STUDY ON DIFFERENT BEAM MODELS IN MOVING FORCE IDENTIFICATION

Several methods have been developed recently to identify moving forces on a beam. The results from all of them exhibit fluctuations at the beginning and end of the time history and they are natural outcomes of the ill-conditioned inverse problem. Accuracy of the identified forces also depends on the parameters of the beam–force system which have not been systematically studied. This paper employs Tikhonov regularization technique to provide bounds to the ill-conditioned results in the identification problem. The calculation of the optimal regularization parameter is discussed. The improved algorithm is then used to study the effect of different parameters on the identification results through single-force and multi-forces identification. Both the Timoshenko beam theory and the Euler–Bernoulli beam theory are included in the study for a comparison. Laboratory results on two moving forces identification are also presented for illustration. Timoshenko beam model is found better than the Euler–Bernoulli beam model in the identification, and the limits of applications of the moving force identification method are also reported.