多输入多输出混合振动试验中正弦与随机信号分离方法

研究了多输入多输出正弦加随机混合振动试验的控制方法，指出混合信号中正弦信号和随机信号的精确分离是提高控制精度的关键因素。提出了具有滤波特性的不相关积分法在时域中识别正弦信号，避免了频域识别的泄露误差问题，详细地推导了将给定频率的正弦信号从混合信号中分离出来的公式；数值计算显示该方法的识别精度达到0.44%。以一悬臂梁作为研究对象建立两输入两输出振动试验系统模型，使用比例均方根控制算法和正弦幅值修正法分别对随机振动和正弦振动进行修正，将随机信号控制在参考谱的±3dB以内，将正弦信号的幅值控制在参考值的±10%以内，满足振动试验要求。

Separation method of sinusoidal and random components in MIMO mixed vibration test

The control method of Multiple Input Multiple Output sine-on-random (SoR) vibration test is systematically discussed in the paper. It is pointed out that the accurate separation of mixed signals, which consist of sine and random signals, is the key factor to improve the control precision. A new separation method called Uncorrelated Integration is proposed out of its filtering characteristics to identify the amplitudes or phases of sine signals in the time domain, and the method can avoid the problem of energy leakage in frequency domain. The formula that sine signals of certain frequency separated from the mixed signals is derived in detail, and numerical calculation shows that the identification accuracy is 0.44%. A two input two output vibration system model is established. The random signals and sine signals are controlled by proportional RMS updating algorithm and amplitude updating algorithm respectively. The power spectrum density (PSD) of random signal is limited to ?3 dB of the reference PSD, and the amplitudes of sine signals are limited to ?10% of the reference value. The results show that the method proposed can meet the engineering demand.