Mechanical anomaly detection method based on potential feature coding

Abstract

The invention discloses a mechanical anomaly detection method based on potential feature coding. The mechanical anomaly detection method based on potential feature coding disclosed by the invention comprises the following steps: data preprocessing: preprocessing vibration signal data, comprising Fourier transform and normalization; and forward propagation: inputting the preprocessed signal into afirst full convolutional network to code the data. The mechanical anomaly detection method disclosed by the invention has the beneficial effects that in view of the problem of anomaly sample missing,the data distribution of normal signal samples is learned by using the feature mining capability of a deep network, and the signals are transferred into a potential space to perform data distributioncomparison by performing coding-decoding-re-coding on the signals.

Description

technical field [0001] The invention relates to the field of mechanical diagnosis, in particular to a mechanical abnormality detection method based on latent feature coding. Background technique [0002] Rotating mechanical equipment is developing towards large-scale, precision and automation, which puts forward stricter requirements for the manufacture, installation and daily maintenance of each component in the entire equipment system. A slight damage or vibration of any component Misalignment may affect the normal operation of the entire system and even cause major accidents. In order to ensure the healthy operation of machinery and equipment, the health monitoring system needs to collect massive data to reflect the health status, prompting the field of machinery health monitoring to enter the "big data" era. Mechanical big data has the characteristics of large capacity, variety and high speed. Mining information from the big data of machinery and equipment to efficient...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide a mechanical anomaly detection method based on latent feature coding. Aiming at the problem that the fault identification ability of the diagnostic method of signal time-domain statistical feature indicators is weak, the present invention is based on a deep learning network, through the whole volume The product network and the generative confrontation network are used to identify and learn the normal signal characteristics of the rotating machinery, and then compare the data distribution of the normal signal and the fault signal, so as to realize anomaly detection

Images