大型结构分散控制系统的优化研究

STUDY ON OPTIMIZATION OF LARGE-SCALE STRUCTURAL DECENTRALIZED CONTROL

  • 摘要: 针对大型工程结构的分散控制方法进行研究,引入完全分散、部分独立分散及部分重叠分散控制的概念,分别以子系统控制效果局部最优及整体最优为目标设计分散控制结构的子系统控制器,并利用基于差分进化的混合粒子群算法对子系统进行优化,综合考虑子系统划分数量、子系统所含层数以及重叠部分所含层数的影响。对9层benchmark模型进行了多工况分散控制设计与数值分析。结果表明,整体最优目标下,采用部分重叠分散控制策略,能获得与集中控制策略接近的控制效果,相对于集中控制,重叠分散控制系统具有更高的可靠性,且不需要更多的控制能量。

     

    Abstract: Decentralized control method for large-scale structures is studied. The concept of fully decentralized control, partially independent and overlapping decentralized control is introduced. Subsystem controller is designed by different optimization targets (local and global optimization), and the PSO-DE-based hybrid algorithm is employed to obtain the optimal weight matrices in this study. Considering the number of subsystem, the DOFS of subsystem and the floors of the overlapping sections, the ASCE 9-story benchmark building is selected as an example to evaluate the control performances among different conditions. Numerical simulation results indicate that the overlapping decentralized control can reduce structural vibration as effectively well as the traditional centralized control, but is more reliable and does not require additional control energy compared with the traditional one.

     

/

返回文章
返回