【摘要】 频域子空间辨识方法具有数值稳定、辨识精度高等特点,但是其受噪声的影响会造成较大误差.本研究提出一种加权的频域子空间辨识算法,根据输出噪声方差对数据进行加权分配,以提高噪声条件下频域子空间辨识的精度,并采用数值模拟对改进算法进行了仿真验证.最后将其应用到某传动部件的模态参数辨识,获得了该部件的模态参数.更多还原

【Abstract】 The frequency state-space algorithm for mode identification is a algorithm with high accuracy and stability of numerical operation.But it need to pay attention to the effect from noise.A new frequency state-space algorithm with weight is put forward in this paper,in which the weight right is allocated according to the noise square variance.The new method is applied in model parameters identification of a transmission component.更多还原