基于逆系统方法的多输入多输出非高斯驱动信号生成

多输入多输出(MIMO)振动环境试验是振动试验的新方法,该方法比传统的单输入单输出(SISO)方法更真实地模拟了结构在实际工作下的振动环境。传统随机振动试验的频域方法所模拟的是平稳高斯信号,但是实际振动环境往往是超高斯的,且超高斯激励和高斯激励对结构的损害差异很大,因此对MIMO超高斯振动环境试验的研究十分重要。驱动信号的产生是振动环境试验的关键。提出一种基于逆系统的MIMO超高斯驱动信号生成方法。首先,根据给定的响应参考功率谱密度(PSD)和参考峭度值采用相位调节法生成参考响应超高斯伪随机信号;其次,利用逆系统方法生成超高斯伪随机驱动信号;最后,对伪随机驱动信号进行时域随机化处理得到超高斯真随机驱动信号。对一根两输入两输出的悬臂梁模型进行了仿真验证,结果表明,采用本文方法所得到的驱动信号加载在梁上后,其输出响应谱与给定参考谱之间的误差完全满足工程中常用的±3 dB的要求,而且峭度也与给定的参考值十分接近,证实了所提方法的有效性。

Generation of multi-input multi-output non-Gaussian driving signal based on inverse system method

The multi-input multi-output (MIMO) vibration environment test is a new method in the vibration testing area, which can replicate the vibration environment endured by a structure under the practical situation more accurately than the traditional single-input single-output (SISO) manner. The traditional random vibration test with frequency domain method aims to generate a stationary and Gaussian vibration environment. But the practical vibration environments are always non-Gaussian, which can cause different damages to the structures compared with Gaussian excitations. Thus, it is significant to study the MIMO non-Gaussian test method. The generation of the driving signals is the key in the vibration environment test. In this paper, a method for generating MIMO non-Gaussian driving signals based on inverse system and phase manipulation is proposed. First, phase manipulation is used to generate the pseudo random non-Gaussian responses according to the given reference power spectrum densities (PSD) and kurtoses. Then inverse system method is used to generate the pseudo driving signals. Finally, the real random non-Gaussian driving signals are obtained by time domain randomization. A two-input two-output cantilever beam model is selected as an example. The simulation results indicate that the error between the output response spectra and the reference ones meets the ±3 dB requirements in engineering practice and the values of kurtoses of the outputs are very close to the reference ones. Thus, the proposed method is valid.