The conventional multi-zone HVAC (Heating Ventilation and Air Conditioning) system includes a lot of Variable Air Volume (VAV) boxes to regulate the corresponding zone temperature, and centralized equipment (air supply fans and ducts, etc.) to supply conditioned air in the air-side. With the limitation of the total supply air mass rate in centralized equipment and the priority of zones’ requirement, it becomes a question of resource allocation and schedule on how to coordinate air mass rate to each zone. It is also intricate to control each zone’s temperature in a fully distributed fashion due to this global constraint. In this paper, a distributed model predictive control (DMPC) with priority coordination is proposed for multi-zone HVAC systems. This strategy is composed of a priority coordinated layer and a lower DMPC layer. In the coordinated layer, a thermal bound calculation mechanism with priority is presented to adjust thermal bounds for DMPC to resolve the resource allocation problem when the subsystems are competing for the resource. In DMPC, an operational cost-saving controller is put forward, and the global resource limit is guaranteed by transforming the optimization problem into a consensus problem, which is running in a fully distributed way. In this control scheme, the original infeasible DMPC problem is relaxed, and the entire system is functioning well with specified performance compromise. The effectiveness of the proposed scheme is proved by simulation examples.

传统的多区域 HVAC（暖通空调）系统包括大量的可变风量（VAV）箱来调节相应的区域温度，以及集中设备（送风风扇和管道等）.) 在空气侧供应调节空气。由于集中设备总送风量率的限制和分区需求的优先级，如何协调各分区的风量率成为资源分配和调度的问题。由于这种全局约束，以完全分布的方式控制每个区域的温度也很复杂。在本文中，针对多区域 HVAC 系统提出了一种具有优先协调的分布式模型预测控制 (DMPC)。该策略由一个优先协调层和一个较低的DMPC层组成。在协调层中，提出了一种具有优先权的热边界计算机制，用于调整DMPC的热边界，以解决子系统竞争资源时的资源分配问题。在 DMPC 中，提出了一种节省运营成本的控制器，通过将优化问题转化为共识问题来保证全局资源限制，该共识问题以完全分布式的方式运行。在这种控制方案中，原来不可行的 DMPC 问题得到了缓解，整个系统在特定的性能折衷下运行良好。通过仿真算例证明了所提方案的有效性。