Simultaneous identification of structural local damage and impact force using sparse regularization method

Abstract

In the field of structural health monitoring (SHM), simultaneous identification of local damage and impact force is a research topic with important theoretical significance and engineering value. However, due to the limited measurement data, nonlinear structural system, and noise interference, this simultaneous identification faces considerable challenges. The damage is characterized by the stiffness reduction coefficient of the finite element model based on sparse regularization in this article. Under the presupposition of structural damage, the whole structure is divided into several independent units, and the units are grouped and independent optimization tasks are set up. In each task, a sparse regularization method is used to reconstruct the impact force and act as an optimization constraint. In the process of optimization, the gradient descent method with a greedy strategy is used to gradually determine the optimal damage identification scheme. Sparse regularization based on K sparsity criterion is used to solve the optimization problem, so that the optimization process can obtain accurate and relatively sparse solutions, and has good noise resistance. Numerical simulation and experimental research show that this method can effectively identify both impact force and local damage of structures, and has good recognition accuracy and stability.