桥梁索力数据异常监测系统OA

In order to meet the urgent need for bridge safety monitoring and intelligent operation and maintenance(O&M)in China,this paper proposes a bridge anomaly monitoring system based on multi-sensor spatiotemporal collaborative learning,which differs from traditional low-efficiency,high-cost manual inspection methods.To ensure the system's continuous stability and effective generalization,a lightweight clustering anomaly data detection algorithm based on DBSCAN is proposed to effi-ciently and real-time detect sensor anomaly data.Meanwhile,a multi-sensor data repair network based on LSTM is developed,collaboratively mining the associated information from multi-sensors under varying spatiotemporal conditions to repair missing or inaccurate data caused by sensor failures.During the bridge anomaly detection process,a bridge structural anomaly determina-tion method based on the Naive Bayes classifier is adopted to generate accurate structural safety assessment results.Considering the integration of software and hardware in anomaly detection,this paper leverages the complementary advantages of real-time data collection and processing on the embedded end,as well as the rich computational resources of the host computer,decou-pling and deploying the anomaly detection methods to both the embedded system and host computer.The entire process is visual-ized through a GUI interface.The experimental results show that the system can effectively identify and repair anomaly data,im-prove the accuracy of bridge structural safety assessment,and perform well in anomaly detection accuracy and data repair in prac-tical applications,which provides a reliable safeguard for intelligent O&M and structural safety early warning of bridges.