混合量子-经典算法的配电网灾后拓扑重构方法  

Post-Disaster Topological Reconfiguration Approach for Distribution Networks Using Hybrid Quantum-Classical Algorithm

作　　者：付炜 谢海鹏 王鹤峰 陈晨[1] 别朝红[1] FU Wei;XIE Haipeng;WANG Hefeng;CHEN Chen;BIE Zhaohong(School of Electrical Engineering,Xi’an Jiaotong University,Xi’an 710049,China;School of Physics,Xi’an Jiaotong University,Xi’an 710049,China)

机构地区：[1]西安交通大学电气工程学院,西安710049 [2]西安交通大学物理学院,西安710049

出　　处：《西安交通大学学报》2025年第1期1-16,共16页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(52307134);中央高校基本科研业务费专项资金资助项目(xzy022024069)。

摘　　要：为及时有效地制定配电网拓扑重构策略以提升负荷快速恢复能力,基于量子计算的优越性,提出混合量子-经典(HQC)算法的弹性配电网灾后拓扑重构方法。首先,构建基于HQC算法的灾后配电网拓扑重构模型,以实现实际场景、优化问题、嵌入算法相应模块在量子计算和经典计算环境下的协作交互过程。然后,将配电网拓扑重构问题构造为无约束离散优化子问题和有约束连续优化子问题,提出量子退火嵌入式交替方向乘子(QA-ADMM)算法,将离散子问题等效映射成量子可解释的伊辛模型后,部署在D-Wave量子退火计算机中,并与经典计算机中连续子问题迭代求解,采用自适应惩罚因子调节机制加速算法收敛。最后,通过IEEE 14、33、69、123以及改进的205节点的不同规模配电系统,分析验证了QA-ADMM算法的有效性、稳定性与可扩展性。结果表明,惩罚因子、目标函数惩罚项系数、量子退火中采样读取次数会影响所提算法的精度和收敛速度;优化问题规模扩大后,所提混合量子-经典算法计算优势更加明显,205节点配电系统算例下,计算效率较经典计算可提升约34%。To promptly and effectively devise topological reconfiguration strategies for distribution networks to boost rapid load recovery capabilities,an elastic post-disaster topological reconfiguration method for distribution networks using a hybrid quantum-classical(HQC)algorithm is introduced,with a specific focus on the advantages of quantum computing.Firstly,a post-disaster topology reconfiguration model for distribution network based on HQC algorithm is established to facilitate interactive processes among real-world scenarios,optimization problems,and embedded algorithm modules in both quantum and classical computing environments.Then,the topological reconfiguration problem for distribution networks is structured into discrete unconstrained optimization sub-problems and continuous constrained optimization sub-problems.A quantum annealing-embedded alternating direction method of multipliers(QA-ADMM)algorithm is proposed,which maps discrete sub-problems into quantum-interpretable Ising models.This algorithm is implemented on the D-Wave quantum annealing computer and iteratively solved on classical computers for continuous sub-problems.An adaptive penalty factor adjustment mechanism is utilized to hasten algorithm convergence.Through analyses of various distribution systems,including IEEE 14,33,69,123 and an enhanced 205-node distribution system,the effectiveness,stability,and scalability of the QA-ADMM algorithm are validated.The findings suggest that penalty factors,penalty term coefficients,and quantum annealing sampling read times influence the accuracy and convergence speed of the algorithm.The computational benefits of the hybrid quantum-classical algorithm become more pronounced with larger optimization problem scales.In the case of a 205-node distribution system,computational efficiency using the hybrid approach can be boosted by around 34%compared to classical computing.

关 键 词：弹性配电网 拓扑重构 混合量子-经典算法 量子计算 量子退火 

分 类 号：TM72[电气工程—电力系统及自动化]