Fault diagnosis for induction generator-based wind turbine using ensemble deep learning techniques

Most cases of electrical machines are connected directly to the low voltage power system without inverter, which can be subjected to a harmonic pollution at any time. The presence of these harmonics can compromise the reliability of the fault detection. In this paper, we present the winding fault diagnosis of induction machine connected to a harmonically polluted low voltage power system. The detection method is based on the Hilbert transform. This proposed approach allows the suppression of the harmonics from the network while keeping all possible clues to the presence of winding fault in the induction machine. The proposed approach is tested experimentally to an induction machine subjected to a 2.5% of inter-turns short circuit fault.

The aim of this paper is to introduce a new methodology for the fault diagnosis of induction machines working in the transient regime, when time-frequency analysis tools are used. The proposed method relies on the use of the optimized Slepian window for performing the short time Fourier transform (STFT) of the stator current signal. It is shown that for a given sequence length of finite duration, the Slepian window has the maximum concentration of energy, greater than can be reached with a gated Gaussian window, which is usually used as the analysis window. In this paper, the use and optimization of the Slepian window for fault diagnosis of induction machines is theoretically introduced and experimentally validated through the test of a 3.15-MW induction motor with broken bars during the start-up transient. The theoretical analysis and the experimental results show that the use of the Slepian window can highlight the fault components in the current’s spectrogram with a significant reduction of the required computational resources.

Induction machine model with space harmonics for fault diagnosis based on the convolution theorem

The development of new methods for the fault diagnosis of induction machines (IMs) require fast models of the machine under fault conditions that are able to reproduce the harmonics generated by different kinds of faults. These models must take into account the space harmonics of the air gap magnetomotive force (MMF) generated by the machine windings, due to the complex interactions between spatial and time harmonics in a faulty machine. A common fault in induction machines is the rotor eccentricity, which can cause significant radial forces, and destructive rotor-stator rub. On the one hand, FEM models can simulate the eccentric machine, but, due to the asymmetry of the air gap length, they require high computational resources. On the other hand, analytical models of the eccentric machine are very fast, but the computation of the inductances needed in analytical models is very complex, due to the non-uniformity of the air gap. In this paper, a new method is proposed to obtain the inductances of an eccentric induction machine. It is based on the combination of a conformal mapping, which transforms the eccentric machine into a non-eccentric one, followed by a convolution that is able to extract from this conformal mapping model the inductances of the eccentric machine. The proposed method is theoretically justified, and is validated using a FEM model.

Condition monitoring of induction machines is crucial to prevent costly interruptions and equipment failure. Mechanical faults such as misalignment and rotor issues are among the most common problems encountered in industrial environments. To effectively monitor and detect these faults, a variety of sensors, including accelerometers, current sensors, temperature sensors, and microphones, are employed in the field. As a non-contact alternative, thermal imaging offers a powerful monitoring solution by capturing temperature variations in machines with thermal cameras. In this study, we propose using 2dimensional Self-Organized Operational Neural Networks (SelfONNs) to diagnose misalignment and broken rotor faults from thermal images of squirrel-cage induction motors. We evaluate our approach by benchmarking its performance against widely used Convolutional Neural Networks (CNNs), including ResNet, EfficientNet, PP-LCNet, SEMNASNet, and MixNet, using a Workswell InfraRed Camera (WIC). Our results demonstrate that Self-ONNs, with their non-linear neurons and self-organizing capability, achieve diagnostic performance comparable to more complex CNN models while utilizing a shallower architecture with just three operational layers. Its streamlined architecture ensures high performance and is well-suited for deployment on edge devices, enabling its use also in more complex multi-function and/or multi-device monitoring systems.

Motor Current Signature Analysis has been successfully used for fault diagnosis in induction machines. However this method does not always achieve good results with not constant load torque. This paper proposes a different signal processing method, by combination of Wavelet and Power Spectral Density techniques. It presents good theoretical and experimental results.

A vibration analysis based interturn fault diagnosis of induction machines is proposed in this paper, using a neural-network-based scheme, constituting of two parts. The first part finds out the optimum network size of the probabilistic neural network (PNN) using the Orthogonal Least Squares Regression algorithm. This judges the size of the PNN, with an effort to reduce the computation. The feature extraction to model the PNN is made meaningful using dual tree complex wavelet transform (DTCWT), which is nearly shift invariant analytical wavelet transform, giving a true representation of the input space. In the second part, preprocessing using principal component analysis is suggested as an effective way to further reduce the dimension of the feature set and size of the PNN without compromising the performance. The sensitivity, specificity, and accuracy show that the vibration signatures capture the fault more effectively (especially by the axial and radial ones), under varying supply-frequency and load conditions. A comparison with traditional discrete wavelet transform proves the applicability of the proposed scheme. A comparative evaluation with feedforward neural network and naïve Bayes scheme brings out the advantage of the proposed optimized DTCWT-PNN based technique over other machine learning approaches.

