Selection of Fitness Function of Genetic Algorithm for Optimal Sensor Placement for Estimation of Vibration Pattern of Structures

It is often necessary to predict the vibration patterns of the structures from the signals of finite number of vibration sensors. This study presents the optimal placement of vibration sensors by applying the genetic algorithm and the modal expansion method. The modal expansion method is used to estimate the vibration response of the whole structure. The genetic algorithm is used to estimate the optimal placement of vibration sensors. Optimal sensor placement can be obtained so that the fitness function is minimized in the genetic algorithm. This paper discusses the comparison of the performances of two types of fitness functions, modal assurance criteria(MAC) and condition number( CN). As a result, the estimation using MAC shows better performance than using CN.