A high efficiency impact force identification method based on non-convex sparse regularization

Sparse regularization methods have been extensively used in impact force identification to accurately determine locations and the time history of impact forces. It is known that the accuracy of localization can be significantly improved by increasing the location of the potential impact action, which also leads to a significant increase in the dimensionality of the transfer matrix. Consequently, this reduces computational efficiency substantially and makes it difficult to achieve online monitoring. Considering a non-convex sparse regularization method, such as lp- norm (0 < p < 1), guarantees that the identified more accurate amplitude of the impact force. In this paper, a reduced iteratively reweighted l 1 - norm minimization algorithm (rIRL1) is proposed, which combines the lp- norm regularization with the reduced matrix to reconstruct the reduced transfer matrix for alleviating computational difficulty. Several simulations and experiments are conducted on an edge-fixed plate to compare the performance of iteratively reweighted l 1 - norm minimization algorithm (IRL1) and rIRL1 algorithm. Results indicate that the proposed rIRI1 algorithm accurately obtains the impact force and its location under different noise levels. Furthermore, the proposed method has faster solution efficiency.