网络攻击下虚拟电厂互动影响评估与深度弹性调控方法

Considering the significant feature of cyber-physical fusion of virtual power plants (VPP) in Energy Internet and the great harm caused by cyber attacks, the impacts of cyber attacks on market transaction and system operation and control of VPPs will be researched in the framework of the integrated cyber-physical system (CPS) in this project as well as the scheduling-control countermeasures. The research contents mainly include: 1. Establishing the GAN-based “Mechanism+Data” information-energy interaction model of VPPs, which will improve the accuracy and computational efficiency, and accurately reflect the information-physical interaction and the ability of flexible resources; 2. On this basis, proposing the impact analysis approach of cyber attacks on market transaction, system operation response and control of distributed energy resources (DER) of VPPs, and mastering the interaction influence mechanism; 3. Further, for the harm of cyber attacks, proposing the deep neural network (DNN)-observer based resilient distributed cooperative scheduling-control method for VPPs participating in the market interaction, which enables fast and accurate detection and estimation of cyber attacks and elimination of the impact of cyber attacks. Thus, a complete and effective solution to the problem of cyber-attack impact analysis and resilience scheduling-control will be established. The proposed theoretical method will be verified via simulation and experiment through the constructed hardware-in-the-loop simulation system. The research results of this project will lay a theoretical and technical foundation for the safety market trading, system operation, scheduling and control of VPPs.

针对能源互联网中虚拟电厂的信息-物理深度融合特征及网络攻击带来的巨大危害，本项目在信息物理一体化系统框架下研究网络攻击对虚拟电厂市场交易与系统运行的影响及运行调控对策。主要研究内容包括：1.基于GAN构建虚拟电厂的“机理+数据”信息-能量交互模型，提升其精确性与运算效率，并准确反应信息物理交互作用和灵活性资源调控能力；2.在此基础上，提出网络攻击对虚拟电厂市场交易、系统运行响应和DER控制的影响分析方法，掌握其交互影响机理；3.进一步，针对网络攻击的危害，研究提出基于DNN和观测器的虚拟电厂参与市场互动的弹性分布式协同调控方法，使其能够快速精准检测和估计网络攻击，消除网络攻击的影响。从而构成一套完整有效的网络攻击影响分析与弹性调控解决方案。结合实际对象，在已构建的半实物仿真系统上，对所提理论方法进行仿真和应用验证。项目研究成果将为虚拟电厂安全市场交易与系统运行调控奠定理论与技术基础。

针对网络攻击和低惯量给虚拟电厂及系统带来的巨大危害，本项目在信息物理一体化系统框架下研究了网络攻击和低惯量对虚拟电厂优化调控及系统安全稳定运行的影响及对策。主要研究内容包括：1.基于深度神经网络技术，研究和提出了“机理+数据”的虚拟电厂聚合等效鲁棒动态模型的建立与惯量时空分布的估测方法，并准确反应灵活性资源调控能力，解决了系统建模和感知问题；2.推导和分析了网络攻击与惯量时空分布对虚拟电厂及系统安全稳定的影响机理；3.针对网络攻击和低惯量的危害，提出了基于虚拟电厂参与频率控制的虚拟电厂的数据驱动分布式最优鲁棒包含控制策略，提升了系统频率调节能力，提出了基于分布式观测器和自适应补偿的虚拟电厂参内DER的弹性分布式协同调控方法，消除了网络攻击的影响。项目的研究构成了一套完整有效的系统建模与感知——影响机理分析——弹性协调调控的解决方案，为虚拟电厂及系统的安全稳定运行奠定了基础。