Optimal Sensor Placement Based On Pigeon Colony Algorithm(PCA)

Optimal placement of sensors is an important part in structural health monitoring. It is important to arrange the sensors in given numbers in the actual project. Swarm intelligence optimization algorithm has strong global optimization ability, which can find the global optimal solution for the optimization criterion. But most of the presence of swarm intelligence optimization algorithms have premature convergence, large cycle indexes and slow convergence and easily fall into local optimums for high-dimensional, multi-peak and complicated functions. In this paper, for solving global numerical optimization problems, a new swarm intelligence optimization algorithm is proposed:PCA. and the algorithm is successfully applied to the study of the project. The contents are as follows:(1) Propose a new swarm intelligence optimization algorithms:PCA. Introduce the PCA concept in detail, and give three process of PCA algorithm:take-off, flight and homing. The concrete steps and formulas of algorithm are given. The detailed flowcharts and pseudo code are given.(2) Analyze the parameters of PCA. Divided the parameters of PCA in groups, given the parameters selected range. Demonstrate the iterative process of PCA and confirmed the applicability of the parameters through a numerical example.(3) Use low-dimensional, high-dimensional and nonlinear equations to analyze the optimal result of PCA. Compared PCA with the PSO and standard genetic algorithm. The results show that PCA has features:1) Since the algorithm only need to compare the value of the objective function, less demanding on the nature of the objective function, the function expression can also be expressed in the form of a non-functional form; 2) strong global convergence, fewer cycles algorithm, fast convergence for low-dimention functions; 3) PCA has good global convergence, small cycle indexes, and strong stability to find optimal solutions in high-dimensional, multi-peak and complicated problems:(4) The PCA algorithm is introduced to optimal arrangement of the sensor. Three-dimensional modal assurance criteria is selected as the optimization objective function. Using a reference model of University of Florida to test the main parameters of the algorithm. The considering the redundancy of information and without redundancy considered are compared respectively. The results show that, PCA algorithm can be applied in Optimal placement of sensors, in The considering the redundancy of information of the three-dimensional modal criteria can better fit the modal shape, such that the sensor arrangement is more reasonable.(5) Use Runyang Yangtze River Bridge as a practical engineering example. The main beam of the South Bridge of Runyang Yangtze River Bridge is modeled. The main beam has 93 nodes to arrange the sensors. The minimum of maximum value of non-diagonal element and average of non-diagonal element of non-dimensional modal assurance criterion are selected as optimization objective function. The optimized result of PC A are compared. By MSE value and fitting curve, confirming considering the redundancy of information of optimized arrangement is better, and the sensor arrangement is more reasonable. The results show that it is feasibility to use PCA in sensor optimal placement.