ABSTRACT
Moving force identification (MFI) has attracted increasing attentions in recent years. However, the MFI results obtained from the existing methods are inaccurate under measurement noises. One of reasons is fact that the characteristics of moving forces are not properly explored. To fix the problem, a novel MFI method is proposed by combining an alternating iterative method (AIM) and regularizations here. Firstly, each of moving forces is decomposed into constant force and time-varying force components respectively. The prior knowledge that time-varying force component usually fluctuates around constant force component is then adopted. Rather than convex optimization problem, a semi-convex constrained estimation problem is defined and solved by the AIM. Numerical simulations on MFI in a simply supported beam are conducted to assess its effectiveness. The results show that the proposed method can provide a higher MFI accuracy than the truncated generalized singular value decomposition method and the prior knowledge is reasonable.
