欢迎来到IDEAL Lab！
北京大学IDEAL Lab 秉承智能决策驱动可持续未来的愿景，致力于通过前沿智能决策理论与方法，推动全球能源体系向可持续、安全及低碳化转型。
核心使命
- 开展前沿研究: 聚焦可信人工智能、运筹优化大模型、大数据分析、信息-物理安全等关键领域，研究可再生能源并网、电力系统优化调度及能源系统安全运行等核心问题，提升低碳能源系统的效率、安全性与可靠性；
- 推动可持续发展: 面向"双碳"目标与全球气候治理需求，构建数据驱动的智慧能源管理方案，通过先进优化算法降低碳排放强度，促进可再生能源的规模化应用；
- 培养专业人才: 为本科生、研究生、博士后及访问学者提供高水平科研训练，培养兼具理论基础与工程实践能力的复合型人才，使其具备引领能源转型的专业素养；
- 促进学术合作: 与国内外高校、科研机构、能源企业及政府部门建立合作关系，开展跨学科协同创新，推动研究成果向工程应用转化；
- 服务政策制定: 基于定量分析与实证研究，为低碳能源发展规划、电力系统安全运行及碳减排路径等政策制定提供科学依据与决策支持。
资助机构
研究成果
By integrating cutting-edge research in artificial intelligence, data analytics, and energy systems engineering, the IDEAL Lab aims to be a global leader in fostering innovations that promote environmental sustainability, economic viability, and energy security.
Physics-integrated Data-Driven Methods for Low Carbon Energy Systems Operation and Control with Safety Guarantee
The real-world implementation of DRL-based operation and control methods is facing the following fundamental challenges: 1) performance guarantee under modeling and parameter uncertainties between simulated and real-world inverter-connected energy systems; 2) safety guarantee (do not violate the energy system's physical constraints) during the learning and decision-making processes; and 3) adaptability against the dynamic system operating conditions. To this end, we propose a series of physics-informed Adaptive and Safe-Certified DRL (AdapSafe) algorithms for system operation and control to address the aforementioned challenges simultaneously.
- X. Wan, M. Sun*(Corresponding Author), etc., “AdapSafe: Adaptive and Safe-Certified Deep Reinforcement Learning-Based Frequency Control for Carbon-neutral Power Systems”, in AAAI-23, Washington, US, 2023.
- X. Wan, M. Sun*(Corresponding Author), “AdapSafe2: Prior-free Safe-certified Reinforcement Learning for Multi-Area Frequency Control”, in IEEE Transactions on Power Systems, 2025.
- L. Zeng, M. Sun*(Corresponding Author), “Bridge the Sim-to-Real Gap for Deep Reinforcement Learning-based Frequency Control”, in IEEE Transactions on Power Systems, Under Revision.
- X. Wan, C. Yang, C. Yang, J. Song, M. Sun*(Corresponding Author), “Fuzzify Uncertainty: Leverage Fuzzy-Logic for Robust Safe Reinforcement Learning”, in NeurIPS 2025, 2025.
- Y. Wang, D. Qiu, M. Sun, G. Strbac, Z. Gao, “Secure energy management of multi-energy microgrid: A physical-informed safe reinforcement learning approach”, in Applied Energy, vol. 335, 120759, 2023.
Probabilistic Deep Learning-based Dynamic Security Assessment of Large-Scale Low Carbon Electrical System with High Interpretability
The ongoing decarbonization of modern electricity systems has led to a substantial increase in operational uncertainty, particularly due to the large-scale integration of renewable energy generation. However, the expanding space of possible operating points renders necessary the development of novel data-driven probabilistic security assessment approaches. Furthermore, a better understanding of confidence in the security assessment result is of key importance for Transmission System Operators (TSOs) to use and rely on these deep learning methods with high interpretability.
- M. Sun, I. Konstantelos and G. Strbac, "A Deep Learning-Based Feature Extraction Framework for System Security Assessment," in IEEE Transactions on Smart Grid, vol. 10, no. 5, pp. 5007-5020, Sep. 2019.
