基于串联结构的分布式模型预测控制

分布式模型预测控制(Distributed model predictive control, DMPC)是一类用于多输入多输出的大规模系统的控制方式.每个智能体通过相互协作完成整个系统的控制. 已有的分布式预测控制算法可以划分为迭代式算法和非迭代算法:迭代算法在迭代到收敛情况下,具有集中式预测控制(Centralized model predictive control, CMPC)算法的性能,但迭 代次数过多,子系统间通信量大;非迭代算法不需要迭代,但性能有一定损失.本文提出了一种基于串联结构的非迭代分布式预测控 制算法.本文算法在串联结构系统中可以有效减少计算量,并结合氧化铝碳分解(Alumina continuous carbonation decomposition process, ACCDP)这一串联过程,通过仿真验证了算 法的有效性;同时分析了算法运用在串联结构下的性能并证明了其稳定性.

Distributed Model Predictive Control Based on Cascade Processes

Distributed model predictive control (DMPC) is a useful control theme which is usually used to control large scale systems with multiple inputs and multiple outputs. Every agent communicates with the other in order to control the whole system. The algorithms for distributed model predictive control can be divided into two categories, one is iterative and the other is non iterative. The iterative ones can reach the same performance as the centralized model predictive control (CMPC) when they converge, however, because of the large number of iterations, the communication burden is heavy; while the non iterative ones do not need iteration, the performance is not as good as centralized algorithms. This article proposes a non iterative algorithm of distributed model predictive control based on cascade processes. Our algorithm can save computational burden for cascade processes. Finally, the alumina continuous carbonation decomposition process (ACCDP) is used to prove the effectiveness for the algorithm. We also analyse the performance and proof the stability of the algorithm.