一种基于频率响应函数的刚体惯性参数改进识别方法

Improved identification method for structural inertial parameters based on FRF

质量、质心位置、转动惯量和惯性积等惯性参数是结构动态仿真、性能优化、有限元分析、悬置安装位置和角度选取等过程中所需的重要参数。对三种常用的基于试验模态的结构惯性参数识别方法进行仿真比较,包括直接物理参数识别法,剩余惯量法(或称质量线法)和模态模型法。仿真发现,三种方法在识别精度、实验条件和实现难度上各有优缺点。通过对一平板结构进行仿真分析发现,三种方法单独使用时,会存在识别结果受到结构支撑系统刚度大小、噪声、测点坐标误差、模态参数识别结果等因素影响,或需要提前预知某些参数才能求解等问题。针对测试工作量和三种方法各自的特点,提出一种将直接参数识别法和剩余惯量法组合使用的改进识别方法。通过某型列车转向架结构试验进行了工程验证,证明了改进方法在含有较大噪声的实际工程环境下的适用性和可靠性,在保证精度的同时,还能减少了测试点数量,大幅降低了测试工作量,加快了实验进程。

Rigid body inertial parameters including mass, mass center, inertia moment and inertia product are essential ones for structural dynamic simulation, performance optimization, finite element analysis, mounting installation position and angle choosing. Here, simulations were performed contrastively for 3 commonly used methods for structural inertial parametric identification based on tested modes including the direct physical parametric identification(DPPI) method, the residual inertia(RI) one or the mass line(ML) one and the modal model(MM) one. The results showed that the 3 methods have advantages and disadvantages in identification accuracy, test conditions and realization difficulty, respectively. Through simulation analysis for a flat plate structure, it was shown that when each method is used singly, the recognition results can be affected by structural support stiffness, noise, coordinate error of measured point, and modal parametric identification results, or some parameters need to be known in advance. Aiming at problems mentioned above, an improved identification method combining DPPI and RI was proposed. Through tests of a certain type train bogie structure, the proposed method was verified in engineering. The result showed that the proposed method is applicable and reliable even in an environment of higher noise;it can reduce the number of measured points but keep the accuracy unchanged.