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

1D convolutional neural networks and applications: A survey

During the last decade, Convolutional Neural Networks (CNNs) have become the de facto standard for various Computer Vision and Machine Learning operations. CNNs are feed-forward Artificial Neural Networks (ANNs) with alternating convolutional and subsampling layers. Deep 2D CNNs with many hidden layers and millions of parameters have the ability to learn complex objects and patterns providing that they can be trained on a massive size visual database with ground-truth labels. With a proper training, this unique ability makes them the primary tool for various engineering applications for 2D signals such as images and video frames. Yet, this may not be a viable option in numerous applications over 1D signals especially when the training data is scarce or application-specific. To address this issue, 1D CNNs have recently been proposed and immediately achieved the state-of-the-art performance levels in several applications such as personalized biomedical data classification and early diagnosis, structural health monitoring, anomaly detection and identification in power electronics and motor-fault detection. Another major advantage is that a real-time and low-cost hardware implementation is feasible due to the simple and compact configuration of 1D CNNs that perform only 1D convolutions (scalar multiplications and additions). This paper presents a comprehensive review of the general architecture and principals of 1D CNNs along with their major engineering applications, especially focused on the recent progress in this field. Their state-of-the-art performance is highlighted concluding with their unique properties. The benchmark datasets and the principal 1D CNN software used in those applications are also publically shared in a dedicated website.

1D Convolutional Neural Networks and Applications: A Survey

Medium-voltage (MV) induction motors are widely used in the industry and are essential to industrial processes. The breakdown of these MV motors not only leads to high repair expenses but also causes extraordinary financial losses due to unexpected downtime. To provide reliable condition monitoring and protection for MV motors, this paper presents a comprehensive survey of the existing condition monitoring and protection methods in the following five areas: thermal protection and temperature estimation, stator insulation monitoring and fault detection, bearing fault detection, broken rotor bar/end-ring detection, and air gap eccentricity detection. For each category, the related features of MV motors are discussed; the effectiveness of the existing methods are discussed in terms of their robustness, accuracy, and implementation complexity. Recommendations for the future research in these areas are also presented.

A Survey of Condition Monitoring and Protection Methods for Medium-Voltage Induction Motors

Condition monitoring of induction motors is a fast emerging technology in the field of electrical equipment maintenance and has attracted more and more attention worldwide as the number of unexpected failure of a critical system can be avoided. Keeping this in mind a bearing fault detection scheme of three-phase induction motor has been attempted. In the present study, Support Vector Machine (SVM) is used along with continuous wavelet transform (CWT), an advanced signal-processing tool, to analyze the frame vibrations during start-up. CWT has not been widely applied in the field of condition monitoring although much better results can been obtained compared to the widely used DWT based techniques. The encouraging results obtained from the present analysis is hoped to set up a base for condition monitoring technique of induction motor which will be simple, fast and overcome the limitations of traditional data-based models/techniques.

Bearing fault detection of induction motor using wavelet and Support Vector Machines (SVMs)

Bearing faults are one of the major causes of motor failures. The bearing defects induce vibration, resulting in the modulation of the stator current. In this paper, the stator current is analyzed via wavelet packet decomposition to detect bearing defects. The proposed method enables the analysis of frequency bands that can accommodate the rotational speed dependence of the bearing defect frequencies. The wavelet packet decomposition also provides a better treatment of nonstationary stator current than currently used Fourier techniques.

Bearing damage detection via wavelet packet decomposition of the stator current

This paper provides the theoretical basis for the measurement of reactive and distortion powers from the wavelet transforms. The measurement of reactive power relies on the use of broad-band phase-shift networks to create concurrent in-phase currents and quadrature voltages. The wavelet real power computation resulting from these 90/spl deg/ phase-shift networks yields the reactive power associated with each wavelet frequency level or subband. The distortion power at each wavelet subband is then derived from the real, reactive and apparent powers of the subband, where the apparent power is the product of the v, i element pair's subband rms voltage and current. The advantage of viewing the real and reactive powers. In the wavelet domain is that the domain preserves both the frequency and time relationship of these powers. In addition, the reactive power associated with each wavelet subband is a signed quantity and thus has a direction associated with it. This permits tracking the reactive power flow in each subband through the power system.

Reactive power measurement using the wavelet transform

Three-phase induction motors are the workhorses of industry because of their widespread use. They are used extensively for heating, cooling, refrigeration, pumping, conveyors, and similar applications. They offer users simple, rugged construction, easy maintenance, and cost-effective pricing. These factors have promoted standardization and development of a manufacturing infrastructure that has led to a vast installed base of motors; more than 90% of all motors used in industry worldwide are ac induction motors. Causes of motor failures are bearing faults, insulation faults, and rotor faults. Early detection of bearing faults allows replacement of the bearings, rather than replacement of the motor. The same type of bearing defects that plague such larger machines as 100 hp are mirrored in lower hp machines which has the same type of bearings. Even though the replacement of defective bearings is the cheapest fix among the three causes of failure, it is the most difficult one to detect. Motors that are in continuous use cannot be stopped for analysis. We have developed a circuit monitor for these motors. Incipient bearing failures are detectable by the presence of characteristic machine vibration frequencies associated with the various modes of bearing failure. We will show that circuit monitors that we developed can detect these frequencies using wavelet packet decomposition and a radial basis neural network. This device monitors an induction motor's current and defines a bearing failure.

Detecting motor bearing faults

Adjustable-speed drives perform many vital control functions in the industry, serving in such diverse applications as rolling mills, variable-speed compressors, fans, and pumps. When an adjustable-speed drive fails due to a bearing failure, it is usually catastrophic. Bearing defects introduce vibration anomalies that alter the current characteristic frequencies. This paper addresses the application of motor current signature analysis using wavelet packet decomposition to detect bearing faults in adjustable-speed drives.

Adjustable-Speed Drive Bearing-Fault Detection Via Wavelet Packet Decomposition

The fast Fourier transform (FFT) is the most widely used power system harmonic analysis tool in real-time power metering due to its computational efficiency. Recently, an alternate method, i.e., wavelet packet decomposition (WPD), has been applied to power system signals to meter the voltage and current harmonics. Although the new method provides better analysis, the computational complexity of WPD places a limitation on its use in real-time metering. This paper proposes the use of all-pass-implemented special half-band elliptic infinite-impulse-response filters in the WPD of power system signals. The proposed implementation reduces the computational complexity to levels comparable to FFT.

M.J. Devaney

Papers

Bearing damage detection via wavelet packet decomposition of the stator current

Reactive power measurement using the wavelet transform

Detecting motor bearing faults

Adjustable-Speed Drive Bearing-Fault Detection Via Wavelet Packet Decomposition

Harmonic Analysis Via Wavelet Packet Decomposition Using Special Elliptic Half-Band Filters