Study On The Optimal Sensors Placement And Damage Identification Problems In The Health Monitoring System Of Large Span Bridges

During the operating period of a bridge,it is inevitably to suffer damages with the adding of the dynamic and static loads,the changing of the material when time flies and the influences caused by the environment,and such damages will finally result in function degeneration of the structure.Bridge health monitoring system is an important tool in catching the bridge information,evaluating the bridge health status during its lifetime.A comprehensive system would reflect the problems that bridges have,guarantee the safety of people and avoid substantial economic loss in time.The sensors subsystem and diagnose and forewarning subsystem are the two crucial parts in it.It is the hotspots that how to save the cost of the sensors subsystem under the condition of the bridge information be completely collected,how to optimize the quantity and locations that sensors be used and to identify the bridge damages through the information collected by the system.In order to solve the Optimal Sensors Placement(OSP)problem,this thesis proposes a Nested-Stacking Genetic Algorithm(NSGA),which combined the Genetic Algorithm(GA)with the Nested Partition Algorithm(NPA),to deal with some mathematical difficulties such as problems with huge solution domain,large number of discrete variables or dense extremum values exist,and a full mathematical structure is constructed in this thesis.Subsequently,two benchmark test functions are used to evaluate the performance of the new algorithm by comparing with GA,and the conclusion that the optimal ability of this new algorithm is better than GA can be obtained.Subsequently,according to a real engineer project named Wuhai Yellow River extrodosed cable-stayed bridge,ANSYS was used to build the finite element entity model and its corresponding dynamic modal analysis were carried out.According to the structural and dynamic characteristics of the bridge,the main monitoring objects of this bridge were determined.Lastly,adopting different sensor optimal criteria based on different monitoring objects,the data of the bridge is calculated by NSGA to obtain a comprehensive sensor placement scheme.In order to popularize the new algorithm to different problems,the parameters included in the new algorithm are analyzed for their sensitivity in this thesis based on the data of simplified model.The instructional value ranges of each parameter in NSGA are provided as theoretical foundation for computations in similar issues.Under the determined acceleration sensors arrangement scheme,three damage identification indices,curvature modal method,flexibility curvature method and uniform load surface method,which were researched recently are tested for damage identification.To partition the bridge through its shape and stress features,several working conditions with single damage and two points damage are formulated theoretically through simulation methods to discuss the indices sensitivity to damages,respectively.Through substantial theoretical tests,it is obviously that the curvature modal method cannot identify the damage effectively under the sensors arrangement scheme.The combining of the flexibility curvature method with the uniform load surface method can orientate most damages roughly,but the precise locations and extents of damages cannot be identified.Therefore,these aspects need further research in the future.This thesis is a part of the Mongolian transportation technology project named research on the health monitoring and security evaluating system of Wuhai Yellow River bridge.And the research achievements can be applied as the guidance in formulating the health monitoring system of this project.