The Research On Optimization Methods Of The Dynamic Force Identification

At present, the frequency domain method is one of the most widely used methods of dynamics force identification, which began relatively earlier. The main principle of the method is to use inverse matrix of frequency response function in frequency domain and the measured response to identify the unknown force. Due to the linear relationship of the system input-output dates in the method, the complicated problems on the dynamics force identification become simplified, which makes the method more convenient for engineering application. In the process of the dynamic force identification, the trouble of the frequency domain method lies in how to get frequency response function (FRF) matrix with smaller condition number to reduce systematic error and identify the dynamic force accurately.However, there is a relationship between the characteristic of FRF matrix and response points. FRF matrix with smallest condition number can be selected when trying all combination of response points, but the calculated amount will be too massive to meet the case of more measured response points.According to the situation of more measured response points in engineering, this paper establishes the optimal model of response point selection for the dynamic force identification by analyzing the FRF matrix of the single-degree-of-freedom system and the multiple-degree-of-freedom system. Meanwhile, the condition number of FRF matrix will be considered as the objective function. The optimization model of the response point selection belongs to discrete optimization problems, and its establishment relies on the 0-1 integer programming model. In order to solve the problem, this article introduced an improved discrete particle swarm algorithm, which can solve the large-scale combinatorial optimization problems. Based on the research of the optimization model, the paper made comparison and analysis on several common setting strategies of inertia weight.Finally, results of the matlab simulation experiments prove that the model is able to select the best combination of response points quickly and effectively, even under the situation of more response points. Besides, the results also indicate that the model and its solution method have a high practical value in engineering application.