A Bayesian Regularization-based Method For Identification Of Impact Force On Composite Structure

Nowadays,the increasing usage of composite materials in flight vehicle structures has become a trend.However,composite materials are especially vulnerable to low-velocity impacts,which may introduce invisible internal damages,thus significantly reducing the strength and stability of the structure and even causing serious consequences.Therefore,it is very important to develop structural health monitoring system to identify the impact forces acting on composite structure and evaluate the damage status.This can help to make maintenance decision to increase the structural safety.This thesis performs a study on identification of impact force for composite structures.Firstly,a time domain model for impact force identification is established by discretization of the Duhamel integral equation,and regularization is introduced to solve the ill-conditioned problem after its ill-posedness is analyzed.Secondly,based on Bayesian statistical theory,a Bayesian regularization(augmented Tikhonov regularization)-based impact force identification method is proposed,which can adaptively determine an optimal regularization parameter from the measured data while reconstruct the impact force history.A finite element model of a composite wing structure is constructed by software MSC.Patran/Nastran to perform numerical study.The newly developed Bayesian regularization method is used to reconstruct the impact force on the wing structure model,and the results are compared with those obtained by Tikhonov regularization whose regularization parameters are given by L-curve method and generalized cross validation(GCV)method.Finally,a test setup is established to perform experimental study of impact force identification for a sandwich composite structure.Combining with linear interpolation,a method by minimizing an objective function is proposed to identify the impact location and then Bayesian regularization method is used to reconstruct the impact force history.The numerical results and experimental results have demonstrated the applicability and effectiveness of the proposed impact force identification method.