An Adaptive Two-Level Alarm Method for Axial Displacement Sudden Monitoring

Abstract

An adaptive two-level alarm method for axial displacement mutation monitoring. First collect displacement signals, establish a data set, and preprocess to obtain a training data set of displacement; then extract features from the training data set to form feature data composed of N data According to the time information, the training data set and the feature data set are fused; then the statistical parameters of the feature data set are estimated according to the probability density, and a first-level alarm model is established; then all the data in the training data set are sorted in ascending order to obtain the data set, Calculate the quantile range and median of the data set, determine the initial threshold value of the displacement training data set according to the principle of multiple quantile range, and establish a two-level alarm model; as the engine running time goes by, judge and Update; the present invention realizes the self-adaptive alarm of the axial signal of the engine, the calculation amount of the method is small, the required historical data is less, and online monitoring can be realized.

Description

technical field [0001] The invention belongs to the technical field of equipment signal processing and monitoring and diagnosis, and in particular relates to an adaptive two-level alarm method for axial displacement mutation monitoring. Background technique [0002] The rocket engine is a kind of jet engine used in the aerospace field. It uses the discharge of high-temperature and high-speed gas to generate thrust. It is characterized by high cost, complex design structure and harsh working environment. Before the rocket engine is delivered to use, it often faces extremely rigorous tests. In order to prevent rocket motor explosion and casualties during the test, the status monitoring and fault warning of rocket motors are particularly important. However, the current rocket engine has a series of problems such as complex design structure, difficult signal acquisition, poor working conditions of sensors, limited types of collected signals, and low signal-to-noise ratio, so th...

AI Technical Summary

Problems solved by technology

The traditional method is based on the probability density estimation to learn the probability distribution to form an adaptive alarm curve, but this method requires a large amount of previous data accumulation and cannot effectively monitor engine state changes under small sample conditions

Images