基于微分神经网络的风电机群低电压穿越特性建模

doi:10.16182/j.issn1004731x.joss.22-FZ0928

系统仿真学报 ›› 2022, Vol. 34 ›› Issue (12): 2546-2556.doi: 10.16182/j.issn1004731x.joss.22-FZ0928

基于微分神经网络的风电机群低电压穿越特性建模

赖启平(), 肖谭南, 李东晟, 沈沉()

清华大学电机工程与应用电子技术系，北京 100084

收稿日期:2022-08-07 修回日期:2022-09-21 出版日期:2022-12-31 发布日期:2022-12-21
通讯作者: 沈沉 E-mail:lqp22@mails.tsinghua.edu.cn;shenchen@mail.tsinghua.edu.cn
作者简介:赖启平(2000-)，男，博士生，研究方向为风电场等值建模与控制。E-mail：lqp22@mails.tsinghua.edu.cn
基金资助:
国家自然科学基金集成项目(U2166601);国家自然科学基金(52107104)

Low Voltage Ride-through Modeling for Wind Turbines Based on Neural ODEs

Qiping Lai(), Tannan Xiao, Dongsheng Li, Chen Shen()

Department of Electrical Engineering, Tsinghua University, Beijing 100084, China

Received:2022-08-07 Revised:2022-09-21 Online:2022-12-31 Published:2022-12-21
Contact: Chen Shen E-mail:lqp22@mails.tsinghua.edu.cn;shenchen@mail.tsinghua.edu.cn

摘要/Abstract

摘要：

针对目前缺少分析风电机群低电压穿越特性高效建模方法的问题，提出了一种基于微分神经网络的风电机群低电压穿越特性建模 方法。模型输入为风电场并网点电压与各台风机风速，输出为并网点电流。该模型能够较好地表征其强非线性切换过程，刻画风电场不同风速分布场景下风电机群的低电压穿越特性。在CloudPSS云仿真平台上建立了包含3台双馈风力发电机的风电机群仿真算例，对所提方法进行测试，测试结果验证了微分神经网络模型的泛化能力与有效性。

关键词: 双馈风力发电机, CloudPSS建模仿真, 低电压穿越特性, 微分神经网络, 数据驱动建模

Abstract:

Considering the difficulty of equivalent modeling of low voltage ride-through(LVRT) characteristics of a wind farm, a neural ordinary differential equation(ODE)-based wind farm LVRT modeling methodis proposed. The input of the model is the voltage and wind speed of each wind turbine at the grid connection point of wind farm, and the output is the current at the grid connection point. The model can better characterize the strong nonlinear switching process and describe LVRT characteristics of wind farms under different wind speed scenarios. A simulation example of a wind farm including three doubly-fed induction generators(DFIGs) is established on CloudPSS simulation platform to test the proposed method. The test results verify the generalization ability and effectiveness of Neural ODEs model.

Key words: DFIGs(doubly-fed induction generators), CloudPSS simulation, LVRT(low voltage ride-through) characteristics, Neural ODEs(ordinary differential equation), data-driven modeling

中图分类号:

TP391.9

赖启平, 肖谭南, 李东晟, 沈沉. 基于微分神经网络的风电机群低电压穿越特性建模[J]. 系统仿真学报, 2022, 34(12): 2546-2556.

Qiping Lai, Tannan Xiao, Dongsheng Li, Chen Shen. Low Voltage Ride-through Modeling for Wind Turbines Based on Neural ODEs[J]. Journal of System Simulation, 2022, 34(12): 2546-2556.

