工业大系统双层结构预测控制的集中优化与分散控制策略

为降低工业大系统模型预测控制(Model predictive control,MPC)在线计算复杂度,同时保证系统的全局优化性能,提出一种集中优化、分散控制的双层结构预测控制策略.在稳态目标计算层(Steady-state target calculation, SSTC),基于全局过程模型对系统进行集中优化,将优化结果作为设定值传递给动态控制层;在动态控制层,将大系统划分为若干个子系统,每个子系统分别由基于各自子过程模型的模型预测控制进行控制,为减少各子系统之间的相互干扰,在各个子系统之间添加前馈控制器对扰动进行补偿,提高系统的总体动态控制性能.该策略的优点在于能确保系统全局最优性的同时降低了在线计算量,提高了工业大系统双层结构预测控制方法的实时性.仿真实例验证该方法的有效性.

Strategy of Centralized Optimization and Decentralized Control for Two-layered Predictive Control in Large-scale Industrial Systems

Aiming at reducing the on-line computational burden of model predictive control (MPC) for large-scale systems and maintaining the global optimization performance, a new two-layered MPC strategy is proposed. In the steady-state target calculation (SSTC) layer, the centralized optimization of the system is based on the global process model. The optimal results are sent to the dynamic control layer as the set points. In dynamic control layer, the large-scale system is divided into several subsystems which will be controlled with separate MPCs. To compensate the interference between the subsystems, feedforward controllers are added between the subsystems, so that the overall dynamic control performance of the system will be improved. The advantages of the proposed strategy are that, the global optimal performance of the system is guaranteed, while the on-line computational complexity is reduced, and the realtime property of the two-layered MPC for large industrial system is enhanced. Simulation examples are given to verify the proposed method.