Search

Recently, precise and deterministic feature extraction is one of the current research topics for bearing fault diagnosis. For this aim, an experimental bearing test setup was created in this study. In this setup, vibration signals were obtained from the bearings on which artificial faults were generated in specific sizes. A new feature extraction method based on co-occurrence matrices for bearing vibration signals was proposed instead of the conventional feature extraction methods, as in the literature. The One (1) Dimensional–Local Binary Patterns (1D-LBP) method was first applied to bearing vibration signals, and a new signal whose values ranged between 0–255 was obtained. Then, co-occurrence matrices were obtained from these signals. The correlation, energy, homogeneity, and contrast features were extracted from these matrices. Different machine learning methods were employed with these features to carry out the classification process. Three different data sets were used to test the proposed approach. As a result of analysing the signals with the proposed model, the success rate is 87.50% for dataset1 (different speed), 96.5% for dataset2 (fault size (mm)) and 99.30% for dataset3 (fault type – inner ring, outer ring, ball) was found, respectively. [ABSTRACT FROM AUTHOR]

This article demonstrates a technique for diagnosis of fault type and faulty phase on overhead transmission lines. A method for computation of fault location is also incorporated in this work. The proposed method is based on the multi-resolution S-transform, which is used for generating complexS-matrices of the current signals measured at the sending and receiving ends of the line. The peak magnitude of the absolute value of everyS-matrix is noted. The phase angle corresponding to every peak component is obtained from the argument of the relevantS-matrix. These features are used as input vectors of a probabilistic neural network for fault detection and classification. Detection of faulty phase(s) is followed by estimation of fault location. The voltage signal of the affected phase is processed to generate theS-matrix. The frequency components of theS-matrices for different fault locations are used as input vectors for training a back-propagation neural network. The results are obtained with satisfactory accuracy and speed. All the simulations have been done in MATLAB (The MathWorks, Natick, Massachusetts, USA) environment for different values of fault locations, fault resistances, and fault inception angles. The effect of noise on both the current and voltage signals has been investigated. [ABSTRACT FROM PUBLISHER]

In this paper, we propose a highly reliable state monitoring system for induction motors. The proposed system utilizes vibration signals to analyze characteristics of the induction motor and extract features for classifying abnormal states from normal ones. To extract the features of faulty and healthy signals, we first convert one-dimension vibration signals into two-dimension gray images to utilize the relationship between each element and its neighboring elements, and we calculate the number of significant pixels in these converted images. We then use multiclass support vector machines to distinguish between abnormal data and normal data. The experimental results indicate that the proposed state monitoring system achieves 100% classification accuracy. In addition, we explore the effects of the noise components inherent in the vibration signals by adding white Gaussian noise to the vibration signals to obtain signal-to-noise ratios (SNRs) of 10 dB, 15 dB, 20 dB, 30 dB, and 40 dB, respectively. The experimental results show that the proposed approach continues to achieve 100% classification accuracy in noisy environments with SNRs of at least 15 dB. Furthermore, the experimental results show that the proposed approach outperforms a conventional state-of-the-art algorithm in both noisy and noiseless environments. [ABSTRACT FROM PUBLISHER]