A Novel Anti-Noise Fault Diagnosis Approach for Rolling Bearings Based on Convolutional Neural Network Fusing Frequency Domain Feature Matching Algorithm

Abstract

The development of deep learning provides a new research method for fault diagnosis. However, in the industrial field, the labeled samples are insufficient and the noise interference is strong so that raw data obtained by the sensor are occupied with noise signal. It is difficult to recognize time-domain fault signals under the severe noise environment. In order to solve these problems, the convolutional neural network (CNN) fusing frequency domain feature matching algorithm (FDFM), called CNN-FDFM, is proposed in this paper. FDFM extracts key frequency features from signals in the frequency domain, which can maintain high accuracy in the case of strong noise and limited samples. CNN automatically extracts features from time-domain signals, and by using dropout to simulate noise input and increasing the size of the first-layer convolutional kernel, the anti-noise ability of the network is improved. Softmax with temperature parameter T and D-S evidence theory are used to fuse the two models. As FDFM and CNN can provide different diagnostic information in frequency domain, and time domain, respectively, the fused model CNN-FDFM achieves higher accuracy under severe noise environment. In the experiment, when a signal-to-noise ratio (SNR) drops to -10 dB, the diagnosis accuracy of CNN-FDFM still reaches 93.33%, higher than CNN’s accuracy of 45.43%. Besides, when SNR is greater than -6 dB, the accuracy of CNN-FDFM is higher than 99%.

Bearing Fault Diagnosis Using Lightweight and Robust One-Dimensional Convolution Neural Network in the Frequency Domain

“…In general, the time domain or the frequency domain can be used to analyse the signals. The signal properties are more distinct when time-domain signals are converted to frequency-domain signals since the latter signals are less influenced by noise compared to the former signals [ 43 ]. The presence of a fault characteristic frequency would amplify the signal component amplitudes that are associated with the characteristic component amplitudes that are associated with the characteristic fault frequency, which allows the detection of the failed bearing location through the initial vibration signal of the frequency components, which corresponds to the nature of the bearings [ 38 , 44 , 45 ].…”

“…While these approaches have achieved significant progress, the accuracy of these models is reduced when the noise level exceeds −4 dB. Zhou et al [ 43 ] proposed an integrated framework via the Convolutional Neural Network and Frequency-Domain Feature Matching (CNN-FDFM) algorithm to assist in sustaining the excessive noise levels. Essential frequency features from the frequency-domain signals are captured by FDFM and retained at high accuracy with limited samples under high noise conditions.…”

2D CNN-Based Multi-Output Diagnosis for Compound Bearing Faults under Variable Rotational Speeds

“…For instance, Shao et al [39] proposed a hybrid model that combined 1D CNN and SVM with the support of an improved swarm optimization algorithm to enhance the performance and convergence speed. Zhou et al [40] proposed an 1D CNN-based fusing frequency feature matching algorithm to extract key frequency features in the signal spectrum for bearing fault diagnosis under noisy environments. Ince et al proposed an efficient 1D CNN-based method that adapts an inherent adaptive design to combine a feature extractor and classifier into a single learning body, with the input data being raw signals [41].…”

A Soft Sensor for Measuring the Wear of an Induction Motor Bearing by the Park’s Vector Components of Current and Voltage

“…The use of a neural network to detect the status of the stator current sensor was suggested. The work convolutional neural network (CNN) fusing a frequency domain feature matching algorithm (FDFM) is used for the diagnosis of rolling bearings [ 25 ].…”