模型-数据混合驱动的电网安全特征选择和知识发现关键技术与工程应用

随着可再生能源的大规模并网、需求响应的逐步实现,电网运行方式的复杂性和波动性不断攀升,电力系统的安全运行正面临新的需求与挑战。因此,基于人工智能技术,在广东电网建立了"模型—数据混合驱动的电网安全特征选择和知识发现平台",保证电网安全、稳定、经济运行。文中首先定义了电网安全特征和知识,阐述了模型—数据混合驱动的思想与具体实现方法,并分析了降低误差的手段;其次阐释了平台的并行计算技术;接着设计了平台的软硬件架构;最后,展示了平台在广东电网的实际应用效果,结果表明:(1)从运行规则制定层面,将运行专家离线制定粗放运行规则的模式,变革为人工智能在线发现精细运行规则的模式;(2)从运行规则应用层面,将调度员人工判定运行规则的模式,变革为人工智能实时判定运行规则的模式。

Hybrid Model and Data Driven Concepts for Power System Security Feature Selection and Knowledge Discovery: Key Technologies and Engineering Application

With the integration of large-scale renewable energy and the implementation of demand response, the power system operation scenarios have become increasingly complicated and variable, leading to new requirements and challenges in power system security operation. Therefore, based on artificial intelligences, a hybrid model and data driven platform for power system security feature selection and knowledge discovery is established in the Guangdong Power Grid in China, to keep power system in secure, stable and economic conditions. Firstly, the definitions of power system security feature and knowledge are put forward, the concepts for the hybrid model and data driven platform are described, and the methods for reducing error are analyzed. Secondly, parallel techniques for the platform are discussed. Thirdly, the software and hardware architecture of the platform is designed. Finally, the application results in the Guangdong Power Grid are demonstrated, which show that: (1)from the perspective of rule-making, the platform transfers the pattern that conservative operation rules should be made by expert offline to the pattern that specific operation rules can be discovered by artificial intelligences online; (2)from the perspective of rule-application, the platform transfers the pattern that operation rules used online should be determined by operators artificially to the pattern that the operation rules can be determined by artificial intelligences automatically.