Fault diagnosis of Fuel Pump Based on Wavelet Denoising and Deep Learning. Issue 1 (1st March 2022)

• Record Type: Journal Article
• Authors: Guo, Yudi; Chen, Xin; Wang, Weizhen
• Is Part Of: Journal of physics. Volume 2216:Issue 1(2022)
• Publication Date: 2022-03-01
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/2216/1/012050
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: Aiming at the problem that the signal collected by the sensor of aircraft fuel system contains much noise, which masks the effective fault characteristics, a fault diagnosis method of aircraft fuel pump based on wavelet threshold denoising and convolutional neural network is proposed. Wavelet threshold denoising is added to the traditional data preprocessing to eliminate the noise contained in the data; the convolution neural network is used to build a fault detection model to identify and locate the fault of the fuel pump. The experimental results show that this method can accurately and effectively identify and locate the fuel pump fault based on the noisy signal, and the recognition accuracy is more than 85%, which proves the effectiveness of this method for aircraft fuel pump detection.