The inter-turn fault is a common fault of oil-immersed transformers. Due to the deterioration of inter-turn insulation, its resistance decreases, which leads to an increase in the winding coil current and loss, causing the winding temperature to rise. Therefore, analyzing the electrothermal characteristics of inter-turn short circuit faults in oil-immersed transformers can provide a theoretical basis for the diagnosis of inter-turn faults in transformers. This paper studies the equivalent circuit under inter-turn faults and establishes a field–circuit coupling model to calculate the current characteristics under different inter-turn short circuit faults of transformers. The distribution of coil losses under faults is calculated as thermal source excitations, and a thermal–fluid-coupled calculation model for oil-immersed transformers is established. After the experimental verification, it is found that both the errors between the winding and tank surface temperatures do not exceed the acceptable limits, validating the effectiveness of this model. Using this model to calculate the temperature distribution for various inter-turn short circuit faults in transformers reveals their hot-spot characteristics. The results indicate that under inter-turn fault conditions, the temperature at faulty turns will gradually increase towards a limit, eventually leading to coil damage. This research provides theoretical support for enhancing transformer detection and diagnostic capabilities.

This paper presents a performance comparison among some of the most effective spectral estimation techniques applied to the fault diagnosis of induction machines. The diagnostic test is based on the analysis of the current space vector during motor starting via short-time analysis, using a sliding window and different spectral estimation algorithms. Differently from most of the diagnostic techniques already proposed in the technical literature, the approach, presented in this work, is effective regardless of the load condition of the machine. Algorithms based on the FFT or optimal band-pass filters (nonparametric methods), on the estimation of a linear time-invariant model of the signal (parametric methods), and on the eigenanalysis of the autocorrelation matrix (high-resolution methods) have been used to process the motor current space-vector. Experiments prove that both parametric and high-resolution methods overcome the FFT-based approaches, keep only the principal frequency components of the signal and decrease the noise influence, thus permitting a better interpretation of the current vector spectrum and an automatic fault detection procedure.

Approximately 35% of all induction motor defects are caused by stator inter-turn faults. In this paper a novel algorithm has been proposed to analyze the three-phase stator current signals captured from the motor while it is in operation. The suggested method seeks to identify stator inter-turn short circuit faults in early stage and take the appropriate action to prevent the motor's condition from getting worse. Three-phase current signals have been captured under healthy and faulty conditions of the motor. Involving discrete wavelet transform (DWT) based decomposition followed by reconstruction using inverse DWT (IDWT), 50 Hz fundamental component has been removed from the captured raw current signals. Subsequently, from each phase current 15 statistical parameters have been retrieved. The statistical parameters include mean, standard deviation, skewness, kurtosis, peak-to-peak, root mean square (RMS), energy, crest factor, form factor, impulse factor, and margin factor. At the end, a standard machine learning algorithm namely error correcting output codes-support vector machine (ECOC-SVM) has been employed to classify six different severity of stator winding faults. The proposed fault diagnosis method is load and motor-rating independent.

Deep learning models, especially deep convolutional neural networks (CNNs), have been intensively investigated for hyperspectral image (HSI) classification due to their powerful feature extraction ability. In the same manner, ensemble-based learning systems have demonstrated high potential to effectively perform supervised classification. In order to boost the performance of deep learning-based HSI classification, the idea of deep learning ensemble framework is proposed here, which is loosely based on the integration of deep learning model and random subspace-based ensemble learning. Specifically, two deep learning ensemble-based classification methods (i.e., CNN ensemble and deep residual network ensemble) are proposed. CNNs or deep residual networks are used as individual classifiers and random subspaces contribute to diversify the ensemble system in a simple yet effective manner. Moreover, to further improve the classification accuracy, transfer learning is investigated in this study to transfer the learnt weights from one individual classifier to another (i.e., CNNs). This mechanism speeds up the learning stage. Experimental results with widely used hyperspectral datasets indicate that the proposed deep learning ensemble system provides competitive results compared with state-of-the-art methods in terms of classification accuracy. The combination of deep learning and ensemble learning provides a significant potential for reliable HSI classification.

