Structural Parameter Identification And Sensor Layout Optimization Based On Extended Kalman Filter

Civil engineering structures are affected by natural disasters such as earthquakes and long-term loads,and they will be damaged,and the damage will continue to accumulate,which can easily lead to sudden damage to the structure and cause hidden social safety hazards.Therefore,it is very necessary to develop structural health monitoring technology,and carry out parameter identification and real-time monitoring of the structure.The Kalman filter algorithm has the advantage of real-time identification of structural physical parameters in the case of incomplete observation response data,so it has become the most commonly used physical parameter identification method.Considering the limited number of sensors in the actual project,different layout schemes have a greater impact on the accuracy of the recognition results.Therefore,based on the extended Kalman filter algorithm,this paper studies the optimization of the location and number of sensors in the structure under partial observation response data.The paper mainly completed the following tasks:1.The basic principle of Kalman filter is described in detail,and its recurrence formula is derived,and the recognition process of the extended Kalman filter method applied to shear-type structure is sorted out,which provides a theoretical basis for subsequent research.The numerical simulation results of different structures are used to identify the physical parameters such as structural stiffness and damping.The results show that the extended Kalman filter algorithm can well identify the physical parameters of linear and nonlinear structures.2.In order to study the sensor placement problem of the structure with a given number of sensors,the 8-story shear type structure is numerically simulated,so that the number of accelerometer sensors is 1,2,3,and 4;and the number of sensors remains unchanged.The schemes of different observation positions are identified,and a reasonable layout position recommendation is given by analyzing the error curve of the stiffness and damping identification results.In order to avoid the contingency of the calculation results,a 13-story structure is used as a model to verify the obtained structure and conclusions.3.In order to study the number of sensors in different structures that can meet the requirements of parameter identification accuracy and balance economic conditions,5,9,12,and 16 stories of structures are selected for identification,and the number of acceleration sensors increases from 1 to the number of structural stories and the location of sensors are selected according to the principle that the bottom and top storiess are arranged separately and the rest are close to the top story of the structure.Through the comparative analysis of the recognition error,the optimal sensor arrangement quantity of each layer structure is obtained.Finally,suggestions for the optimal number of sensors for structures within 20 stories are given,and the rationality of the suggestions is verified by the recognition results of 8-story and 13-story structures.