摘要
湍流控制涉及流体力学和控制理论,对航空航天、运载工具、风力发电等众多领域具有重要的科学意义和应用价值.由于湍流的复杂性,传统控制方法在湍流控制领域面临很多瓶颈,人工智能技术的发展为突破这些瓶颈提供了工具.本文简要综述了文献报道的有模型和无模型的人工智能控制湍流的进展,总结了研究中采用的典型的人工智能控制系统、算法、在不同湍流控制应用中取得的突出成果.作者所在团队在国际上首次尝试对人工智能控制系统产生的海量数据进行分析,从而挖掘出重要的信息乃至发现控制相似律.对面临的挑战和未来的展望进行了分析.
Turbulence control involves fluid dynamics and control theory,and is of great importance to many fields such as aeronautics and astronautics,vehicle,wind power generation,etc.Due to the complexity of turbulence,traditional control methods face many bottlenecks in the field of turbulence control.The development of artificial intelligence(AI)technology provides a tool to break through these bottlenecks.This paper briefly summarizes the applications of AI in turbulence control reported in the literature,focusing on AI control systems,algorithms,and the outstanding achievements achieved in different turbulence control applications,as well as the first attempt by the author's team to analyze the big data generated by the AI control system to discover important information and even the control scaling law.The challenges and future prospects are also analyzed.
作者
吴智
范德威
周裕
WU Zhi;FAN De Wei;ZHOU Yu(Center for Turbulence Control,Harbin Institute of Technology(Shenzhen),Guangdong,China)
出处
《力学进展》
EI
CSCD
北大核心
2023年第2期273-307,共35页
Advances in Mechanics
基金
自然科学基金委通过项目编号91952204、12202123
深圳科创委通过项目编号JCYJ20190806143611025、JCYJ20210324132816040和GXWD20220811152110003所给予的支持。
关键词
人工智能
大数据
流动控制
湍流
神经网络
遗传编程
深度强化学习
s artificial intelligence
big data
flow control
turbulence
neural network
genetic program-ming
deep reinforcement learning