摘要
在能源技术变革日新月异、人工智能技术快速发展的背景下,研究具有高适应性、高精度的机组组合决策方法具有重要意义。该文基于长短时记忆网络,通过构建面向机组组合问题的深度学习模型,提出一种基于数据驱动具有自我学习能力的机组组合智能决策方法。首先基于K-means算法对历史调度数据进行聚类预处理;然后构建基于长短时记忆网络的机组组合深度学习模型,通过历史数据训练建立系统负荷与调度决策结果之间的映射模型,以此为基础进行机组组合决策;最后通过积累历史数据实现对模型的持续修正,从而赋予其自我进化、自我学习的能力。基于标准算例、实际电网数据的一系列仿真结果表明:相比于传统决策方法,该方法不仅可以在实际使用过程中不断提升其决策精度或效率,且在面对不同类型的机组组合问题时适应性更好。
It is now a background of the energy technologies change rapidly and the artificial intelligence technologies develop quickly. It is significant to study the dispatching methods of unit commitment with high adaptability and high precision. A data-driven intelligent securityconstrained unit commitment dispatching method with self-learning ability was proposed. In the method, the long-short term memory neural network was utilized to construct a deep learning model, which was focus on the problem of the unit commitment. Firstly, the K-means algorithm was used to do a cluster preprocessing for the historical dispatching data. Then, the unit commitment deep learning model was constructed by long-short term memory neural network. The mapping model between system load and dispatching results was constructed by historical data training. After that, the all above process was used as a foundation to do the unit commitment dispatching. Finally, the model was kept revised through the accumulation of historical data, which may give it the abilities of selfevolution and self-learning. A series of simulation results based on the standard calculation example and the practical grid data indicate some information. Compared with traditional dispatching methods, the method can not only improve the accuracy and efficiency of the method in practical usage, but also adapt to different kinds of unit commitment problems.
作者
杨楠
叶迪
林杰
黄禹
董邦天
胡文斌
刘颂凯
YANG Nan;YE Di;LIN Jie;HUANG Yu;DONG Bangtian;HU Wenbin;LIU Songkai(New Energy Micro-grid Collaborative Innovation Center of Hubei Province (China Three Gorges University),Yichang 443002, Hubei Province, China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2019年第10期2934-2945,共12页
Proceedings of the CSEE
基金
国家自然科学基金项目(51607104)~~
