Abstract: To solve the ill-posed problem, a two-step ideology is proposed to reduce the ill condition of transmissibility function and weaken the influence of noise in force identification research. First, the optimal combination of measurement points is calculated based on the conditional number theory of transfer function, and a transfer function matrix with the lowest degree of ill condition is obtained. Then, Tikhonov regularization is adopted to identify input excitation. In the procedure mentioned above, B-spline function is introduced to interpolate L-curve to acquire more exact regularization parameters. Simulation results show that the method proposed is able to effectively reduce force identification error and achieve a more accurate force time history.

Key words: force identification, optimum measuring point, B-spline function, L-curve, regularization