With the development of machine learning in recent years, the application of machine learning to machine fault diagnosis has become increasingly popular. Applying traditional feature extraction methods for complex systems will weaken the characterization capacity of features, which are not conducive to subsequent classification work. A reciprocating compressor is a complex system. In order to improve the fault diagnosis accuracy of complex systems, this paper does not use traditional fault diagnosis methods and applies deep convolutional neural networks (CNNs) to process this nonlinear and non-stationary fault signal. The valve fault data is obtained from the reciprocating compressor test bench of the Daqing Natural Gas Company. Firstly, the single-channel vibration signal is collected on the reciprocating compressor and the one-dimensional CNN (1-D CNN) is used for fault diagnosis and compared with the traditional model to verify the effectiveness of the 1-D CNN. Next, the collected eight channels signals (three channels of vibration signals, four channels of pressure signals, one channel key phase signal) are applied by 1-D CNN and 2-D CNN for fault diagnosis to verify the CNN that it is still suitable for multi-channel signal processing. Finally, further study on the influence of the input of different channel signal combinations on the model diagnosis accuracy is carried out. Experiments show that the seven-channel signal (three-channel vibration signal, four-channel pressure signal) with the key phase signal removed has the highest diagnostic accuracy in the 2-D CNN. Therefore, proper deletion of useless channels can not only speed up network operations but also improve diagnosis accuracy.

With the development of information and technology, especially with the boom in big data, healthcare support systems are becoming much better. Patient data can be collected, retrieved, and stored in real time. These data are valuable and meaningful for monitoring, diagnosing, and further applications in data analysis and decision-making. Essentially, the data can be divided into three types, namely, statistical, image-based, and sequential data. Each type has a different method of retrieval, processing, and deployment. Additionally, the application of machine learning (ML) and deep learning (DL) in healthcare support systems is growing more rapidly than ever. Numerous high-performance architectures are proposed to optimize decision-making. As reliability and stability are the most important factors in the healthcare support system, enhancing the predicted performance and maintaining the stability of the model are always the top priority. The main idea of our study comes from ensemble techniques. Numerous studies and data science competitions show that by combining several weak models into one, ensemble models can attain outstanding performance and reliability. We propose three deep ensemble learning (DEL) approaches, each with stable and reliable performance, that are workable on the above-mentioned data types. These are deep-stacked generalization ensemble learning, gradient deep learning boosting, and deep aggregation learning. The experiment results show that our proposed approaches achieve more vigorous and reliable performance than traditional ML and DL techniques on statistical, image-based, and sequential benchmark datasets. In particular, on the Heart Disease UCI dataset, representing the statistical type, the gradient deep learning boosting approach dominates the others with accuracy, recall, F1-score, Matthews correlation coefficient, and area under the curve values of 0.87, 0.81, 0.83, 0.73, and 0.91, respectively. On the X-ray dataset, representing the image-based type, the deep aggregation learning approach shows the highest performance with values of 0.91, 0.97, 0.93, 0.80, and 0.94, respectively. On the Depresjon dataset, representing the sequence type, the deep-stacked generalization ensemble learning approach outperforms the others with values of 0.91, 0.84, 0.86, 0.8, and 0.94, respectively. Overall, we conclude that applying DL models using our proposed approaches is a promising method for the healthcare support system to enhance prediction and diagnosis performance. Furthermore, our study reveals that these approaches are flexible and easy to apply to achieve optimal performance.

An intelligent fault detection and diagnosis monitoring system for reactor operational resilience: Fault diagnosis

Deep learning (also known as deep structured learning, hierarchical learning, or deep machine learning) is a branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data by using a deep graph with multiple processing layers, composed of multiple linear and nonlinear transformations. Deep learning has been characterized as a class of machine learning algorithms with the following characteristics [257]: • They use a cascade of many layers of nonlinear processing units for feature extraction and transformation. Each successive layer uses the output from the previous layer as input. The algorithms may be supervised or unsupervised and applications include pattern analysis (unsupervised) and classification (supervised). • They are based on the (unsupervised) learning of multiple levels of features or representations of the data. Higher-level features are derived from lower-level features to form a hierarchical representation. • They are part of the broader machine learning field of learning representations of data. • They learn multiple levels of representations that correspond to different levels of abstraction; the levels form a hierarchy of concepts. These definitions have in common: multiple layers of nonlinear processing units and the supervised or unsupervised learning of feature representations in each layer, with the layers forming a hierarchy from low-level to high-level features. Various deep learning architectures such as deep neural networks, convolutional deep neural networks, deep belief networks (DBN), and recurrent neural networks have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition, and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. In this chapter we start in Section 7.1 with an introduction, giving a brief history of this field, the relevant literature, and its applications. Then we study some basic concepts of deep learning such as convolutional neural networks, recurrent neural networks, backpropagation algorithm, restricted Boltzmann machines, and deep learning networks in Section 7.2. Then we illustrate three examples for Apache Spark implementation for mobile big data (MBD), user moving pattern extraction, and combination with nonparametric Bayesian learning, respectively, in Sections 7.3 through 7.5. Finally, we have summary in Section 7.6.