基于遗传算法的超导磁储能装置H2/H∞鲁棒控制器设计

提出了应用GA（遗传算法）设计H2／H∞鲁棒控制器的一般方法，使得在满足闭环系统的H∞范数小于指定的工程要求的条件下误差信号的H2范数最小，该问题可归结为一个带约束条件的极小值优化问题，并通过一种改进的遗传算法加以解决，该设计方法能对具有任意控制器结构的控制系统进行参数优化，与传统的鲁棒控制器设计方法相比，所提出的方法不仅可以提高控制器的鲁棒性，还具有易于进行参数优化的优点，将所提出的方法应用于超导磁储能装置（SMES）PI型鲁棒控制器的设计，时域仿真结果表明所设计的控制器具有很好的鲁棒性，在不同的运用工况下均能有效阻尼电力系统的振荡，此外，控制器还具有结构简单，易于工程实现的优点。

A GENETIC APPROACH TO MIXED ROBUST CONTROL DESIGN FOR SMES

A general approach is proposed to design a mixedrobust controller via genetic algorithms. The main objective of the proposed mixed control is to find out a suitable controller that minimizes the performance index of error signal subject to an unequal constraint on the norm of the closed loop system. This is firstly interpreted as a problem of optimal tracking subject to a robust stability constraint. A modified genetic searching algorithm is used to solve this problem. Compared with the conventional design methods for robust controller, the proposed approach is simple and easy to be employed for the industrial applications. The proposed algorithms are used to design a PI controller for SMES. The effectiveness of the controller is verified by a simple power system with the proposed controller. Simulation results show that the proposed PI controller can damp out the oscillations of the power system rapidly after a large disturbance under different system operating conditions.