A method for early fault detection of compressor bearings

Abstract

The invention discloses a method for early fault detection of compressor bearings, which includes obtaining the autocorrelation length of the vibration data sequence through an autocorrelation function, compressing the vibration data sequence by an autoencoder, suspending and processing the data sequence set, and then calculating the threshold value to convert the health Compare the index with the threshold value to judge the early fault detection, and finally find out whether the compressor bearing is faulty; the present invention uses the autocorrelation function and the bearing data compression method of the autoencoder to solve the problem of different vibration data lengths; and then through the data sequence The integrated processing and technology can prevent the detection results from being affected by noise and cause large fluctuations, which will affect the early fault detection; at the same time, the early fault detection method based on the adaptive threshold and historical data and the decision logic of early fault detection are used to reduce False positives and false negatives.

Description

technical field [0001] The invention relates to the technical field of mechanical equipment fault detection, in particular to an early fault detection method for compressor bearings. Background technique [0002] The bearing is one of the most easily damaged components in the compressor, and its safe operation has a crucial impact on the compressor. In a large-scale wind test, if the bearing in the compressor is damaged, the entire compressor will fail and cause economic losses. If the fault can be detected in the early stage of the bearing fault, and the damage of the bearing can be found as soon as possible, the harm caused by the fault will be greatly reduced, which is of great significance to the large-scale wind test. Contents of the invention [0003] Therefore, in order to solve the above-mentioned deficiencies, the present invention provides a kind of compressor bearing early fault detection method here, through the bearing data compression method of autocorrelati...

AI Technical Summary

Problems solved by technology

[0003] Therefore, in order to solve the above-mentioned deficiencies, the present invention provides a kind of compressor bearing early fault detection method here, through the bearing data compression method of autocorrelation function and autoencoder, to solve the problem that the length of vibration data is different; The processing and technology of the sequence set can prevent the detection results from being affected by the noise and cause large fluctuations, which will affect the early fault detection; at the same time, the early fault detection method and the decision logic of early fault detection based on the adaptive threshold and historical data are adopted. Reduce false positives and false negatives

Images