- I. Konstantelos, M. Sun*(Corresponding Author), etc., "Using Vine Copulas to Generate Representative System States for Machine Learning," in IEEE Transactions on Power Systems, vol. 34, no. 1, pp. 225-235, Jan. 2019.
- T. Zhang, M. Sun*（Corresponding Author), etc., "A Confidence-Aware Machine Learning Framework for Dynamic Security Assessment," in IEEE Transactions on Power Systems, vol. 36, no. 5, pp. 3907-3920, Sept. 2021.
- F. Bellizio, J. Cremer, M. Sun*（Corresponding Author), G. Strbac, "A Causality Based Feature Selection Approach for Data-Driven Dynamic Security Assessment", in Electric Power Systems Research, vol. 201, 107537, 2021.
- Z. Zhang, M. Sun*（Corresponding Author), etc., "Physics-Constrained Robustness Evaluation of Intelligent Security Assessment for Power Systems," in IEEE Transactions on Power Systems, vol. 38, no. 1, pp. 872-884, Jan. 2023.
Objective-based Vulnerability Assessment and Resilience Enhancement for Low Carbon Energy Systems Operation and Control
Although the benefits of using Machine learning (ML) approaches to realize the power system autonomous operation and control have been partially uncovered, the vulnerabilities or risks of ML-based power system operation and control models have not been fully explored. It has been recognized that the ML model is vulnerable to adversarial examples, and the power system communication network is prone to cyber-attacks. Moreover, most of the existing works neglect the bad data detection mechanism and energy system physical constraints among the state variables.
- L. Zeng, M. Sun*(Corresponding Author), etc., "Physics-Constrained Vulnerability Assessment of Deep Reinforcement Learning-Based SCOPF," in IEEE Transactions on Power Systems, vol. 38, no. 3, pp. 2690-2704, May 2023.
- Y. Chen, M. Sun*(Corresponding Author), etc., "Vulnerability and Impact of Machine Learning-Based Inertia Forecasting Under Cost-Oriented Data Integrity Attack," in IEEE Transactions on Smart Grid, vol. 14, no. 3, pp. 2275-2287, May 2023.
- K. Zuo, M. Sun*(Corresponding Author), etc., "Transferability-Oriented Adversarial Robust Security-Constrained Optimal Power Flow," in IEEE Transactions on Smart Grid, doi: 10.1109/TSG.2024.3397654.
- Z. Zhang, K. Zuo, R. Deng, F. Teng and M. Sun*(Corresponding Author), "Cybersecurity Analysis of Data-Driven Power System Stability Assessment," in IEEE Internet of Things Journal, vol. 10, no. 17, pp. 15723-15735, 1 Sept.1, 2023.
- L. Zeng, D. Qiu, M. Sun*（Corresponding Author）, “Resilience enhancement of multi-agent reinforcement learning-based demand response against adversarial attacks,” in Applied Energy, vol. 324, 119688, 2022.
Data-driven Energy Consumer Characterization and Privacy Protection
The wide deployment of Advanced Metering Infrastructures (AMIs) in smart energy systems provides valuable opportunities to characterize energy consumers and fully understand their consumption behaviours by exploiting the massive amount of fine-grained data and cutting-edge artificial intelligence technologies. On the other hand, consumers may not want to share their actual data information with the system operators for decision-making due to privacy concerns.
- M. Sun, T. Zhang, Y. Wang, G. Strbac and C. Kang, "Using Bayesian Deep Learning to Capture Uncertainty for Residential Net Load Forecasting," in IEEE Transactions on Power Systems. vol. 35, no. 1, pp. 188-201, Jan. 2020.
- M. Sun, I. Konstantelos and G. Strbac, "C-Vine Copula Mixture Model for Clustering of Residential Electrical Load Pattern Data," in IEEE Transactions on Power Systems, vol. 32, no. 3, pp. 2382-2393, May 2017.
- M. Sun, Y. Wang, G. Strbac and C. Kang, "Probabilistic Peak Load Estimation in Smart Cities Using Smart Meter Data," in IEEE Transactions on Industrial Electronics, vol. 66, no. 2, pp. 1608-1618, Feb. 2019.
