宁西占 黄伟

摘   要：为解决实时混合试验系统中时变时滞严重影响试验稳定性和精度的问题，以基于混合灵敏度的H∞鲁棒控制方法设计反馈控制器，使试验系统具有稳定的动态特性;以多项式外插进一步消除时滞，并将其应用于线性Benchmark问题.为充分探讨所提方法性能，采用线性和非线性加载系统模型，开展了物理子结构为线性和非线性的虚拟实时混合试验研究.结果显示：作动器的实测位移均与期望位移几乎完全重合;与已有方法的对比显示，采用鲁棒时滞补偿的实时混合试验具有更小的误差;而对非线性物理子结构，试验系统的均方根误差和峰值误差均在0.6%～1.5%之间，且均方根误差大于10%和峰值误差大于6%的概率远小于0.05.研究结果表明，鲁棒时滞补偿方法可大幅提高实时混合试验的模拟精度，并具有较强的鲁棒性.

关键词：实时混合试验;时滞;H∞控制;鲁棒性;不确定性

中图分类号：TU317                          文献标志码：A

Numerical Study of Robust Time-delay Compensation

Method for Real-time Hybrid Simulation

NING Xizhan1，2，HUANG Wei1

（1. College of Civil Engineering，Huaqiao University，Xiamen 361021，China;

2. Key Laboratory for Intelligent Infrastructure and Monitoring of Fujian Province（Huaqiao University），Xiamen 361021，China）

Abstract：To address the problem that the variable time delay seriously affects the stability and accuracy of the real-time hybrid simulation（RTHS），a robust time-delay compensation method was studied. A feedback controller  based on H∞ robust control method of mixed sensitivity was designed to stabilize the dynamics of the testing system，while the polynomial extrapolation method was employed to compensate for the time-delay. The proposed method was applied to a linear Benchmark problem. To further investigate the performance of the proposed method，virtual RTHSs were carried out with linear and nonlinear physical substructure（PS） employing linear and nonlinear models of the loading system. The results showed the measured displacements were almost identical to the desired displacements. Compared with the existing method，RTHS with the robust time-delay compensation method had a smaller error. For nonlinear PS，the root-mean-square error（RMSE） and peak error（PE） were both between 0.6% and 1.5%，and the probability of RMSE greater than 10% and PE greater than 6% was far less than 0.05. The study demonstrated that the robust delay compensation method can improve the accuracy of RTHS greatly and behavior strong robustness.

Key words：real-time hybrid simulation;time delay;H∞ control;robustness;uncertainty

混合试验源于1969年日本学者[1]提出的拟动力试验方法，它将数值计算和物理试验通过在线的方式有机结合在一起，是评估结构在动荷载作用下性能的经济、实用的方法. 通常，混合试验方法将待评估结构中具有较强非线性的部分取出在实验室进行物理加载，剩余部分在计算机中模拟. 前者称为物理子结构，后者称为数值子结构，子结构间通过传递系统（通常是作动器）联系在一起. 混合试验提出之初采用了慢速加载方式，并在逐步积……