风险约束下智能电网主动自愈策略构建的基础性研究

Self-heal, in essence, the immune system, is well identified as the most desirable future of the smart grid. This proposed research project addresses the systematic methodology for a self-healing smart grid construction. To cover the functional requirements of a self-healing smart grid, two novel concepts, i.e., risk-limiting self-healing methodology and proactively self-healing strategy, are proposed and implemented. The comprehensive self-healing methodology is partitioned into five coherent tasks in this proposal. (1) A set of coherent risk metrics and associated PMU-based information acquisition methods are proposed. Mathematical characteristics of the risk metrics will be studied and, furthermore, the emerging risks, as well as potential risks can be identified. (2) A comprehensive self-healing methodology will be proposed. As, a generic and adaptive methodology, it can meet the requirements of system self-heal for different systems and one system with different initial states. Risk-limiting concept and associated approaches will be employed to design the sequential self-healing actions with acceptable risk level of the current action, as well as the remaining actions of the whole process. The proactively self-healing strategy will be designed, i.e., not only the emerging faults will be eliminated, but also potential vulnerability will be mitigated. (3) Sophisticated algorithms for stochastic analysis methods will be proposed to design the risk-limiting proactively self-healing methodology for systems with large-scale renewables and responsive loads integration under new energy paradigm. A set of high-efficient scenario reduction algorithms and stochastic simulation methods will be employed to accomplish this task. (4)A high-efficient CPU-GPU based computing infrastructure will be designed to implement the proposed methods. After identify the characteristics of CPU and GPU computing infrastructure, the computing loads for a self-healing strategy construction will be dispatched to CPU and GPU according to their requirements. (5) Finally, a prototype of the proposed methodology and computing framework will be established. Standard testing and real systems will be used to demonstrate the accomplishments...The present proposed research is built on the experience of our past efforts and is more ambitious. A comprehensive risk-limiting proactively self-healing methodology associated with a set of sophisticated algorithms and high-efficient CPU-GPU based computing infrastructure will be established. It is believed that the accomplishments will be a major step in establishing a comprehensive self-healing smart grid and, ultimately, putting the dream of a self-healing smart grid into practice.

自愈是未来智能电网的核心特征。本项目将系统化地研究电网自愈的基础理论，提出基于风险约束的电网主动自愈方法论，主要目标包括：提出一致性的系统风险测度和信息获取机制，实现系统的风险识别和脆弱性评估，为启动主动自愈并监控自愈过程提供依据；提出基于风险约束的主动自愈策略构建的一般性理论，以自愈过程总体风险为目标，实现显性故障的自愈和潜在脆弱性的消除；提出和改进一系列高效的随机分析方法，实现包含可再生能源和负荷响应等未来智能电网本质特征的系统自愈策略构建，核心技术包括随机场景消减，马尔科夫模拟，随机点估计等；构造CPU和GPU结合的高效计算平台，将自愈策略构建的计算需求合理的分配到CPU和GPU中，实现软硬件结合的高性能计算；在标准和实际系统中完成提出理论和技术的测试。.本课题基于申请人以往大量的国际、国内相关研究经验，其成果将为自愈电网的构建提供理论基础、分析方法和实现手段。

自愈是未来智能电网的核心特征。本项目系统化地研究了电网自愈的基础理论，提出基于风险约束的电网主动自愈方法论。研究成果覆盖主动自愈的全过程，包括：适于自愈的信息获取和风险评估方法，脆弱性的主动消除策略、故障的被动的故障恢复策略以及实现主动自愈策略的核心算法。取得了以下的研究成果： .提出了适用于自愈电网的系统的信息获取和风险评估的方法。首先，建立了互联输电系统在扰动后动态状态识别和估计的方法；第二，提出了输电系统在持续序贯扰动下的灵活域刻画和运行风险的定量评估策略；第三，提出基于自组织临界理论的输电系统风险测度和脆弱环节识别策略；第四，建立了基于系统最大安全运行时间的运行风险评估测度方法。.提出了系统脆弱性的消除策略，实现主动自愈。首先，提出了考虑极端运行环境条件的输电网检修策略，消除潜在的脆弱性；第二，提出考虑极端天气条件下的输电网主动序贯调整策略，减小潜在损失；第三，提出了考虑信息系统脆弱性的主动防御策略；第四，提出微电网脆弱性主动消除策略，减小潜在损失。 .提出了系统化的恢复策略，实现系统的被动自愈。首先，提出了考虑冷负荷、负荷分块特性、系统频率响应特性和发电机爬坡特性的输电系统恢复策略；第二，提出了基于开关模型和原件的可用性的输电网恢复策略构建；第三，提出了配电系统中联合移动发电设备的故障后恢复策略。 .构建自愈策略的高效计算技术。首先提出了基于矢量化技术的含序贯约束的自愈策略高效求解方法；第二，提出了基于潮流方程几何特征的高效自愈策略求解方法；第三，建立了基于CPU-GPU混合的自愈策略验证平台。 .基于理论研究成果，开发了电力系统自愈决策支持系统-System Restoration Navigator，由美国电力科学研究院在北美大规模推广应用。依托本项目培养博士生4名，博士后两名，发表论文27篇，其中SCI检索12篇。出版英文专著一本。.本项目的成果为自愈电网的构建提供理论基础、分析方法和实现手段，并最终可望使自愈电网的梦想逐步变为现实。