Serhat Seker

Feature extraction related to bearing damage in electric motors by wavelet analysis

The purpose of this paper is to extract features from vibration signals measured from motors subjected to accelerated bearing fluting aging and to detect the effects of bearing fluting at each aging cycle of induction motors. This trending of bearing degradation is achieved by monitoring the changes in the vibration signals at various sub-band levels. This is accomplished by the application of wavelet transforms and multi-resolution analysis in order to extract information from selected frequency bands with minimum signal distortion.

Spectral Analysis for Current and Temperature Measurements in Power Cables

Abstract This research aims to detect spectral properties under thermal and current variations for power cables. Therefore, spectral diversities are exposed under current unbalances and different load conditions through the spectral analysis techniques. Also, huge load variations are easily detected from the current signals in the time-frequency plane using the short-time frequency analysis. Hence, this study presents the determination of the frequency characteristics and spectral similarities between the phase currents and thermal variations.

On-line neuro-expert monitoring system for borssele nuclear power plant

Abstract A new method for an on-line monitoring system for the nuclear power plants has been developed utilizing the neural networks and the expert system. The integration of them is expected to enhance a substantial potential of the functionality as operators support. The recurrent neural network and the feed-forward neural network with adaptive learning are selected for the plant modeling and anomaly detection because of the high capability of modeling for dynamic behavior. The expert system is used as a decision agent, which works on the information space of both the neural networks and the human operators. The information of other sensory signals is also fed to the expert system, together with the outputs that the neural networks generate from the measured plant signals. The expert system can treat almost all known correlation between plant status patterns and operation modes as a priori set of rules. From the off-line test at Borssele Nuclear Power Plant (PWR 480 MWe) in the Netherlands, it was shown that the neuro-expert system successfully monitored the plant status. The expert system worked satisfactorily in diagnosing the system status by using the outputs of the neural networks and a priori knowledge base from the PWR simulator. The electric power coefficient is simultaneously monitored from the measured reactive and active electric power signals.

Transfer Function approach based upon wavelet transform for bearing damage detection in electric motors

This study presents a transfer function (TF) approach based on the continuous wavelet transform (CWT) using the vibration measurements for feature extraction. This approach helps to extract the origin of the bearing damage that develops during the aging process and then, it can be used to find the potential defects, which exist in healthy motor bearings as manufacturing defects. In this manner, there are two fundamental steps of the study. They are the definition of the transfer function and feature extraction which is related to the bearing damage characterization in electric motors. The definition of the feature transfer function is done between the first scale of the CWT analysis and original vibration signal in the frequency domain. Hence, it is introduced as a new viewpoint for condition monitoring studies in induction motors.

Multi-resolution wavelet analysis for chopped impulse voltage measurements and feature extraction

In this study, an application of the wavelet transform based on the multi-resolution analysis (MRA) is aimed for evaluation of amplitude and time parameters of the chopped impulse voltage. In terms of getting data set, three types of the chopped lightning impulse voltages with different chopping times are considered and MRA is applied to these data as different case studies. Hence, some characteristic properties are extracted from the impulse waveforms at some special frequency values. In this sense, frequency component of 33 MHz is found with the maximum peak value for each case before the chopping time and their appearing times for three cases are determined using the time-frequency analysis. Hence the difference between the chopping time and occurring time of the maximum peak values at 33 MHz is calculated easily. Also, some physical interpretations of the frequency component of 33 MHz are defined through the circuit parameters of the measurement system. In usual application, the MRA can be presented as a powerful technique to extract the noise effects, which come from different sources, in the high voltage measurements.

Redundancy-Based Predictive Fault Detection on Electric Motors by Stationary Wavelet Transform

In this study, a signal-based predictive fault-detection approach is developed to identify potential faults within an electric motor. In order to evaluate the performance of the proposed approach, first artificial motor vibration data are produced and used as a base line for analysis and assessment of the methodology. After successfully confirming proof of concept by detecting all fault frequencies hidden within the artificial data, the approach is then applied to the experimental data to see whether it can be accepted as a suitable potential fault detection tool for healthy electric motors. As the keystone of the study, algebraic summation operation (ASO) is introduced for predictive fault detection. ASO is built and based upon the redundant nature of stationary wavelet transform (SWT). Although similar to the SWT, the down sampling operation defined for the perfect reconstruction of SWT is omitted in ASO to amplify the redundancy within the transform. In other words, the redundancy obtained during decomposition is conserved on purpose with the goal of amplifying potential fault frequencies in electric motor vibration spectra and allowing for more robust and predictive fault detection.

Wavelet based trend analysis for monitoring and fault detection in induction motors

The aging mechanism of an induction motor is observed using vibration signatures in this study. A wavelet based trending application is used through Multi Resolution Wavelet Analysis. The progress of aging is shown and interpreted based on the trends of different cases of the same induction motor. Aging region concept is introduced to designate a new condition monitoring strategy. So, the vibration monitoring is refined into a smaller scale. By this way, a new simple and feasible monitoring algorithm is proposed. Then a rating parameter is introduced in order to estimate the situation of the motor. Thus, by using this method the aging of any random state can be detected as a percentage.

Generalized 2D lattice structure for causal and noncausal modeling of random fields

In this paper, the authors propose a new method to construct a low complexity and general purpose 2D lattice structure for modeling of random fields. The algorithm for obtaining this structure by employing the auxiliary vertical and horizontal prediction error fields is outlined. It is shown that the proposed 2D lattice structure with a rectangular prediction support region can be used to simultaneously obtain all possible causal and noncausal 2D models.

Using probabilistic neural networks with wavelet transform and principal components analysis for motor fault detection

This study represents an application of probabilistic neural networks along with multi resolution wavelet analysis, and principal components analysis to an induction motor which was applied to an accelerated aging process according to IEEE standard test procedures. In this manner, the algorithm first applies a multiresolution wavelet analysis to the vibration signals with Shannon entropy to calculate the feature vectors Then, principal components analysis is applied to the feature vectors, reducing the dimensionality for the condition monitoring classification that is to be made by the probabilistic neural networks. The application results show extremely high success rate, thus the study is vital in the scope of reliability.

Signal based fault detection for stator insulation in electric motors

Stator winding insulation damage in induction motors is considered in this research and fault features developed during the artificial aging process of the stator insulation are extracted from the collected data by the spectral analysis methods. Even harmonic effects are determined as a common feature between the phase currents and vibration signals using the coherence analysis and hence this feature is interpreted as an indicator of the stator insulation fault.