- M. Sun, Y. Wang, F. Teng, Y. Ye, G. Strbac and C. Kang, "Clustering-Based Residential Baseline Estimation: A Probabilistic Perspective," in IEEE Transactions on Smart Grid. vol. 10, no. 6, pp. 6014-6028, Nov. 2019.
- Q. Yuan, M. Sun*(Corresponding Author), Y. Sheng and Q. Guo, "PrivCPM: Privacy-Preserving Cooperative Pricing Mechanism in Coupled Power-Traffic Networks," in IEEE Transactions on Smart Grid, doi: 10.1109/TSG.2024.3406572.
Scenario Reduction for Large-Scale Energy System Planning
Energy system investment planning problems become intractable due to the vast variability that characterizes system operation and the increasing complexity of the optimization model to capture the characteristics of renewable energy sources (RES). In this context, making optimal investment decisions by considering every operating period is unrealistic and inefficient. Therefore, one of the most effective solutions is to select a limited number of representative scenarios. To this end, three major research questions are pointed out and investigated: 1) Which variables should the clustering be based on? The choice can include combinations of variables in the input domain (e.g. demand and/or renewable injection) and in the decision domain (e.g. lines built or investment cost); 2) Which is the most appropriate clustering technique to be applied for a chosen set of variables (e.g. centroid methods, mixture models, etc.); 3) After clustering the different scenarios, how to select the representative profile of each cluster (e.g. mean value or median point).
- M. Sun, F. Teng, X. Zhang, G. Strbac and D. Pudjianto, "Data-Driven Representative Day Selection for Investment Decisions: A Cost-Oriented Approach," in IEEE Transactions on Power Systems, vol. 34, no. 4, pp. 2925-2936, Jul. 2019.
- M. Sun, F. Teng, I. Konstantelos, G. Strbac, "An objective-based scenario selection method for transmission network expansion planning with multivariate stochasticity in load and renewable energy sources," in Energy, vol. 145, pp. 871-885, Feb. 2018.
- M. Sun, G. Strbac, P. Djapic, D. Pudjianto, "Preheating Quantification for Smart Hybrid Heat Pumps Considering Uncertainty," in IEEE Transactions on Industrial Informatics, vol. 15, no. 8, pp. 4753-4763, Aug. 2019.
教学与课程
工业人工智能安全与隐私
课程号：08617210
Theory and Advanced Technology for Cyber-physical Systems (ZJU)
Introduction to Big Data Analysis and Application (ZJU)
Role and Value of Smart Grid Technologies (ICL)
课程号：EE9-FPN1-01
实验室成员
负责人
孙铭阳博士
孙铭阳，研究员、博雅青年学者、博士生导师, IEEE Senior Member。入选国家高层次青年人才计划，全球2%顶尖科学家，中国科协“青年人才托举工程”、省特聘专家、省“万人计划”青年拔尖人才，兼任英国帝国理工学院Honorary Lecturer。毕业于英国帝国理工学院控制与电力研究所，师从英国能源系统专家Goran Strbac教授。2024年6月加入北京大学工学院、工业工程与管理系, IDEAL-Lab 低碳能源系统智能决策实验室负责人。
