, it is fast and convenient, does not require the spectra peak counterpart, and achieves a relatively high accuracy. It can be used in brain tumor prognostics and in helping to optimize the cutting margins of gliomas.

Tractor condition recognition has important research value in helping to understand the operating status of tractors and the trend of tillage depth changes in the field. Therefore, this article presents a method for recognizing tractor conditions, providing the basis for establishing the relationship between tractor conditions and the tillage depth of the attached agricultural machinery. This study designed a tractor condition recognition method based on neural networks. Using real-world vehicle data to establish a data set, K-means clustering analysis was used to label the data set based on four conditions: “accelerated start”, “constant speed”, “decelerated stop” and “turning”. The learning vector quantization (LVQ) neural network and the VGG-16 model of a CNN were selected for use recognizing the tractor conditions. The results showed that both the neural networks had good recognition effects. The average accuracy rates of the VGG-16 model of CNN and LVQ neural network were 90.25% and 79.7%, respectively, indicating that these models could be applied to tractor condition recognition and provide theoretical support for the correction of angle detection errors.

Transformer inrush currents are high magnitude, harmonic-rich currents generated when transformer cores are driven into saturation during energization. In this paper an efficient method for detection of inrush current in distribution transformer based on wavelet transform is presented. Using this method inrush current can be discriminate from other transients such as capacitor switching, load switching and single phase to ground fault. Wavelet transform is used for decomposition of signals and Learning Vector Quantizer(LVQ) neural network used for classification. Inrush current data and other transients are obtained by simulation using EMTP program. Results show that the proposed procedure is efficient in identifying inrush current from other events.