摘要
电能路由器在未来能源互联网、泛在电力物联网等领域具有广泛的应用前景,具有多级联、多端口、多流向和多形态的特征,这些特征给传统人工设计带来诸多难题。电力电子设计自动化有望克服设计目标权衡困难、裕量设计浪费成本、硬件迭代耗时耗力、设计自动化程度低等挑战。在宽禁带器件蓬勃发展、计算机性能快速迭代、变换器应用场景延伸细化的三重推进下,电力电子设计自动化有望实现设计集成化、精细化和快速化。电能路由器设计问题可以抽象为多学科优化问题,该问题数学性质差、问题规模大,传统优化算法难以胜任,而遗传算法在解决此类优化问题上性能优异。该文在系统总结当前主流的电力电子系统的设计流程和设计软件架构的基础上,分析了遗传算法解决电力电子设计自动化问题的适用性,总结了电能路由器设计自动化潜在的问题和挑战,并给出相应的建议。
Electrical energy routers are promising in energy Internet,ubiquitous electric Internet of Things and other scenarios,and have the features of multi-cascade,multi-port,multi-flow,and multi-modal.These features bring many difficulties to traditional manual design.Design automation for power electronics is expected to overcome the challenges such as difficulty in weighing design objectives,wasting costs on marginal design,time-consuming and labor in hardware iterations,and low degree of design automation.With the rapid development of wide bandgap devices,rapid iteration of computer performance,and extension and refinement of converter application scenarios,design automation for power electronics is expected to achieve design integration,refinement and rapidity.The electrical energy router design problem can be abstracted as a multidisciplinary optimization problem with poor mathematical nature and large scale.Regarding such optimization problems,genetic algorithms have better performance than the traditional optimization algorithms.Based on a systematic summary of the current mainstream design workflow and design software architecture,this paper analyzes the applicability of genetic algorithms to solve design automation problems of power electronics,summarizes the potential problems and challenges in design automation for electrical energy routers,and gives corresponding suggestions.
作者
袁立强
陆子贤
孙建宁
段任之
赵争鸣
Yuan Liqiang;Lu Zixian;Sun Jianning;Duan Renzhi;Zhao Zhengming(State Key Laboratory of Control and Simulation of Power System and Generation Equipment Department of Electrical Engineering Tsinghua University,Beijing 100084 China)
出处
《电工技术学报》
EI
CSCD
北大核心
2020年第18期3878-3893,共16页
Transactions of China Electrotechnical Society
基金
国家自然科学基金资助项目(51577100,51490683)。
关键词
电力电子设计自动化
电能路由器
设计流程
设计软件
优化
遗传算法
Design automation for power electronics
electrical energy router
design workflow
design software
optimization
genetic algorithm
