基于稀疏增强动态模态分解的风电场降阶模型
Reduced-order Wake Model of Wind Farm Based on Sparsity Promoting Dynamic Mode Decomposition
魏赏赏 1顾明轩 1李智寒 1许昌 1赵振宙1
作者信息
- 1. 河海大学新能源学院,南京 210001
- 折叠
摘要
针对基于标准动态模态分解所得风电场尾流场维数高、模态排序不唯一等问题,本文采用稀疏增强动态模态分解方法(SPDMD)对风电场尾流进行降阶建模.通过稀疏增强处理优化尾流场不同模态振幅从而提炼风电场尾流中主导模态,并基于主导模态构建数据驱动范式下的风电场尾流模型.通过WFSim中保真风电场以及SOWFA的高保真风电场进行案例分析,结果表明稀疏增强动态模态分解方法较标准动态模态分解,所得模态时空分布更为集中,且可以利用更少的模态数实现风电场尾流高精度重构,证明了稀疏增强动态模态分解在风电场尾流降阶模型构建中的优越性.
Abstract
To solve the problems of wind farm wake with high dimensionality and non-unique mode ordering based on standard dynamic mode decomposition,sparse enhanced dynamic mode decomposition(SPDMD)is used to model wind farm wake.By optimizing the amplitude of different modes of wake through sparse enhancement processing,the dominant modes in wind farm wake are extracted,and the wind farm wake model under data-driven paradigm is constructed based on the dominant modes.Through the case analysis of the medium-fidelity wind farm in WFSim and the high-fidelity wind farm in SOWFA,the results show that compared with the standard dynamic mode decomposition,the spatiotemporal distribution of the modes obtained by the SPDMD method is more concentrated,and the wind farm wake can be reconstructed with fewer modes.The advantages of sparse enhanced dynamic mode decomposition in the construction of the reduced model of wind farm wake are proved.
关键词
风电场尾流/稀疏增强动态模态分解/尾流模型/降阶模型Key words
wind farm wake/sparsity promoting dynamic mode decomposition/wake model/reduced-order model引用本文复制引用
出版年
2024
NETL
NSTL
万方数据