Dynamic load identification for a power battery pack based on a combined regularization algorithm

Dynamic load identification is often required in engineering to understand and ensure structural integrity and reliability. However, ill-posedness is one of the main challenges to accurate dynamic loads identification, especially for practical large-scale problems, when continuous and impulsive loads are simultaneously contained. In this research, to address such an ill-posed problem occurred to identification of complicated loads, an approach is proposed by combining the Tikhonov regularization and the truncated generalized singular value decomposition (TGSVD). The combined regularization algorithm is exploited to identify the dynamic loads of a power battery pack installed on a commercial electrical vehicle in the time domain. Structural complex modal characteristics are considered in the method, which are the most cases in engineering. A series of moving overlapped windows is adopted to improve computational efficiency for long-time load identification. The identified loads show good consistency with those directly measured results, which demonstrates the applicability of the proposed approach to dynamic load identification for large engineering structures.