Research On Health Monitoring And Evaluation Method Of Long-Span Bridge Based On Machine Learning

The research of large-span bridges health monitoring and evaluation is of great significance in ensuring the safety and reliability of large-span bridges,extending the life of large-span bridges,and conducting scientific research and exploration of large-span bridges.This paper completes the design of the long-span bridge health monitoring platform by combining large-span bridges sensing test technology,broadband network communication technology,signal and information processing technology,.NET C# and Spring-Spring MVC-Mybatis framework,and completes the long-span bridge based on machine learning.Health monitoring and evaluation,taking the Jinmeng Yellow River Bridge as an example,completed the implementation of a large-span bridge health monitoring and evaluation method.The basic functions of bridge health monitoring system are designed by the influence of temperature,displacement,strain,vibration characteristics and other factors on the monitoring of large-span bridges,and the overall architecture of the system is completed with optical fiber sensing testing technology.After the basic function design is completed,the development of the full platform Computer / Server and Browser / Sever architecture is carried out.Through the collection and transmission of data,processing and control,the integration of the real-time system and the information system is completed.Therefore,in the process of designing a large-span bridge health monitoring system,the specific items and contents from the monitoring system should be determined by the structural characteristics of the bridge,the environment,and the investment scale.Based on Neural Network method,Support Vector Machine,Dynamic Bayesian Network,the design of early warning and evaluation analysis module is completed.Through collecting and analyzing of the data,the health status identification,the law and research of evolution and life of the Jinmeng Yellow River Bridge are deeply revealed.Support Vector Machines theoretically have better performance than Neural Networks,so the application in the field of large-span bridges damage identification is gradually paid attention,but Support Vector Machines cannot be extended to complex structures.Good statistical model.The algorithm is more accurate than the neural network algorithm and the support vector machine method,with an average damaged accuracy of 0.987 and an unimpaired accuracy of 0.984.Finally,Jinmeng Yellow River Bridge has adopted many new structures,new technologies,new materials,and new processes in structural design and construction.Therefore,it is undoubtedly of great significance to carry out the health monitoring and safety assessment of the Jinmeng Yellow River Bridge.Developed a series of experiments and explorations for the development of the Jinmeng Yellow River Bridge Bridge Health Monitoring System,completed the data collection and transmission,processing and control,and bridge health monitoring and evaluation software for the Jinmeng Yellow River Bridge Bridge Health Monitoring System,and concluded the long-span bridge health Monitoring system design and implementation methods and prospects.