机器学习在能源与电力系统领域的应用和展望

新一代人工智能(AI)近年来成为国内外研究的热点,其中的典型代表机器学习(ML)作为一个算法范畴,通过分析和学习大量已有或生成数据形成预测和判断以做出最佳决策。中国的新一代AI正处于快速发展的关键期,目前已在能源与电力系统中得到初步应用。基于此,文中以新一代AI中的ML为代表,重点综述了强化学习、深度学习、迁移学习、平行学习、混合学习、对抗学习和集成学习等7种代表性ML在能源与电力系统调度优化和控制决策等方面的应用。最后,对未来ML的发展进行了思考与展望。

Machine Learning for Energy and Electric Power Systems: State of the Art and Prospects

The new generation of artificial intelligence(AI), i. e. , AI2. 0, has become a research highlight in recent years. Among AI2. 0, machine learning(ML)as a typical representative is an algorithm category that completes predictions and judgments for optimal decision-making through analyzing and learning a large amount of existing or generated data. AI2. 0 is developing rapidly in China, and it has been preliminarily applied to the energy and electric power system(EEPS)that contains smart grid(SG)and energy interconnection(EI)fields. To this end, this paper takes ML in AI2. 0 as an example to review the current application of seven representative MLs in EEPS from aspects of dispatch optimization and control decision-making, including reinforcement learning, deep learning, transfer learning, parallel learning, hybrid learning, adversarial learning, and ensemble learning. Finally, the prospects for the future development of ML are conducted, trying to provide some reference for the theoretical, technical and application studies of AI2. 0, especially ML in the field of EEPS in the future.