Optimal Sensor Placement On Bridge Based On Genetic Algorithm And Monitoring Signal Analysis For Bridge

In company with scale-up of the bridge construction,the novel and unique architectural style often cause complicated mechanical behavior of the designed structure.The vibration monitoring system is a necessary component part of bridge health monitoring system.Theoptimized arrangement of acceleration sensor points is directly related to the collecting effect of monitoring system information,so it needs to satisfy the need of bridge health monitoring with the most limited measurement points.When the optimum layout scheme of acceleration sensor is determined and the bridge health monitoring system is running,the acceleration signals collected by the correlative points are analyzed.Combining with numerical simulation,experimental analysis and engineering case,this thesis studies several different points of measurement optimization layout algorithm and analyzes the acceleration signals collected by the vertical and transverse measurements in the real bridge,using 3 different signal analysis methods,the main research of this thesis are listed below.1.This thesis summarizes the status quo of the research on the health monitoring of the bridge structure and constructs the mathematic model of the optimum layout of the sensor.Finally summarizes the signal analysis methods commonly used in the present.Aiming at the optimum layout of sensors:effective independent algorithm,genetic algorithm,simulated annealing algorithm and ant colony algorithm principle are studied in detail.The method of signal analysis:Fast Fourier transform,wavelet transform and HHT transform are studied in detail.2.The convergence of Soft computing(GA,simulated annealing algorithm and ant colony algorithm)are analyzed in detail and the MATLAB program of Soft computation are compiled.With 4 typical functions as the numerical experiment platform,the comparison experiment analysis of these 3 algorithms are used respectively and the optimum performance and application range are analyzed.The practical application performance of soft computing is verified by the problem of optimal path optimization with complex measurement points.When there are only 10 test points,the 3 algorithms can converge to the global optimal solution quickly,but when there are 100 measuring points,approximate solution can be obtained.3.For the Shanghai-Tong Yangtze River Bridge,the main span 1092m using MIDAS CIVIL2013 to establish finite element model,then modal analysis,extracting relevant modal information,and then using effective independent method,improved genetic algorithm and simulated annealing algorithm to optimize the vertical acceleration measurement point and the longitudinal inclination point of Bridge tower in main truss.According to the optimization results,the optimization results and characteristics of different algorithms are compared and the comprehensive layout scheme are determined.4.Optimum arrangement of vertical acceleration sensor for main truss of Shanghai Tong Yangtze River Bridge,'in the MIDAS finite element model of Shanghai-Tong Yangtze River Bridge,the unit impulse load and white noise excitation are applied to the bridge structure respectively,the acceleration response signal is extracted from the sensor point position,the signal time domain analysis is carried out,and the modal parameters of the bridge are identified.5.In allusion to the non-stationary acceleration signals collected by the vertical and transverse acceleration sensors in a bridge,the HHT transform is used to analyze the measurement points of the vehicle loads and the white noise environment.By comparing the analysis results of HHT Transform and FFT and wavelet transform,it proves that the non-stationary signal collected in the health monitoring system can be obtained by using HHT transform to analyze the frequency-time-amplitude graph of Hilbert spectrum,which has high resolution.Simultaneously can effectively and quickly identify the different spectral bands in the original signal.