摘要
优化问题是电力系统领域中最常见的问题之一,大多数工程问题都可以归结为优化问题。为了解决这些优化问题,人们不断地提出各种优化算法。然而,随着人类需求的快速提高和可再生能源的高度渗透,高度复杂的优化问题不断出现,使得传统的优化算法在求解精度和速度上都面临着巨大挑战。近年来,新一代人工智能(artificial intelligence,AI)的发展为优化算法的升级创造了新的机遇。智能电网是未来电网的发展方向,对电网的安全性、稳定性、可靠性、弹性、可持续性、高效性等都有很高的要求。这些要求使得智能电网成为一个具有多目标和多约束的高度复杂的优化问题。最典型的问题是潮流控制。无论是使用传统的还是现代的AI优化算法,已有许多研究人员致力于解决各种潮流控制问题。该文旨在全面、清晰地介绍这一研究领域的最新进展。首先,介绍优化算法的进化过程、分类和最先进的方法。然后,全面回顾传统和现代AI优化在控制潮流可解性、功率与频率、电压、安全性和稳定性方面的应用。这些应用均取得了良好的性能。尽管如此,该文仍然在高效性、可解释性、可迁移性、稳定性、经济性和鲁棒性方面点明一些关键挑战。为了克服这些挑战,给出在未来的研究中有价值的潜在方向,以使现代AI优化更适用于实际工程。
The optimization problem is one of the most common problems in the power system field,and most engineering problems can be attributed to optimization problems.In order to solve these optimization problems,various optimization algorithms have been proposed.However,with the rapid improvement of human needs and highly renewable energy penetrations,highly complex optimization problems appear constantly,making the traditional optimization algorithm challenging to meet the needs in accuracy and speed.In recent years,the development of a new generation of artificial intelligence(AI)has created a new opportunity to upgrade optimization algorithms.Smart grid is the future development direction of the power grid,which has high demand in security,stability,reliability,elasticity,sustainability,and efficiency.Due to these requirements,smart grid is a highly complex optimization problem with multiple objectives and constraints,in which the most typical problem is power flow control.Many researchers have been committed to solving various power flow control problems,whether using traditional or modern AI optimization algorithms.This paper aims to give a comprehensive and clear picture of recent advances in this research area.First,the evolution process,classification,and state-of-the-art methodologies of optimization algorithms are introduce.Then,a comprehensive review of traditional and modern AI optimization applications to control solvability,power and frequency,voltage,security,and stability is given.These applications have achieved good performance.Nevertheless,some critical challenges in efficiency,interpretability,transferability,stability,economy,and robustness are put forward.To overcome these challenges,critical potential directions in future research to make modern AI optimization more suitable for practical engineering are given.
作者
王甜婧
汤涌
王兵
黄彦浩
郭强
陈兴雷
文晶
李文臣
WANG Tianjing;TANG Yong;WANG Bing;HUANG Yanhao;GUO Qiang;CHEN Xinglei;WEN Jing;LI Wenchen(Laboratory of Power Grid Safety and Energy Conservation(China Electric Power Research Institute),Haidian District,Beijing 100192,China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2023年第5期1799-1817,共19页
Proceedings of the CSEE
基金
国家自然科学基金项目(U1866602)。
关键词
优化
传统方法
人工智能
潮流
深度强化学习
optimization
traditional methods
artificial intelligence
power flow
deep reinforcement learning
