Optimization Of Sensor Placement In Long Cable-stayed Bridges By PSO Algorithm

With the increasing of sophisticated bridge structures, the implementation ofstructural health monitoring and its safe operation has become the widespreadattention in the field of general engineering. A good sensor system directly affectsthe reliability of data, which related to the effect of structural health monitoring, butthe optimal allocation of sensors depend on the layout strategy and methods. As anew optimization algorithm, the advantages of particle swarm optimizationalgorithm are fast convergence and high accuracy. It adopts random transformmethod and searches for the direction through the objective function value, Whichprovides a feasible solution for complex multi-objective optimization problem. So itespecially suitable for dealing with the sensor placement problem of large spanspace structure which has multiple nodes. This paper puts forward a multi-objectiveoptimization layout method based on particle swarm optimization (PSO) algorithm,and achieves the optimal placement of sensors of the reduced scale model ofBinzhou Yellow River highway bridge.And compared with the traditional methodsand relevant test, the feasibility and validity of this method was verified.In this paper, the number of cable-stayed bridge sensor is optimized. Based onFisher information matrix2-norm method, this article selects the number of targetmode of girder and the main tower. It determines monitoring mode order number ofthe girder vertical and horizontal and the main tower longitudinal, which depends onModal Strain Energy. The QR decomposition is made on the vibration mode of themeasurable points obtained on the basis of the finite element modal analysis todetermine the initial positions of sensors. Then determining the number of sensorsof the main tower longitudinal and girder vertical, horizontal.Particle swarm optimization algorithm is proved to be correct throughnumerical simulation and comparison analysis. To determine the optimal sensorlocation by three kinds of traditional methods such as Effective Independence,EFI-Drive Point Residue and QR decomposition method. Three deferent fitnessfunctions were built using root mean square of MAC matrix off-diagonal elementsand sum of Gramian matrix elements. Using improved particle swarm optimizationalgorithm to optimize algorithm to realize the sensor optimization placement ofcable-stayed bridge. Analysis of the advantages of PSO algorithm by comparing theevaluation index of traditional optimization algorithm.An experimental test platform is built up with the help of scale model ofBinzhou Yellow River highway bridge in laboratory. Using the improved PSO algorithm, an optimal sensor placement is worked out. Then we can extract themodal parameters of cable-stayed bridge model. Finite element analysis results andthe experiment data show that this method is effective.