A noise detection method for rotating machinery monitoring data based on SES-LOF

Abstract

A noise detection method for rotating machinery monitoring data based on SES-LOF is provided. Firstly, a monitoring signal of the rotating machinery is obtained as the original signal, and the local abnormal factor value of each signal point is calculated according to the amplitude value. All signal points whose local anomaly factor value is greater than 1 are marked as suspicious noise points, and a set of suspicious noise points is obtained. Then, the suspicious noise point is selected from the set, and the suspicious noise point is removed from the original signal to obtain a new signal. Then Hilbert transform and Fourier transform are applied to the new signal, and Shannon entropy of the new signal is calculated. Whether the suspicious noise point is a real noise point is judged according to the 3 signma criterion; after traversing all the suspicious noise points in the set, the real noise point is retrieved finally. The method of the invention improves the ability of detecting noise points of rotating machinery monitoring data based on the traditional LOF algorithm, and has an ideal effect on the detection of noise points of rotating machinery monitoring data.

Description

technical field [0001] The invention belongs to the technical field of improving the quality of monitoring data of rotating machinery, and in particular relates to a noise point detection method of monitoring data of rotating machinery based on SES-LOF. Background technique [0002] Rotating machinery is widely used in fields such as electric power, petrochemical, metallurgy and transportation, and is an important equipment in industrial production. In order to ensure the safe and reliable operation of rotating mechanical equipment, it is necessary to monitor its health status and obtain a large amount of equipment operation monitoring data in real time. By analyzing and processing the monitoring data, equipment failures can be detected in time or even in advance. However, the working environment of rotating machinery equipment is often very harsh, resulting in a large number of drifting, distorted, and incomplete dirty data mixed in the monitoring big data. These dirty data...

AI Technical Summary

Problems solved by technology

LOF (Local Outlier Factor) is a commonly used algorithm for outlier detection. When the traditional LOF method is directly applied to the noise detection of rotating machinery monitoring data, the impact component in the rotating machinery monitoring signal is often mistakenly identified as noise. Noise elimination will destroy the typical fault information in the original data, resulting in the quality of monitoring data not improving but degrading

Images