广大师生：

华南理工大学“海内外优秀青年学者论坛”，旨在面向全球邀请拥有不同学术背景的青年才俊，围绕国际科学前沿、热点研究领域以及行业产业的技术问题等展开探讨和交流。通过这个平台，互相启迪、开拓视野，增强国际交流与合作，促进双方共同发展。

一、论坛时间：2024年1月3日（星期三）

二、论坛地点：华南理工大学五山校区9号楼225

三、论坛议程

欢迎广大师生参加！

                                                       华南理工大学电力学院

                                                                                                                                          2023年12月29日

学术报告1

讲座题目：Applications of Reinforcement Learning on Power Market Simulation and Power System Dispatch

内容摘要：As China's electricity market continues to undergo de-regulation reforms, the design and optimization of the electricity market mechanism have emerged as critical challenges. An effective electricity market mechanism ensures incentive compatibility, where market participants, while pursuing their self-interests, inadvertently contribute to the maximization of market operational efficiency. This synergy enhances the fairness and liquidity of the electricity market. However, the current design methods for electricity market mechanisms are predominantly based on theoretical arguments, without reliable simulation as verifications. This gap often results in suboptimal real-world implementation and hinders the development of China's electricity market to some extent.

This report introduces a suite of electricity market simulation algorithms based on multi-agent reinforcement learning techniques, proposing an optimization method for market mechanisms derived from these simulations. This approach enables the modeling of power producers' decision-making behaviors and market equilibrium under various market mechanisms. Its architecture supports the flexible modification of market rules, allowing for the simulation of electricity markets under any national context or mechanism. Additionally, the multi-agent reinforcement learning approach offers a distributed computational framework and privacy safeguards, aligning the algorithm's operation with the fundamental principles of electricity market functioning. This demonstrates the reliability and interpretability of the algorithm. The simulation results produced by this algorithm enable regulatory agencies and trading centers to estimate market equilibrium and performance metrics under different market designs, thereby optimizing market design options and enhancing market efficiency.

Finally, the report will discuss future research prospects of reinforcement learning in power system dispatch.

报告人简介：朱子晴，香港浸会大学计算机系研究助理教授，博士毕业于香港理工大学，硕士毕业于英国曼彻斯特大学，主要研究方向是人工智能与机器学习在电力系统优化运行和电力市场中的应用。目前以第一/通讯作者身份发表13篇SCI论文，其中8篇JCR一区/中科院一区论文，并曾获得IET APSCOM会议最佳论文奖，IEEE Transactions on Sustainable Energy杰出审稿人奖。

学术报告2

讲座题目：Nonlinear Behavior of Grid-Connected Converters and Stability of Converter-Based Systems

内容摘要：With the increasing use of power electronics in the power grid, the dynamics of the grid are significantly altered, which may affect the normal operation of the grid and increase the likelihood of instability events and power blackout. There are two parts in this talk. In the first part, some instability phenomena, such as sub-synchronous oscillations, will be traced back to the essential complex nonlinear behavior of the power electronic converters. I will show the dual properties of the homoclinic bifurcation behaviors in the grid-forming and grid-following converters, and the impact of the homoclinic bifurcation on the transient synchronization stability. Moreover, the new route to generate the sub-synchronous oscillations will be demonstrated in terms of hidden attractors, and the physical characteristics of these hidden attractors will be presented as well. In order to study the large-scale multi-converter systems, we need to develop appropriate model order reduction techniques for grid-forming/following converters and transmission networks. Some research results will be summarized in this talk. Besides, focusing on the active loads, I will show the connections between the converters’ bifurcation behavior and the grid’s voltage stability. In the second part of this talk, the stability of multi-input-multi-output voltage balance control systems, the aliasing phenomena under commonly used modulation-waveforms updating schemes, and the noise propagation mechanism and related quantitative index will be presented, which may be extended to study the characteristics of the cyber-physical multi-converter systems.

报告人简介：杨景熙，香港城市大学副研究员，本科及博士毕业于北京交通大学电气工程学院，主要从事并网变换器的控制、电力电子化电网的非线性动力特征和稳定性等方面的研究。自2018年起以第一/通讯作者发表一区Top期刊论文9篇，谷歌学术引用率670+，参与申请10余项专利，核心参与国家自然科学基金重点项目、国家重点研发计划等。2017年获得第三届“直流输电与电力电子创新杯大赛”二等奖，2019年获得北京交通大学五四奖章，2021年被评为IEEE Transactions on Power Electronics最佳审稿人，2022年被评为Journal of Modern Power Systems and Clean Energy最佳审稿人，担任2024 ECCE Asia专栏联合主席，担任IEEE PECAS技术委员会成员，担任多个IEEE Top期刊、MPCE、CSEE JPES等的审稿人。

学术报告3

讲座题目：Energy Storage Overview and State-of-the-Art

内容摘要：Using low-carbon renewable energy to reduce carbon emissions has become a general trend and a pressing need. However, most renewable energy is intermittent and therefore its efficient utilization relies on highly reliable energy storage technologies. In the presentation, ten different mainstream energy storage technologies will be overviewed and discussed, including their key problems and solutions.

报告人简介：田原，英国德蒙福特大学副教授，机械系主管教学副主任，华威大学博士。曾任赫特福德大学和阿斯顿大学讲师（终身教职），博导和课程主任。为英国高等教育协会 Fellow，国际能源署储能组专家，英国储能智囊团科技委员会委员，英国物理和工程科学研究基金会（EPSRC）高级评审（优秀评委奖），英国科学与技术设施理事会（STFC），英国医学研究基金会（MRC），智利科技部（CONICYT）和欧盟委员会等多国科研基金项目评审专家。在新能源高效存储等领域发表学术论文50篇，被引用 6200余次，担任六十余家国际能源类学术杂志的编委或审稿人。主讲课程：新能源、流体力学、传热学和热力学。