多年来一直从事智能电网和能源系统领域优化和智能决策等关键问题的研究，主要研究内容包括基于人工智能的能源系统调度控制、能源大数据分析、建模与预测、能源系统优化规划等国际前沿热点课题。受邀担任电力系统领域国际顶级期刊IEEE Trans. on Power Systems、工业信息领域国际顶级期刊IEEE Trans. on Industrial Informatics、IET Smart Grid编委和Applied Energy、Advances in Applied Energy首席客座编委。研究成果发表/录用Cell姊妹刊Joule、The Innovation、Nature Communications、中国工程院院刊Engineering、能源与电力顶刊IEEE Trans. on Power Systems、IEEE Trans. on Smart Grid、Applied Energy、人工智能旗舰会议NeurIPS （CCF-A）、ICML（CCF-A）、AAAI(CCF-A)、IJCAI (CCF-A)、安全四大顶会USENIX Security (CCF-A)、NDSS (CCF-A)等论文100余篇。ESI高被引论文4篇，入选Applied Energy近五年高被引论文TOP25。获得IEEE Trans. on Smart Grid 2023年度最佳论文奖、IEEE电力与能源协会最高级别学术年会2016 IEEE PES-GM会议最佳论文奖、2016 PMAPS会议最佳论文奖及IEEE TrustCom 2022 最佳论文奖。
任中国自动化学会新能源与储能系统控制专委会委员、青年工作委员会委员、IEEE PES 中国区电力系统运行、规划与经济技术委员会常务理事，2019 Global Automation & Control Workshop程序主席，中国计算机学会CCF YOCSEF杭州副主席，Peking University IEEE Power & Energy Society Student Branch Chapter Advisor。近年来主持项目经费4580余万元，其中纵向经费3400余万元。具体而言，主持国家自然科学基金优秀青年科学基金项目（海外）、重大项目课题、面上项目、青年科学基金（C类）项目、联合重点项目课题、国家自然科学基金委员会与荷兰研究理事会合作研究项目课题等共6项，国家科技重大专项课题1项、国家重点研发计划项目子课题2项、省重点项目1项，CCF科研获奖项目1项等。曾参与承担多个英国、欧盟重大科研项目，其中包括欧盟“地平线2020计划”项目EU-SysFlex(2000万欧元)、泛欧洲电网动态安全风险评估项目iTesla（欧盟第七框架计划FP7支持, 1940万欧元）等，并与英国、法国、澳大利亚、美国等国家的科研机构与知名学者保持联系与密切合作。
科研秘书
岳菲
毕业于北京理工大学，获得国际经济与贸易专业学士学位。现担任北京大学IDEAL Lab科研秘书，负责实验室日常运营及行政工作。
邮箱: yuefei@pku.edu.cn
博士后研究员
张宗衍博士
现任北京大学先进制造与机器人学院助理研究员，博士毕业于华南理工大学计算机科学与工程学院。张宗衍博士在生成式人工智能、大模型与神经网络优化等领域开展了一系列研究工作。目前以核心骨干成员参与国家重点研发计划1项，主持国家科技重大专项子课题1项，参与多项国家自然科学基金项目。在人工智能领域发表多篇顶级期刊论文（IEEE TEVC，IEEE TNNLS，IEEE TCSVT，IEEE TAFFC）与人工智能顶级学术会议论文（ICCV，ICME），尤其在大模型算法领域提出了一系列优化算法法与高效推理框架。
吴佳耕博士
吴佳耕博士，现任北京大学先进制造与机器人学院助理研究员，博士毕业于吉林大学计算数学专业。主要研究方向为通信感知一体化系统的优化建模与算法设计，涉及非凸优化、凸松弛理论、波束赋形优化等。博士期间在 IEEE Transactions on Signal Processing、Journal of the Operations Research Society of China 等国际期刊发表多篇论文，并参与国家重点研发计划。
博士研究生
左可
于2020年获得浙江大学电气工程及其自动化专业学士学位，现于浙江省杭州市浙江大学控制科学与工程学院攻读博士学位。自2023年起，他在帝国理工学院控制与电力研究组担任访问研究员，师从Goran Strbac教授。其主要研究方向包括数据驱动的电力系统稳定性评估与控制。
陈焱
于2020年获浙江大学控制科学与工程学院学士学位，现于该校同一学院攻读博士学位。2023年10月至2024年10月，她在伦敦大学学院电子电气工程系担任访问研究员，导师为Boli Chen教授。其主要研究方向包括人工智能安全及电力系统时序预测中的机器学习技术。
万旭
于2021年获中国地质大学（武汉）自动化专业学士学位。现为浙江大学控制科学与工程学院博士研究生。其主要研究方向包括鲁棒与安全强化学习理论及其在能源电力系统中应用。
沈超
于2022年获华中科技大学水利水电工程学士学位，现为浙江大学控制科学与工程学院博士研究生。研究方向涵盖电力系统动态安全与稳定分析、电力系统建模与调度优化。其近期研究聚焦于物理信息深度学习的建模框架，以及大规模时序模型在电力系统预测与动态建模中的方法论探索，旨在推动机理驱动与数据驱动的融合发展，以提升电力系统运行的安全性与智能化水平。
余泽
于2022年获浙江工商大学信息安全学士学位，现于浙江大学控制科学与工程学院攻读博士学位。其研究方向聚焦于逆变器资源高渗透率下的新型电力系统信息物理安全。
杨铮昊
于2025年在北京大学获得理论与应用力学学士学位。他目前正在北京大学先进制造与机器人学院攻读博士学位。其研究兴趣包括电力系统中的能源感知型强化学习和基于大语言模型的运筹学。
甄红伟
于2024年获得浙江大学电子信息工程学士学位。目前，正在浙江大学电气工程学院攻读博士学位。他的研究方向包括高比例新能源电力系统中的安全攻击与防御，以及融合电力电子特性的网络脆弱性分析方法设计。
张一帆
于2026年获得西安交通大学电气工程及其自动化学士学位，目前于北京大学先进制造与机器人学院攻读博士学位。他的研究方向为基于大语言模型的电力系统自建模理论。
Master Students
方海豹
