动态载荷识别的优化方法研究

发布时间：2018-09-05 19:33

【摘要】：目前,动态载荷识别领域应用较为广泛的一类方法是频域法,该方法起步较早,主要原理是对频域中的频响函数矩阵求逆,然后通过测得的响应和频响函数矩阵的逆矩阵求出未知载荷。在该方法中,系统输入输出呈线性关系,因此能够将动态载荷识别这一复杂问题简单化,从而方便工程应用。在识别过程中,频域法的困难在于如何获得条件数较小的频响函数矩阵,使得系统误差较小,从而准确的识别动态载荷。而频响函数矩阵的特性与参与识别的响应点特性有着一定的关系。遍历所有响应点组合的方式虽然可以从中选出条件数最小的频响函数矩阵,但却因计算量庞大(计算量随响应点数量增加呈指数级增长),而不适用于响应点数量较多的情况。本文针对工程上实测响应点较多的情况,通过分析单自由度系统及多自由度系统的频响函数矩阵,以响应点组合的频响函数矩阵的条件数作为目标函数,建立了动态载荷识别响应点选择的优化模型。该优化模型的数学模型属于整数规划中的0-1规划,其优化问题则属于离散领域的优化问题。本文用了一种基于改进离散粒子群算法的求解方法对该优化问题进行求解,该算法能够快速地解决庞大规模的组合优化问题。在对本文中优化模型研究的基础上,对几种常见的惯性因子设置策略做了说明和对比。最后,通过matlab仿真实验,分析了仿真的结果。实验结果验证了该模型能够在响应点数目较多的情况下,快速有效的选择出较优响应点组合,从而能够说明该模型及其求解方法在实际工程应用中具有较高的实用价值。
[Abstract]:At present, the frequency domain method is widely used in the field of dynamic load identification. The main principle of this method is to inverse the frequency response function matrix in the frequency domain. Then the unknown load is obtained by the measured response and the inverse matrix of the frequency response function matrix. In this method, the input and output of the system is linear, so the complex problem of dynamic load identification can be simplified, which is convenient for engineering application. In the recognition process, the difficulty of frequency-domain method is how to obtain the frequency response function matrix with small condition number, so that the system error is smaller, and the dynamic load can be accurately identified. The characteristics of the frequency response function matrix are related to the response point characteristics of the recognition. Although the method of traversing all the response points can select the frequency response function matrix with the smallest number of conditions, it is not suitable for the large number of response points because of the large amount of calculation (the amount of calculation increases exponentially with the number of response points). In this paper, the condition number of the frequency response function matrix of the response point combination is taken as the objective function by analyzing the frequency response function matrix of the single degree of freedom system and the multi-degree of freedom system. An optimization model for dynamic load identification response point selection is established. The mathematical model of this optimization model belongs to 0-1 programming in integer programming, and its optimization problem belongs to the optimization problem in discrete domain. In this paper, an improved discrete particle swarm optimization algorithm is used to solve the optimization problem. The algorithm can solve the combinatorial optimization problem on a large scale quickly. Based on the research of optimization model in this paper, several common inertial factor setting strategies are explained and compared. Finally, the simulation results are analyzed by matlab simulation experiment. The experimental results show that the model can quickly and effectively select the optimal response point combination when the number of response points is large, which shows that the model and its solution method have higher practical value in practical engineering application.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB123

【参考文献】