基于平均化理论的PWM变流器电磁暂态快速仿真方法 (三)适用于图像处理器的改进EMTP并行仿真算法
Q1 Engineering
引用次数: 8
Abstract
智能电网技术的发展需要快速电磁暂态程序（EMTP），而日益广泛应用的图像处理器（GPU）为电磁暂态仿真提供了高效的仿真环境和平台。文中首先提出了细粒度并行算法的运算级并行策略，即基于单指令多数据流（SIMD）的运算级并行策略和基于共享内存的运算级并行策略。随后，设计了应用这两种并行策略的改进电磁暂态细粒度并行算法。三相脉宽调制（PWM）变流器仿真测试表明，适用于GPU的细粒度并行算法能够在保证仿真正确性的同时，显著提高仿真效率，从而验证了基于GPU的细粒度并行仿真算法适用于带有开关过程和复杂控制的大规模电力系统快速电磁暂态仿真应用的可行性。基于平均化理论的PWM变流器电磁暂态快速仿真方法 (三)适用于图像处理器的改进EMTP并行仿真算法
智能电网技术的发展需要快速电磁暂态程序（EMTP），而日益广泛应用的图像处理器（GPU）为电磁暂态仿真提供了高效的仿真环境和平台。文中首先提出了细粒度并行算法的运算级并行策略，即基于单指令多数据流（SIMD）的运算级并行策略和基于共享内存的运算级并行策略。随后，设计了应用这两种并行策略的改进电磁暂态细粒度并行算法。三相脉宽调制（PWM）变流器仿真测试表明，适用于GPU的细粒度并行算法能够在保证仿真正确性的同时，显著提高仿真效率，从而验证了基于GPU的细粒度并行仿真算法适用于带有开关过程和复杂控制的大规模电力系统快速电磁暂态仿真应用的可行性。
本文章由计算机程序翻译，如有差异，请以英文原文为准。
求助全文
约1分钟内获得全文
求助全文
来源期刊
电力系统自动化
Energy-Energy Engineering and Power Technology
CiteScore
8.20
自引率
0.00%
发文量
15032
期刊介绍：
Founded in 1977, Power System Automation is a well-known journal in the discipline of electrical engineering in China. At present, it has been issued to all provinces, cities, autonomous regions, Hong Kong, Macao and Taiwan, and abroad to dozens of countries in North America, Europe and Asia-Pacific region, with a large number of readers at home and abroad. Power System Automation takes “based on China, facing the world, seeking truth and innovation, promoting scientific and technological progress in the field of electric power and energy” as the purpose of the journal, mainly for the professional and technical personnel, teachers and students engaged in scientific research, design, operation, testing, manufacturing, management and marketing in the electric power industry and higher education institutions as well as electric power users, and focuses on hotspots of the industry's development and the It focuses on the hot and difficult issues of the industry. It focuses on the hot and difficult issues of the industry, both academic and forward-looking, practical and oriented, and at the same time emphasizes and encourages technical exchanges of experiences, improvements and innovations from the front line of scientific research and production.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
