Time–frequency analysis

About: Time–frequency analysis is a research topic. Over the lifetime, 5407 publications have been published within this topic receiving 104346 citations.

A new perspective for the IEEE standard 1459-2000 via stationary wavelet transform in the presence of nonstationary power quality disturbance

Abstract: Power components, power factors and pollution factor are defined according to the IEEE Std. 1459-2000 based on the FFT. However, the FFT in the presence of nonstationary Power Quality (PQ) disturbances results in inaccurate values due to its sensitivity to the spectral leakage problem. In this paper, a new perspective for the IEEE Std. 1459-2000 definitions is introduced using the Stationary Wavelet Transform (SWT). As a time-frequency transform, SWT can provide variable frequency resolution while preserving time information with no spectral leakage as the FFT. Moreover, unlike other time-frequency transform such as Discrete Wavelet Transform (DWT), SWT possess the time-invariance property that keeps the time and frequency characteristics throughout all the decomposition levels. Results of different case studies including stationary, nonstationary of synthetic and real PQ disturbances proves the effectiveness of applying the SWT over FFT or DWT.

A Multi-Input and Multi-Task Convolutional Neural Network for Fault Diagnosis Based on Bearing Vibration Signal

Abstract: Bearing fault diagnosis is essential for the safe and stable operation of rotating machinery. Existing methods use signals from a single dimension, limiting diagnostic generality and accuracy. To address these limitations and make improved use of signal features from multiple dimensions, a novel convolutional neural network model with multi-dimensional signal inputs and multi-dimensional task outputs called MIMTNet is proposed. First, frequency domain signals and a time frequency graph are obtained by using the short-time Fourier transform and a wavelet transform to process original time domain signals simultaneously. Then, the time domain signals, the frequency domain signals, and the time frequency graph are fed into the model and a special aggregation is performed after extracting features from the three corresponding branches. Finally, the outputs of the three-dimensional tasks are acquired by different full connection layers to process the aggregated features of bearing position, damage location within the bearing, and the damage size. Two common bearing vibration signal datasets are used to verify the generalization ability of our proposed method. And experimental results show that the proposed method effectively improves the bearing diagnosis capability of the deep learning model.