Distributed Model Predictive Control for Building Energy Systems

建筑供暖、通风和空调系统由许多不同的设备组成，这些设备是单独控制的。通过实时协调这些系统的操作，可以显著节省能源。然而，建筑系统不断变化的性质要求这种协调控制策略具有灵活性和模块化，能够适应新设备或建筑操作的变化。该奖项支持基础和应用研究，以创建和演示能够协调建筑系统的算法，以减少能源使用，同时保持或提高建筑物居住者的舒适度。这些研究的结果将通过降低建筑能源成本和环境影响直接使美国受益。该项目还将有助于扩大代表性不足的群体参与工程研究。本项目研究的控制策略层次包括较低层次的控制策略，这些控制策略补偿了建筑能源子系统固有的非线性行为，同时保留了当前设备常见的直观控制器设计和调谐方法。研究还探讨了一种监督式分布式预测控制算法，该算法能够为低层控制器确定系统级能量最优参考轨迹。分布式控制器需要系统之间最少的通信，并且是模块化的，可以适应建筑设备或操作条件的变化。协调控制器最大限度地减少能源消耗，同时保持乘员的舒适性。通过在预测控制器中包含乘员需求，系统还可以向用户提供与特定行为相关的能源成本反馈，并帮助塑造乘员行为。

Building heating, ventilation, and air-conditioning systems are comprised of many different pieces of equipment, which are controlled individually and separately. By coordinating the operation of these systems in real-time, significant energy savings are possible. However, the changing nature of building systems require that this coordinated control strategy be flexible and modular, with the ability to adapt to new equipment or changes in building operations. This award supports both fundamental and applied research to create and demonstrate algorithms capable of coordinating building systems to reduce energy usage, while maintaining or improving the comfort of building occupants. Results of these research will directly benefit the United States by reducing building energy costs and environmental impact. The project will also help broaden participation of underrepresented groups in engineering research.The hierarchy of control strategies investigated in this project includes lower level control strategies that compensate for the inherently nonlinear behavior of building energy subsystems, while preserving the intuitive controller design and tuning methods common to current equipment. The research also investigates a supervisory distributed predictive control algorithm that is capable of determining to the system-level energy optimal reference trajectories for lower-level controllers. The distributed controller requires minimal communication between systems, and is modular and adaptable to changes in building equipment or operational conditions. The coordinated controller minimizes energy consumption while maintaining occupant comfort. By including occupant demands in the predictive controllers, the system can also provide feedback to users regarding the energy costs associated with specific behaviors, and help shape occupant behavior.