基于大语言模型的新型电力系统生成式智能应用模式初探  

Exploration of Generative Intelligent Application Mode for New Power Systems Based on Large Language Models

作　　者：丁俐夫 陈颖[1] 肖谭南 黄少伟[1] 沈沉[1] DING Lifu;CHEN Ying;XIAO Tannan;HUANG Shaowei;SHEN Chen(Department of Electrical Engineering,Tsinghua University,Beijing 100084,China;State Key Laboratory of Power System Operation and Control(Tsinghua University),Beijing 100084,China)

机构地区：[1]清华大学电机工程与应用电子技术系,北京市100084 [2]新型电力系统运行与控制全国重点实验室(清华大学),北京市100084

出　　处：《电力系统自动化》2024年第19期1-13,共13页Automation of Electric Power Systems

基　　金：国家自然科学基金联合基金资助项目(U22B2096)。

摘　　要：新型电力系统数字化转型需要智能化应用体系支撑。首先,文中探讨了新型电力系统智能应用构建的关键难题,提出了基于大语言模型的新型电力系统生成式智能应用模式及其逻辑框架,并探讨了此模式下实现智能应用全生命周期自动化构建的技术方案;其次,针对系统仿真需求,介绍了典型生成式智能应用案例,并说明了相关技术方案特点;最后,展望了生成式智能应用发展的趋势和研究方向,供后续研究者借鉴。The digital transformation of new power systems needs the support of intelligent application system.Firstly,the critical challenges in the construction of intelligent application for the new power system are discussed.A generative intelligent application mode and its logical framework for new power systems based on large language models are proposed.And the technical schemes for realizing the automatic construction of intelligent application in the entire lifecycle in this mode are explored.Secondly,according to the requirements of system simulation,typical cases of generative intelligent applications are introduced,and the characteristics of the related technical schemes are explained.Finally,the development trends and research directions of generative intelligent applications are prospected,which provide reference for future researchers.

关 键 词：新型电力系统 大语言模型 智能应用 预训练模型 数字化转型 

分 类 号：TM73[电气工程—电力系统及自动化]