Axle temperature fault detection method based on NEST and SPRT fusion algorithm

Abstract

The invention relates to the field of industrial equipment fault detection, and discloses an axle temperature fault detection method based on an NEST and SPRT fusion algorithm, and the method comprises the steps: obtaining a motor train axle temperature historical data feature set, carrying out the feature dimension reduction of the motor train axle temperature historical data feature set, and obtaining a motor train axle temperature feature parameter after dimension reduction; obtaining a bullet train axle temperature historical normal sample set corresponding to the bullet train axle temperature characteristic parameters after dimension reduction, establishing an NSET model, and obtaining a memory matrix D of the NSET model by utilizing the bullet train axle temperature historical sampleset; obtaining bullet train axle temperature test data, and outputting a bullet train axle temperature residual error sequence by the NSET model according to the memory matrix D; and carrying out variance test on the axle temperature residual error sequence of the motor train by using an SPRT test method to obtain an axle temperature state fault detection result. According to the invention, statedetection of all axle temperature can be realized through one-time modeling, fault early warning of the axle temperatures can be indirectly realized, and the invention has general universality for solving motor train unit equipment faults.

Description

technical field [0001] The invention relates to the field of industrial equipment fault detection, in particular to a shaft temperature fault detection method based on NEST and SPRT fusion algorithms. Background technique [0002] The status detection of the axle temperature plays a vital role in the safety of the actual operation of the high-speed EMU. Current axle temperature anomaly detection technology generally detects axle temperature changes by setting a fixed alarm threshold for a single axle temperature parameter, but this method does not consider the impact of axle temperature-related factors on itself, and the fixed threshold is based on human subjective Empirically determined, no exact theoretical basis. [0003] For example, the national patent publication "CN210363861U" discloses "a shaft temperature detection device and shaft temperature pre / alarm system". The utility model shaft temperature detection device includes: a first shaft temperature detection syste...

AI Technical Summary

Problems solved by technology

This utility model detects changes in shaft temperature by setting a fixed alarm threshold with a single shaft temperature parameter. This shaft temperature detection method relies on human experience, and the accuracy of shaft temperature detection is relatively low.

Images