Identification of Multi-Point Dynamic Load Positions Based on Filter Coefficient Method

Dynamic load identification plays an important role in practical engineering. In this paper, a novel fast algorithm is investigated to identify the multi-point dynamic load positions in frequency domain. For any given frequency, the amplitude of each load spectrum is relatively constant. By solving the kinetic equation set with the elimination method, the relationship of the true dynamic load positions can be expressed as the form of filter coefficients, then many dynamic load position combinations can be found that they do not satisfy the relationship so they can be excluded from the possible true position combinations. Compared to the traditional method, the novel algorithm only needs to sort out the true positions from a few dynamic load position combinations by the minimum determination coefficient method, which reduces the number of matrix inversion operations and improves the speed of the identification of load positions. Through a numerical simulation and an identification test on the simply supported beam structure, the high accuracy and effectiveness of the novel algorithm are successfully demonstrated, while the rapidity of the novel algorithm is shown by comparing the computation time of the novel algorithm with that of the traditional method.