虚拟电厂是新型电力系统获取灵活性、保证安全消纳新能源的关键手段。通信技术是虚拟电厂的重要支撑，亟需从信息物理耦合角度，重新审视其运行和调控机制。虚拟电厂运行需考虑通信网的实际限制因素；通信系统也应在准确刻画海量异质信息重要性及语义的基础上，实现海量数据高效传输并主动协同虚拟电厂，保证其安全、可靠、经济运行。本文对集中控制型、集中协调型、分散联邦型虚拟电厂中的信息物理协同调控问题展开研究，主要成果如下： 本文解析了通信网络与虚拟电厂决策、调控耦合的机理和作用路径，构建了考虑信息异质性的虚拟电厂信息物理协同调控优化模型，借由分解方法提出了解析异质数据传输特性变化对于虚拟电厂影响的定量分析手段，并在信息维度创新地提出了通信协同调控方法，从新的科学视角为耦合协同运行奠定了理论基础。针对集中控制型虚拟电厂的分钟级精准需求响应服务，本文解析了异质信息的可达性与资源可观测性、可调控性的耦合机理，定量表征了信息传输对聚合调控效果的影响，并从耦合角度建立了分布式资源无线通信接入的协调优化模型及其并行求解算法，最大限度减少因信息不可达而导致的不可靠响应损失。 针对集中控制型虚拟电厂的秒级紧急频率支撑需求，本文提出了计及时滞的紧急频率支撑性能参数化评估方法，建立了内嵌性能约束的备用成本优化模型，解析了分布式资源时滞变化对聚合成本的影响，并构建了通信协同时滞优化模型，以提升虚拟电厂资源的快速响应能力，同时减少由实时时滞波动引起的风险损失。针对集中协调型虚拟电厂中的高并发平台能量共享需求，本文提出了面向海量用户的有限连接度的平台能量共享算法，在有效减少了通信复杂度的基础上防止了通信拥塞；并从理论上解析了用户的异质信息对能量共享迭代效率的影响，提出了优化的并发连接协同调控方法，显著提高了能量共享迭代效率。针对分散联邦型虚拟电厂中的高并发对等能量交易需求，本文提出了面向大规模用户自主对等交易的异步算法，赋予用户在交易中进行同步链路选择的权利；理论解析了不同交易用户异质同步信息对交易竞争均衡收敛的影响，提出了异步高频交易下的同步链路协同调控方法，显著减少均衡迭代轮数和总时长。本文为解析虚拟电厂信息物理耦合机理、建立信息物理协同调控方法提供了理论依据和科学方法，为实现可靠、可信、可持续的虚拟电厂发展提供有益参考。

Virtual power plants (VPPs) are crucial for realizing flexible future electrical systems and securely integrating renewable generators. Communication technology stands as a central support for VPPs. It is pivotal for VPPs to reshape their operational and control paradigms from a cyber-physical perspective. On one hand, VPPs should incorporate communication networks' effects into their operation and control strategies. On the other hand, the communication system should consider the heterogeneous nature and semantics of data from various distributed energy resources (DERs) and users, which is key to achieving effcient data transmission and actively support VPPs. Thus, VPPs' safety, reliability, and economic efficiency can be ensured. This thesis explores challenges in cyber-physical collaborative control across various VPP types, including centralized, distributed, and federated VPPs, presenting key contributions:This thesis delves into the interplay between communication networks and VPPs' decision-making and control, revealing underlying mechanisms and coupling effects. It develops an optimization model for cyber-physical coordination, considering information semantics, and introduces a quantitative approach to evaluate data transmission effects on VPPs via the decomposition method. Furthermore, it introduces a method for collaborative communication control in support of VPPs. They together offer a new perspective and establishes a new foundation for cyber-physical system operation and control. This thesis first focuses on the minute-level accurate demand response services offered by centralized virtual power plants, exploring how communication unreliability affects their observability and controllability. It presents a quantitative method to measure the effects of unreliable communication. It also introduces an optimization model for coordinating distributed energy resources' wireless access and a parallel algorithm to mitigate communication failures' adverse effects.For centralized VPPs offering rapid second-level emergency frequency support, this thesis introduces a parametric method for evaluating emergency frequency support performance that accounts for time delays. It formulates an optimization model to minimize VPPs' reserve costs with performance constraints and examines how delays affect overall costs. Additionally, a communication coordination delay optimization model is established to refine response performance and minimize risks associated with fluctuations in delays. Responding to high-concurrency energy sharing in distributed VPPs, this thesis introduces a novel algorithm for numerous users under limited connectivity. This approach significantly reduces communication complexity and prevents congestion. It theoretically analyzes how the information from different users affects energy sharing efficiency. It then puts forward a strategy for communication connection control apporach that significantly enhances communication efficiency. For federated VPPs involved in peer-to-peer energy trading, this thesis introduces an innovative fully asynchronous algorithm for autonomous trading among a large number of users. It explores the impact of users' heterogeneous information on searching for the equilibrium of competitive trading, and proposes a synchronization coordination method for peer-to-peer connection selection, which substantially reduces the time and iterations needed for achieving trading equilibrium. This thesis offers a theoretical foundation and scientific methodologies for understanding cyber-physical coupling in VPPs and establishing collaborative communication control strategies. It contributes valuable insights into the development of reliable, efficient, and sustainable VPPs.