基于自适应最小二乘支持向量机逆系统的链式STATCOM控制策略

对链式静止同步补偿器(static synchronous compensator,STATCOM)建立了非线性数学模型,采用逆系统反馈线性化的方法,构造了伪线性系统,并且将系统实现解耦。由于开关管IGBT的变化随时性及模型的不确定性,常常引起非线性模型参数的变化。为了解决这一问题,采用最小二乘支持向量机自适应地估计动态的模型非线性部分,通过反馈的方式加入到逆系统的输入端,从而补偿掉实际系统模型中的非线性部分,实现真正意义上的伪线性模型。由于是线性系统的解耦模型,采用线性二次型最优控制算法,实现配电网的无功补偿。仿真和实验结果表明,所提的算法很好地解决了配电网的无功补偿问题,且具有很好的鲁棒性和动态性能。

Control strategy of cascade STATCOM based on adaptive least square support vector machine and inverse system

A nonlinear mathematical model was established for the chain static synchronous compensator (STATCOM),and the whole control of the system was decoupled by using the method of inverse system with feedback linearization.Due to change of the switch tube (Insulated Gate Bipolar Transistor,IGBT)at any time and disturbance uncertainty,these will often lead to changes in the parameters of the nonlinear model.To solve this problem,least squares support vector machine (LSSVM) adaptively estimated the nonlinear dynamic part,and eliminated it by way of feedback.The whole system was a decoupling model of the linear system,so,linear quadratic regulator(LQR) was applied to distribution network for reactive power compensation of the load.Simulation results show that the proposed algorithm solves the problem of reactive power compensation in distribution network,and has good robustness and dynamic performance.