Haibao obtained his Bachelor's degree in Electrical Engineering and Automation from Soochow University in 2022. Currently, he is pursuing a Master's degree at Zhejiang University and is expected to graduate in 2025. His research focuses on reinforcement learning for low-carbon power management and smart city development.
郭子涵
毕业于苏州大学物理学专业，获理学学士学位。现就读于北京大学先进制造与机器人学院，其主要研究方向是大语言模型解决电网优化问题建模和求解。
敖维蕗
于2024年获得哈尔滨工业大学（深圳）计算机科学与技术学院学士学位，现于北京大学工学院攻读专业硕士学位，其主要研究方向聚焦人工智能系统安全风险治理。
Ximing Huang
Ximing obtained his Bachelor's degree in Mathematics and Applied Mathematics from China University of Mining and Technology-Beijing. Currently, he is pursuing a Master's degree at Peking University. His research focuses on robust optimization and LLM for power systems.
芦晓涵
于2023年获得中南大学工程管理学士学位，他目前是北京大学力学与工程科学学院的一名硕士研究生，研究兴趣包括大语言模型和智能体在电力系统中的应用。
Graduated Students
Haochi Wu @ Stanford University
Haochi is currently a Postdoctoral Researcher in the Department of Civil and Environmental Engineering at Stanford University. Before this, he was a doctoral researcher and research associate in the ASSET Lab and Center for Sustainable Systems at the University of Michigan, working with Prof. Michael Craig. He completed my PhD research in energy systems at Zhejiang University, under the supervision of Prof. Mingyang Sun at Peking University. His research interests include macro energy systems, efficient data-driven energy system operation, water-energy nexus, and emerging energy conversion and management technologies
Lanting Zeng @ Alibaba
Lanting received her Bachelor of electrical engineering and automation from Xiamen University in 2020. She obtained her Ph.D. degree at the School of Control Science and Engineering, Zhejiang University, in 2025. Since 2023, she has been a visiting researcher at IEPG group at the Delft University of Technology. Her research focuses on robust reinforcement learning-based power systems operation and control against model uncertainties and cyber attacks.
Tingqi Zhang @ Beijing Electric Power Trading Center Co., Ltd
Dr Tingqi Zhang received the Ph.D. degree from the Department of Electrical and Electronic Engineering, Imperial College London, London, U.K., in 2022. He was a Joint Postdoctoral Fellow with the Electric Power Research Institute, State Grid Liaoning Electric Power Company Ltd., and the Department of Control Science and Engineering, Zhejiang University. His research interests include big data techniques in power systems operation and control.
Juxing Xue @ Huawei Digital Power
Juxing obtained his Master's degree from the College of Control Science and Engineering, Zhejiang University in 2023 and joined Huawei Digital Power Technology. His research focuses on reinforcement learning-based energy management for multi-energy systems.
美好瞬间
