Research On Optimization Of Sensor Placement For Bridge Health Monitoring

Sensor placement optimization is vital to system of bridge health monitoring. Using fewer sensors to get more structure information is preferred. Much research on long span and complicated bridge focuses on optimization of sensor placement under dynamic action while optimization method under static action is rarely presented. On the other hand, research usually deals with place problem but neglects number problem.This paper chooses Genetic Algorithm (GA) as a tool, and Partheno Genetic Algorithm (PGA) and an improved PGA are presented to adapt to character of sensor placement problem. Then an objective function as basis to optimization algorithm is provided and it combines information of sensor and structure, which borrows an idea from wireless sensor network.After the objective function is determined, the method to optimize sensor number and placement under static action is presented, which is as following: getting structure information and detection probability in the first place; and then determining monitor radius; after that using PGA in small size generation to calculate value of fitness while number of sensor changes and get a tendency char to find out a proper number of sensor; in the last step, the improved PGA is used combining objective function to calculate the best value of fitness and determining the place of sensor. At last, the optimization method is applied in Luowanghe Bridge project to optimize sensor placement.