A Method for Reconstructing Information of Faulty Sensors Based on Correlation Degree of Measured Values

Abstract

The present invention provides a faulty sensor information reconstruction method based on the correlation degree of measured values. The method includes the following steps: S1. For the faulty sensor, calculate the correlation degree between the response of the measuring point where the sensor is working normally and the responses of other measuring points; S2. By comparing the magnitude of the correlation degree, determine the response variable required for the establishment of the correlation model; S3, and then use the partial least square method to establish the reconstruction model of the reconstruction variable and the response variable, and use the measured data of structural health monitoring to respond to the fault sensor Information reconstruction. The method provided by the present invention has a good effect on the reconstruction of faulty sensor information, significantly reduces the reconstruction error, and at the same time ensures that the change trend of the reconstructed value is consistent with the change trend of the actual value, and the reliability of the overall structure analysis is remarkable .

Description

technical field [0001] The invention belongs to the fields of civil engineering structure health monitoring and information processing, and in particular relates to a fault sensor information reconstruction method based on the correlation degree of measured values. Background technique [0002] Data acquisition is a very important link in the structural health monitoring system, and the analysis of structural safety performance depends on the reliability of data acquisition. In the actual operation process of civil structure health monitoring, due to factors such as equipment aging, it is inevitable that some sensors will fail and cause data distortion. At the same time, some sensors are not easy to replace, resulting in data loss and reducing the reliability of the overall data. The overall analysis of the structure leads to decision-making mistakes and loses the meaning of structural monitoring; at the same time, structural monitoring will generate a large amount of data, ...

AI Technical Summary

Problems solved by technology

However, this method takes less time, and due to the neural network recovery method, it is mainly aimed at data reconstruction of short-term sensor failures, and has no obvious effect on sensors with high sampling rates and long-term failure prediction.

[0004] By analyzing the research of domestic and foreign scholars on data recovery and prediction, it is found that the current method is still insufficient to establish a structurally stable correlation model, and the effect on long-term fault reconstruction is not obvious. It is necessary to study the method of establishing a correlation model based on monitoring data Perform failure data reconstruction

Images