基于动态优化的多模型广义预测控制器设计

针对工业控制过程中广泛存在系统参数突变的问题,将多模型切换的广义预测控制器引至动态优化策略下的分层式控制系统中,设计了基于动态优化的多模型广义预测控制器.该模型预测控制结构以获取最大经济效益为目标,上层结构对经济目标函数进行动态优化,得到使经济利益最大的关键变量设定值;下层结构中MPC层采用多模型广义预测控制器代替传统单模型广义预测控制器追踪上层得到的设定值,即采用多个固定模型和自适应模型并行辨识系统的动态特性,提高系统暂态性能和模型参数跳变时系统的调节能力;底层为PID控制器用于抑制过程中的扰动.通过仿真验证了该方法的可行性和有效性.

Design of Multi-Model GPC Based on Dynamic Optimization

A generalized predictive control algorithm based on multiple model switching is proposed for the hierarchical model predictive control structure based on dynamic optimization to design multi-model GPC controller and on dynamic optimization for an industrial control process with jumping parameters.For the economical model predictive control structure,taking the maximum economic benefit as the goal,the upper structure dynamically optimizes the economic objective function to get the key variable setpoints：the lower level MPC controller uses a multi-model generalized predictive controller instead of a single model controller in MPC layer to track the setpoint given in parallel to by the upper-level,namely,multiple fixed models and adaptive model are used to identify dynamic systems in parallel to improve the transient performance of the system and the ability to regulate the jumping parameters;The underlying for PID controller is used to restrain the disturbance in the process.Finally,the feasibility and efficiency of the proposed method is illustrated by a case study.