Data Mining Practical Machine Learning Tools and Techniques

With the continuous increase in complexity and expense of industrial systems, there is less tolerance for performance degradation, productivity decrease, and safety hazards, which greatly necessitates to detect and identify any kinds of potential abnormalities and faults as early as possible and implement real-time fault-tolerant operation for minimizing performance degradation and avoiding dangerous situations. During the last four decades, fruitful results have been reported about fault diagnosis and fault-tolerant control methods and their applications in a variety of engineering systems. The three-part survey paper aims to give a comprehensive review of real-time fault diagnosis and fault-tolerant control, with particular attention on the results reported in the last decade. In this paper, fault diagnosis approaches and their applications are comprehensively reviewed from model- and signal-based perspectives, respectively.

/pdf/a-survey-of-fault-diagnosis-and-fault-tolerant-techniques-9uuf1uqysf.pdf

A Survey of Fault Diagnosis and Fault-Tolerant Techniques—Part I: Fault Diagnosis With Model-Based and Signal-Based Approaches

A review on empirical mode decomposition in fault diagnosis of rotating machinery

This paper investigates diagnostic techniques for electrical machines with special reference to induction machines and to papers published in the last ten years. A comprehensive list of references is reported and examined, and research activities classified into four main topics: 1) electrical faults; 2) mechanical faults; 3) signal processing for analysis and monitoring; and 4) artificial intelligence and decision-making techniques.

Advances in Diagnostic Techniques for Induction Machines

Early detection of the motor faults is essential and artificial neural networks are widely used for this purpose. The typical systems usually encapsulate two distinct blocks: feature extraction and classification. Such fixed and hand-crafted features may be a suboptimal choice and require a significant computational cost that will prevent their usage for real-time applications. In this paper, we propose a fast and accurate motor condition monitoring and early fault-detection system using 1-D convolutional neural networks that has an inherent adaptive design to fuse the feature extraction and classification phases of the motor fault detection into a single learning body. The proposed approach is directly applicable to the raw data (signal), and, thus, eliminates the need for a separate feature extraction algorithm resulting in more efficient systems in terms of both speed and hardware. Experimental results obtained using real motor data demonstrate the effectiveness of the proposed method for real-time motor condition monitoring.

Real-Time Motor Fault Detection by 1-D Convolutional Neural Networks

Recently, research concerning electrical machines and drives condition monitoring and fault diagnosis has experienced extraordinarily dynamic activity. The increasing importance of these energy conversion devices and their widespread use in uncountable applications have motivated significant research efforts. This paper presents an analysis of the state of the art in this field. The analyzed contributions were published in most relevant journals and magazines or presented in either specific conferences in the area or more broadly scoped events.

/pdf/advances-in-electrical-machine-power-electronic-and-drive-4jt41t4f8x.pdf

Advances in Electrical Machine, Power Electronic, and Drive Condition Monitoring and Fault Detection: State of the Art

In this paper, the authors propose a method for the diagnosis of rotor bar failures in induction machines, based on the analysis of the stator current during the startup using the discrete wavelet transform (DWT). Unlike other approaches, the study of the high-order wavelet signals resulting from the decomposition is the core of the proposed method. After an introduction of the physical and mathematical bases of the method, a description of the proposed approach is given; for this purpose, a numerical model of induction machine is used in such a way that the effects of a bar breakage can clearly be shown, avoiding the influence of other phenomena not related with the fault. Afterward, the new diagnosis method is validated using a set of commercial induction motors. Several experiments are developed under different machine conditions (healthy machine and machine with different levels of failure) and operating conditions (no load, full load, pulsating load, and fluctuating voltage). In each case, the results are compared with those obtained using the classical approach, based on the analysis of the steady-state current using the Fourier transform. Finally, the results are discussed, and some considerations about the influence of the DWT parameters (type of mother wavelet, order of the mother wavelet, sampling rate, or number of levels of the decomposition) over the diagnosis are done

/pdf/validation-of-a-new-method-for-the-diagnosis-of-rotor-bar-3gsi2u3wrt.pdf

Validation of a new method for the diagnosis of rotor bar failures via wavelet transform in industrial induction machines

In this paper, a general methodology based on the application of discrete wavelet transform (DWT) to the diagnosis of the cage motor condition using transient stator currents is exposed. The approach is based on the identification of characteristic patterns introduced by fault components in the wavelet signals obtained from the DWT of transient stator currents. These patterns enable a reliable detection of the corresponding fault as well as a clear interpretation of the physical phenomenon taking place in the machine. The proposed approach is applied to the detection of rotor asymmetries in two alternative ways, i.e., by using the startup current and by using the current during plugging stopping. Mixed eccentricities are also detected by means of the transient-based methodology. This paper shows how the evolution of other non-fault-related components such as the principal slot harmonic (PSH) can be extracted with the proposed technique. A compilation of experimental cases regarding the application of the methodology to the previous cases is presented. Guidelines for the easy application of the methodology by any user are also provided under a didactic perspective.

A General Approach for the Transient Detection of Slip-Dependent Fault Components Based on the Discrete Wavelet Transform

Transient-based fault diagnosis in induction motors has gained increasing attention over the recent years. This is due to its ability to avoid eventual wrong diagnostics of the conventional motor current signature analysis in certain industrial situations (presence of load toque oscillations, light loading conditions, and so on). However, the application of these transient methodologies requires the use of advanced signal processing tools. This paper presents a detailed comparison between the two main groups of transforms that are employed in transient analysis: discrete and continuous. This paper does not focus on trivial fault cases but on difficult real situations where the application of the conventional methods often leads to false diagnostics (outer bar breakages in double-cage motors, motors with rotor axial duct influence, and combined faults). Indeed, it is the first time that continuous tools are applied to some of these controversial situations. The results in this paper prove the special advantages of the continuous transforms, tearing down some false myths about their use.

Advanced Induction Motor Rotor Fault Diagnosis Via Continuous and Discrete Time–Frequency Tools

In this paper, a new induction motor diagnosis methodology is proposed. The approach is based on obtaining a 2-D time-frequency plot representing the time-frequency evolution of the main components in an electrical machine transient current. The identification of characteristic patterns in the time-frequency plane caused by many of the fault-related components enables a reliable machine diagnosis. Unlike other continuous-wavelet-transform-based methods, this work uses frequency B-spline (FBS) wavelets. It is shown that these wavelets enable an efficient filtering in the region neighboring the main frequency, as well as enable a high level of detail in the time-frequency maps. As a consequence, the evolution of the most important current components is precisely traced. These characteristics make it easy to identify the patterns related to the fault components. The technique is applied to the experimental no-load start-up current of motors in a healthy state and with broken bars; the FBS capabilities are revealed. One of the novelties of this paper is the fact that the diagnosis is carried out via the identification not only of the traditional lower sideband harmonic but also of the upper sideband harmonic and four additional fault-related components.

Jose A. Antonino-Daviu

Papers

Advances in Electrical Machine, Power Electronic, and Drive Condition Monitoring and Fault Detection: State of the Art

Validation of a new method for the diagnosis of rotor bar failures via wavelet transform in industrial induction machines

A General Approach for the Transient Detection of Slip-Dependent Fault Components Based on the Discrete Wavelet Transform

Advanced Induction Motor Rotor Fault Diagnosis Via Continuous and Discrete Time–Frequency Tools

Induction Motor Diagnosis Based on a Transient Current Analytic Wavelet Transform via Frequency B-Splines