Optimal Sensor Placement Of The Cable-Stayed Bridge Based On Intelligent Algorithms

Since the reform and opening up,with the rapid development of the China’s social economy,China’s transportation infrastructure has also flourished,especially since the 1990 s,China has built a large-scale bridge super project,and many long-span bridges with long spans and complex structures have emerged.With the passage of the time,the service time of these bridges has become longer and longer.Coupled with the complexity of the bridge stress and the harsh working environment,some serious bridge collapse accidents have occurred in recent years.Most of the bridges are built on the main roads.Once a safety accident occurs in the future,it will cause serious losses to people’s lives and property.Through the analysis of bridge safety accidents,it can be found that most bridges did not collapse without warning,but ignored these early warning information at that time,so real-time monitoring of the bridge structural safety can effectively avoid people’s lives and property losses.At present,the health monitoring of the bridge structure has become a hot research direction.The key of the health monitoring is to monitor the bridge structure in real time,and arranging sensors on the bridge structure is a necessary means to obtain real-time data.Because of the complex structure of some bridges,it is impossible to arrange sensors at every node.Optimizing the arrangement of sensors is the premise and key to the wide application of the bridge health monitoring system.In this paper,the cable-stayed bridge is taken as an example.Based on the improved genetic annealing algorithm,the sensor layout is optimized,and as few sensors as possible are used to obtain more bridge structure information.The main research work of this paper is as follows:(1)The basic concepts,basic steps,key parameters,advantages and disadvantages of the genetic algorithm and simulated annealing algorithm are described in detail.Aiming at the problem that genetic algorithm has strong global optimization ability but slow convergence speed and is easy to fall into local optimal solution,and compared with simulated annealing algorithm,which has strong local search ability but poor global search ability,an improved genetic annealing algorithm is proposed by combining genetic algorithm with annealing algorithm.Then,the optimization performance of the improved genetic annealing algorithm is tested by using single-peak and multi-peak test functions to verify its reliability and superiority.(2)Based on the laboratory model of the cable-stayed bridge,the corresponding finite element model of the cable-stayed bridge is established by ANSYS APDL finite element software,and then the structural modal analysis of the finite element model is carried out and its modal frequency and vibration mode are extracted.Through observation,it is known that the first 20 vibration modes are mainly vertical and vertical deformation,so the vertical acceleration sensor of the main beam and the longitudinal inclination sensor of the bridge tower are selected as the optimal layout objects.The deflection response of the finite element model and the laboratory model under moving load are analyzed respectively,and the deflection response curves obtained are basically consistent,and the data error is small,which verifies the accuracy of the finite element model and the mechanical characteristics of the cable-stayed bridge under moving load.(3)According to the results of the modal analysis,the vertical acceleration sensor of the main girder and the longitudinal inclination sensor of the bridge tower are selected as optimization objects,and the candidate measuring points of the main girder and the bridge tower are preliminarily determined,and the target modes of the candidate measuring points are extracted;Based on Fisher information matrix,the target modal order of the candidate measuring points of main girder and bridge tower is selected,and then the selected target modal is further optimized and intercepted by referring to structural strain energy on the basis of Fisher information matrix.Aiming at the difficulty in selecting initial measuring points of the modal confidence matrix(MAC),QR decomposition method is introduced to conduct primary selection of measuring points.Finally,the curve of the relationship between the maximum off-diagonal element of MAC matrix and the number of measuring points is drawn by the idea of the gradual accumulation,and the number of sensors is determined according to the curve change.(4)After determining the number of sensors,the program of the improved genetic annealing algorithm is written by MATLAB to optimize the layout of vertical acceleration sensors and longitudinal inclination sensors of the main girder.The results are compared with those of the traditional genetic algorithm and annealing algorithm.After analysis and comparison,the optimized results of the improved genetic annealing algorithm have higher fitness value,better convergence and robustness,and can also have better identification effect for areas with weak sensitivity,which shows that improved genetic annealing algorithm has better optimization ability.Finally,the optimization scheme of improved genetic annealing algorithm is regarded as the best scheme of the sensor optimal arrangement of the cable-stayed bridge.