Autonomous Energy Consumption Management and Load Control in the Smart Grid: a Control-Theoretic Perspective

本研究计划旨在制定智能电网背景下客户侧（需求侧）自主能耗管理（ECM）和负荷控制（LC）的策略。近年来，来自科学界、工业界和公共部门的一项极其重要的努力致力于智能电网范例的发展，旨在升级电力网络，以确保在下个世纪高效可靠的能源生产和输送，同时最大限度地减少和推迟对基础设施的大规模投资。众所周知，智能电网的概念涉及电力网络中的各种元素，从发电到传输、分配和消费。尽管电力网络中不同部分之间存在强烈的相互作用，但该研究计划将重点放在消费部分。为此，研究将在需求侧管理（DSM）的框架下进行。******建议的研究将基于我们最近在相关主题的研究中开发的能源消耗管理和负载控制的分层架构。该体系结构可以封装需求侧能源管理的功能，保证各个组件之间的互操作性，允许不同能源的集成，简化系统维护和升级，并在智能电网的背景下为公用事业和消费者之间的交互提供方便的手段。在本研究计划中，我们将模型预测控制（MPC）方法应用于恒温控制负载（tcl）和暖通空调系统的运行。然后，我们将开发不同类型负载的调度算法以及储能和可再生能源的整合。为了提供一个全面的解决方案，我们将解决与处理有关的问题，特别是需求响应（DR），以实现公用事业和最终用户之间的动态交互。我们还将讨论与实际应用中提出的问题相关的问题，特别是在非线性、建模误差、参数不确定性和通信不完善的情况下的性能保证。为此，我们将考虑鲁棒自适应MPC和李亚普诺夫理论的应用。本项目的另一个研究领域是tcl群体聚集的建模与控制。由于这个问题可能导致由偏微分方程（PDEs）描述的模型，因此将特别努力应用偏微分方程控制理论和技术来实现最优的分布式控制方案。最后，我们计划建立软件平台，以验证开发的解决方案，并通过工业合作对现实系统进行实验。

This research program aims at developing strategies for autonomous energy consumption management (ECM) and load control (LC) from costumer's side (demand side) in the context of the smart grid. In recent years, an extremely important effort from the scientific community, industry, and public sectors has been devoted to the development of the paradigm of smart grids aiming at upgrading electric power networks to ensure an efficient and reliable energy production and delivery in the next century while minimizing and postponing massive investments on infrastructures. It is known that the concept of the Smart Grid is concerned with a wide variety of elements in electric power networks, from power generation to its transmission, distribution, and consumption. This research program will focus on the segment of consumption, despite the strong interactions between different segments in an electric power network. To this aim, the research will be carried out in the framework of Demand-Side Management (DSM). ******The proposed research will be based on a layered architecture for energy consumption management and load control developed in our recent research on the related topics. This architecture can encapsulate the functionality of demand-side energy management, assure the interoperability between various components, allow for the integration of different energy sources, ease system maintenance and upgrading, and provide a convenient means for the interaction of utilities and consumers in the context of the Smart Grid. In the proposed research program, we will apply the method of Model Predictive Control (MPC) to the operation of thermostatically controlled loads (TCLs) and HVAC (heating, ventilation, and air-condition) systems. We will then develop algorithms for the scheduling of different type of loads and the integration of energy storage and renewable energy sources. To provide a comprehensive solution, we will address the issues related to handling, in particular, Demand-Response (DR) to enable a dynamic interaction between utilities and end-use consumers. We will also address issues related to the concerns raised in practical applications, in particular performance guarantees in the presence of nonlinearity, modeling errors, parametric uncertainties, and communication imperfection. For this purpose, we will consider the application of robust and adaptive MPC and Lyapunov theory. Another research venue of this program is the modeling and control of aggregated large population of TCLs. As this problem may lead to models described by partial differential equations (PDEs), a particular effort will be put on the application of PDE control theory and techniques to achieve optimal distributed control schemes. Finally, we plan to build software platforms for the validation of the developed solutions and to conduct experimentations with real-life systems via industrial collaborations.