High-precision bearing signal recovery based on signal fusion and variable stepsize forward-backward pursuit

In multi-sensor, long-distance fault monitoring of rolling bearings, the bearing signals are compressively sampled, transmitted, and reconstructed according to the theory of com-pressive sensing. However, the reconstruction accuracy and speed are limited and are affected by the noise afflicting the collected signals. In this paper, a high-precision signal recovery method, based on signal fusion and the variable stepsize forward-backward pur-suit (VSFBP) algorithm, is proposed. First, the method adaptively adjusts the best estimate of the traditional random weighted fusion algorithm, by using the relative fluctuation value, which can fuse variable signals and reduce the noise component of the detection sig-nal. Second, two fuzzy parameters are used to control the step sizes of the atom selection and deletion in the two-stage matching pursuit algorithm; this improves the reconstruc-tion accuracy and speed of the algorithm under a high compression ratio. Finally, to pre -vent excessive backtracking, in the two-stage matching pursuit algorithm, the observation matrix is updated after each iteration, which improves the reconstruction accuracy of the algorithm further. Simulation and experimental results are compared to verify the effectiveness of the proposed method. (c) 2021 Elsevier Ltd. All rights reserved. Rotating machinery is widely used in aviation, metallurgy, wind power, and other fields [1,2]. Given that rotating machinery usually operates at high speeds, it inevitably causes massive pressure on the key component bearing, which makes the bearing one of the most easily damaged components in the rotating machinery [3?5]. It is worth mentioning that the failure of bearing brings enormous losses to society and the economy [6,7]. Therefore, it is particularly necessary to undertake realtime fault monitoring of bearings [8,9]. Considering that in most cases, the detection system and detected objects are usually located in different spaces, there

Automated summary

This paper proposes a high-precision signal recovery method based on signal fusion and the variable stepsize forward-backward pursuit algorithm, which improves the reconstruction accuracy and speed in multi-sensor, long-distance fault monitoring of rolling bearings.