滑动回归最小二乘算法在某电动振动台模型辨识中的应用

Application of Moving-Horizon Recursive Least squares Algorithmin Model Identification for a Electrodynamic Shaker

摘要:

为了能够准确地建立电动振动台的仿真模型,了解电动振动台系统特性,笔者在分析电动振动台数学模型的基础上,利用先进的滑动回归最小二乘辨识算法,有效地开展了电动振动台这一复杂电力学系统阶数和时延的确认,参数估计,并最终建立了系统的ARMA模型；文章最后,举出了某型电动台空台面和另一型振动台带滑台的模型辨识结果,并通过正弦扫频试验给予验证；结果表明,该算法辨识精度高,收敛速度快,鲁棒性强,能克服一定的噪声,对于高阶线性系统辨识具有良好的辨识效果。

Abstract:

In order to accurately establish the simulation model of electric vibration table, to understand the characteristics of electric shaking table system. Author firstly analyzes the mathematics model of the electrodynamic shaker, and secondly uses the advanced Moving-Horizon recursive least squares algorithm (MRLS) to determine the electrodynamic shaker system order and time delay and estimate the parameters of the system. Finally the ARAM model of the electrodynamic shaker is built. In this paper end, the identification result of a electrodynamic shaker with empty table and another electrodynamic shaker with slip table is given and is verified by making sine sweep vibration tests. The result shows that MRLS algorithm has a good identification capability and strong robustness for the high order line system.