A Method for Optimization of Sensor Locations for Inverse Calculation of Impact Load on Stiffened Panel

Abstract

Abstract There are many scenarios in the area of ships and marine structures where impact on structure members like stiffened panels where the impact loads are difficult to directly measured and one may need to turn to inverse calculation. The inverse calculation of impact load is very important as impact loads are characterized by short action time and large amplitude, which have adverse effects on structural safety. In practice, one of the difficulties in inverse calculation of impact load is the need of a large number of sensors, which means high cost and huge amount of computation. Therefore, it is necessary to optimize the layout of sensors for inverse calculation of impact load. To investigate the method for optimization of sensor locations for inverse calculation of impact load on stiffened panel, the inverse calculation results of different sensor layout of the stiffened panel are analyzed by applying Particle Swarm Optimization. Based on the optimization principle of Particle Swarm Optimization, a fitness function with the overall relative error and cosine similarity as the judgment criteria is established. The inverse calculation of impact loads on the stiffened panel is carried out with different number of calibration points and sensors. The regularization method is applied to deal with the ill-conditioning of matrix. The effects of the number of calibration points and sensors on the inverse calculation of impact loads are discussed. An optimal sensor layout that meets the requirements of the fitness function is found under a calibration point number and sensor number by Particle Swarm Optimization.