基于多模型的单元机组负荷多变量模型算法控制研究

Multi-model based multivariable MAC and its application to unit load system

首先将负荷系统分解为两个双输入单输出系统,并用模型算法控制设计两个子系统的控制器。通过对两个二元一次控制器方程的求解,达到多变量解耦设计的目的。针对机组负荷变动频繁,根据不同负荷下的固定模型设计了不同的模型算法控制器,并在线计算对象和各模型的匹配度以确定各控制器控制量的加权值,达到模型平滑切换的目的。通过对300 MW直流燃煤机组的仿真,证明了该方法的有效性。

The unit load system is decomposed into two two-input and single-output systems. Two sub-controllers are designed with model algorithmic control （MAC）. The multi-variable decoupling design is achieved by solving the equation groups. Different MAC controllers are designed based on fixed models under different loads to overcome the frequent frequency fluctuation of unit load, The matching degrees of objects with fixed models are calculated online to decide the weights of different controllers. The smooth switching among models is realized, Simulation result of 300 MW unit proves that the new method is effective.