Servo system multi-scale health assessment and fault prediction method

Abstract

The invention discloses a multi-scale health assessment and fault prediction method for a servo system, and solves the problem that the reliability of equipment performance monitoring of a servo control system in the prior art still needs to be improved. The method comprises the following steps: performing multi-scale analysis on a preprocessed vibration signal, and converting original vibration sequence data into a plurality of subsequences; carrying out anomaly detection on the sequence data, segmenting the subsequences under different scales into sequence segments, and further obtaining a to-be-detected sequence; detecting an abnormal state in each sub-sequence through a Markov anomaly detection model; the number of abnormal states in a certain time window before a certain moment is calculated, a health index at the moment is calculated, and a health assessment curve of the servo control system is obtained under different scales; monotonicity of the health assessment curves under different scales is calculated, and the scale where the curve with the best monotonicity is located is selected as the best health assessment scale. According to the technology, a health curve of the system is obtained according to an abnormal detection result, and faults are accurately predicted.

Description

technical field [0001] The invention relates to the field of fault data mining of a servo control system, in particular to a method for multi-scale health assessment and fault prediction of a servo system. Background technique [0002] Equipment with servo control systems are widely used in meteorology, national defense, aerospace and other fields. While the performance of these equipment is continuously improved, the problems of reliability, maintainability, failure prediction and maintenance support are becoming more and more serious. As an important part of these servo systems, the stability of their performance will also profoundly affect the reliability of the equipment. Traditional maintenance methods for servo control systems often have the problems of "insufficient maintenance" and "overhaul". The vibration signal in the working process has a time sequence, and the vibration signal can be considered as a set of time series. Anomalies in time series are often reflec...

AI Technical Summary

Problems solved by technology

[0006] However, machine learning-based anomaly detection methods often rely on high-quality labeled data, and such methods cannot accurately locate anomalies from abnormal data.

Images