A sensor fault diagnosis method and system

Abstract

The invention discloses a sensor fault diagnosis method and system. The method includes: determining the first, second and third electrical signals of a sample sensor, and the sample sensor includes a normal working sensor and a faulty sensor; determining the first electrical signal of the sample sensor , the second and third electrical feature vectors, and the fault source label vector; determine the cluster center of each feature vector; determine the first, second and third electrical signals and the first, second and third electrical signals of the sensor to be detected Eigenvectors; according to the first, second, and third electrical eigenvectors of the sensors to be detected and the cluster centers of each eigenvector, determine the fault type of the sensor to be detected; detect the faulty sensor, and use the detection results to perform incremental learning . The invention detects the sensor fault and confirms the source of the fault based on the difference of the electrical characteristics of the power-on instant and the power-off instant caused by the change of the parasitic capacitance of the fault sensor, and the difference of the signal time domain characteristics in the working state.

Description

technical field [0001] The invention relates to the field of sensor fault diagnosis, in particular to a sensor fault diagnosis method and system. Background technique [0002] In industrial production and testing sites, it is often necessary to use various types of sensors that output analog signals to monitor various parameters and indicators in the production process. For example, in automobile brake experiments, temperature sensors, pressure sensors, pedal force sensors, displacement sensors and other sensors are widely used. Due to the harsh environment of industrial production sites, sensors are prone to failures. Timely detection of sensor failures and determination of sensor failure sources can effectively reduce maintenance time, improve production efficiency, and ensure production quality. Traditional sensor fault diagnosis methods, such as patent document CN201810461085.1, use data denoising, data trend analysis, and sliding window diagnosis to complete sensor fau...

AI Technical Summary

Problems solved by technology

[0003] However, this method based on data and rules has many defects in actual use.

First, a failed sensor may have a similar output to a working sensor; second, a sensor failure may be caused by a variety of factors, such as low battery, microprocessor errors, sensor damage, and other factors

Determining the fault source of the sensor requires deep professional background knowledge, and the cost of technicians with professional background knowledge is high

Images