基于最优控制的新型非线性预测控制算法

在电站过程控制中定值调节占很大比例,控制系统的抗阶跃干扰性能非常重要。另外,基于机理模型的电站仿真机,能够模拟真实机组对运行人员进行培训。这种机理模型有很高的动、静态特性,完全可以用来作为预测模型,实现对未来时刻输出的预测。作者根据最优控制的设计思想,设计出了将模型预测控制和常规控制相结合的控制方案。此算法改善了对象特性,具有较强的抗干扰能力和鲁棒性,适用于大惯性、有自平衡能力的大多数热工控制对象,为非线性系统模型预测控制提供了一条新思路、新方法。

A New Algorithm of Non-linear Predictive Control Based on Optimum Control

Tuning of fixed value is a very important part of the process control of power plant,so the capability of anti-interference of step signals in control system is significant. In addition,power plant simulators based on complex mechanism model can simulate real power units and train operators. Because of its good dynamic and static characteristics,mechanism model can be used as predictive models to dope out the output in the future. According to the principles of optimized control,a control scheme which combines the model predictive control with the routine control is designed in this paper. This scheme adapts to most thermodynamic control objects which have big inertia and strong capability of equilibrium for its strong anti-interference and robustness,and put forward a new method of non-linear predictive control.