Fault Diagnosis Methods Based on Machine Learning and its Applications for Wind Turbines: A Review

Abstract

With the increase in the installed capacity of wind power systems, the fault diagnosis and condition monitoring of wind turbines (WT) has attracted increasing attention. In recent years, machine learning (ML) has played a crucial role as an emerging technology for fault diagnosis in wind power systems has played a crucial role. Even though ML methods have shown great potential in dealing with the issues related to the fault diagnosis of WT, there are still some challenges encountered in many aspects. In this paper, typical fault diagnosis methods based on ML methods for wind power systems are thoroughly reviewed in terms of both theoretical fundamentals and industrial applications, including traditional machine learning (TML), artificial neural networks (ANN), deep learning (DL) and transfer learning (TL), in the development line of ML technologies. The advantages and disadvantages of various methods are analyzed and discussed. Meanwhile, a distribution diagram is provided for the discussions of ML methods applied for WT fault diagnosis, and the existing challenges on the applications for fault diagnosis based on ML for wind power generation systems are presented. Moreover, some prospects for future research directions are provided. INDEX TERMSwind turbines, machine learning, fault diagnosis, review Nomenclature AE autoencoder AI artificial intelligence ANN artificial neural networks ART adaptive resonance theory BPNN back propagation neural network CA clustering algorithm CNN convolutional neural network DAE denoising autoencoder DBN deep belief network FCM fuzzy C-means clustering HMM hidden Markov model LSSVM least squares support vector machine LSTM long short-term memory network ML machine learning RBFNN radial basis function neural network RF random forest RLM extreme learning machine RNN recurrent neural network RVM relevance vector machine SAE stacked autoencoder SCADA supervisory control and data acquisition SDAE stacked denoising autoencoders 23 SOM self-organizing map 24 SVM support vector machine 25 t-SNE t-distributed logistic neighbor embedding 26 TL transfer learning 27 TML traditional machine learning 28 VMD variational mode decomposition 29 WT wind turbines 30 I. INTRODUCTION 31 With the increasing consumption of fossil fuels and the 32 gradual deterioration of environmental problems, there is an 33 urgent need to find a clean and renewable energy source. 34 Wind energy is irreplaceable in energy structures owing 35 to its rapid growth. Wind power accounts for 20% of the 36 world's total electricity, and WT are receiving increasing 37 attention as the core components of wind power generators [1], 38 [2]. Usually, wind power generators are installed in remote 39 areas or offshore areas where traffic is inconvenient, and 40 the gearbox is generally installed in a sky above tens or 41even hundreds of meters from the ground. In addition, the 42 blades are often subjected to complex alternating impact loads

A Reliable Deep Learning Approach for Time-Varying Faults Identification: Spacecraft Reaction Wheel Case Study

“…The proposed fault diagnosis approach can be adapted to be applicable to other spacecraft subsystems as well as different safety-critical systems as in the industry. • Computationally expensive [40] • Requires many trainable parameters [72] • Suffer from the issue of gradient disappearing in reverse fine-tuning [73] • Accuracy is less significant when used in fault diagnosis [74] • Not appropriate for safety-critical systems due to the time required for intensive calculations.…”

AI‐Enabled Metal‐Polymer Plain Bearing Based on the Triboelectric Principle

“…Through layer‐by‐layer convolution and pooling of filters, the topological features hidden in the data are extracted step by step, and finally, the characteristics of the input data are obtained during the process of translation, rotation, and scaling. [ 68 ] Here, the proposed AI analysis model is based on a CNN, including the feature extractor and the full connected layer (classifier). The feature extractor contains five convolutional layers, and the classifier contains three layers, as described in Figure 5b.…”

Fault Diagnosis of Rotating Machinery Using Denoising-Integrated Sparse Autoencoder Based Health State Classification

“…With the rise of various analysis tools, the rapid development of data-based mechanical fault diagnosis technology has been promoted. Currently, with benefit from data mining technology to achieve equipment health condition detection has played a decisive role in many key tasks [1][2][3], and has long been valued and studied [4][5][6][7]. However, since large scale and complexity have become the characteristics of current applied machines, the composition and structure of the device are related and affected, and the monitoring signals characterizing the running state exhibit high dimension, sparse, hard-toquantify and indistinguishable properties.…”