图/表 14

图1

图2

图3

图4

图5

表1

表2

表3

图6

图7

图8

图9

表4

表5

参考文献 17

1	Masa A, Kkm B, Sb A, et al. Modified Rotor-Side Converter Control Design for Improving the LVRT Capability of A DFIG-Based WECS-Science Direct[J]. Electric Power Systems Research(S0378-7796), 2020, 186: 106403.
2	葛江北, 周明, 李庚银. 大型风电场建模综述[J]. 电力系统保护与控制, 2013, 41(17): 146-153.
	Ge Jiangbei, Zhou Ming, Li Gengyin. Review on Large-Scale Wind Farm Modeling[J]. Power System Protection and Control, 2013, 41(17): 146-153.
3	崔峰. 双馈风力发电机及风电场建模仿真研究[D]. 北京: 华北电力大学, 2015.
	Cui Feng. Research on Modeling and Simulation of Double-fed Induction Generator and Wind Farm[D]. Beijing: North China Electric Power University, 2015.
4	晁璞璞. 双馈型风电场的机电暂态等值建模方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2018.
	Chao Pupu. Research on Electromechanical Transient Equivalent Modeling of DFIG-Based Wind Farms[D]. Harbin: Harbin Institute of Technology, 2018.
5	胡家兵, 孙丹, 贺益康, 等. 电网电压骤降故障下双馈风力发电机建模与控制[J]. 电力系统自动化, 2006, 30(8): 21-26.
	Hu Jiabing, Sun Dan, He Yikang, et al. Modeling and Control of DFIG Wind Energy Generation System under Grid Voltage Dip[J]. Automation of Electric Power Systems, 2006, 30(8): 21-26.
6	Chen R T Q, Rubanova Y, Bettencourt J, et al. Neural Ordinary Differential Equations[J]. Advances in Neural Information Processing Systems(S1049-5258), 2018, 31: 6571-6583.
7	Zang C, Wang F. Neural Dynamics on Complex Networks[C]//KDD'20: The 26th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York, USA: ACM, 2020: 892-902.
8	沈沉, 陈颖, 黄少伟, 等. 新型电力系统仿真应用软件设计理念与发展路径[J]. 电力系统自动化, 2022, 46(10): 75-86.
	Shen Chen, Chen Ying, Huang Shaowei, et al. Design Idea and Development Path of Simulation Application Software for New Power System[J]. Automation of Electric Power Systems, 2022, 46(10): 75-86.
9	陈颖, 高仕林, 宋炎侃, 等. 面向新型电力系统的高性能电磁暂态云仿真技术[J]. 中国电机工程学报, 2022, 42(8): 2854-2864.
	Chen Ying, Gao Shilin, Song Yankan, et al. High-performance Electromagnetic Transient Simulation for New-type Power System Based on Cloud Computing[J]. Proceedings of the CSEE, 2022, 42(8): 2854-2864.
10	董毅峰, 王彦良, 韩佶, 等. 电力系统高效电磁暂态仿真技术综述[J]. 中国电机工程学报, 2018, 38(8): 2213-2231, 2532.
	Dong Yifeng, Wang Yanliang, Han Ji, et al. Review of High Efficiency Digital Electromagnetic Transient Simulation Technology in Power System[J]. Proceedings of the CSEE, 2018, 38(8): 2213-2231, 2532.
11	宋炎侃, 陈颖, 黄少伟, 等. 大规模电力系统电磁暂态并行仿真算法和实现[J]. 电力建设, 2015, 36(12): 9-15.
	Song Yankan, Chen Ying, Huang Shaowei, et al. Electromagnetic Transient Parallel Simulation Algorithm and Implementation for Large-scale Power System[J]. Electric Power Construction, 2015, 36(12): 9-15.
12	中国国家标准化管理委员会. 风力发电机组故障电压穿越能力测试规程: [S]. 北京: 中国标准出版社, 2018: 1-20.
	Standardization Administration of the People's Republic of China. Wind Turbines-Test Procedure of voltage Fault Ride Through Capability: [S]. Beijing: Standards Press of China, 2018: 1-20.
13	孙大卫, 吴林林, 刘辉, 等. 弱电网直驱风机低电压穿越特性及其对机端暂态电压的影响[J]. 中国电机工程学报, 2021, 41(14): 4777-4786.
	Sun Dawei, Wu Linlin, Liu Hui, et al. Effect of the Low Voltage Ride Through Characteristics on PMSG Terminal Transient Voltage in Weakly-synchronized Gird[J]. Proceedings of the CSEE, 2021, 41(14): 4777-4786.
14	Kidger P. On Neural Differential Equations[D]. Oxford: University of Oxford, 2021.
15	Xiao T, Chen Y, Huang S, et al. Feasibility Study of Neural ODE and DAE Modules for Power System Dynamic Component Modeling[J/OL]. IEEE Transactions on Power Systems(S0885-8950) [2022-08-03]. DOI: 10.1109/TPWRS.2022.3194570 . .
16	Zhou Y, Zhang P. Neuro-Reachability of Networked Microgrids[J]. IEEE Transactions on Power Systems (S0885-8950), 2022, 37(1): 142-152.
17	Pontryagin L S. The Mathematical Theory of Optimal Processes[M]. New York, USA: John Wiley & Sons, Inc, 1962.

参数名称	参数选取	参数名称	参数选取
训练集样本数	200	批大小	4
学习率	0.05	训练轮数	400
网络层数	4	优化器	Adam
隐藏层神经元数	64	激活函数	ELU

风速工况		样本编号	样本数量
对称	000	0 ~ 5	6
	111	6 ~ 13	8
	222	14 ~ 19	6
不对称	000	20 ~ 27	8
	001	28 ~ 35	8
	002	36 ~ 43	8
	011	44 ~ 51	8
	012	52 ~ 59	8
	022	60 ~ 67	8
	111	68 ~ 75	8
	112	76 ~ 83	8
	122	84 ~ 91	8
	222	92 ~ 99	8

风速工况		样本编号	样本数量
对称	000	100 ~ 112	13
	111	113 ~ 126	14
	222	127 ~ 139	13
不对称	000	140 ~ 155	16
	001	156 ~ 171	16
	002	172 ~ 187	16
	011	188 ~ 203	16
	012	204 ~ 219	16
	022	220 ~ 235	16
	111	236 ~ 251	16
	112	252 ~ 267	16
	122	268 ~ 283	16
	222	284 ~ 299	16

测试工况	最终误差
示例工况1	0.000 265 704
示例工况2	0.000 185 900
示例工况3	0.000 405 468
示例工况4	0.000 229 250

循环数	平均误差	方差
100	0.000 265	0.000 162
200	0.000 234	0.000 211
300	0.000 139	0.000 229
400	0.000 115	0.000 218

基于微分神经网络的风电机群低电压穿越特性建模

Low Voltage Ride-through Modeling for Wind Turbines Based on Neural ODEs

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 17

相关文章 3

编辑推荐

Metrics

本文评价

[1]	张洪阳, 时振堂, 张志锋. 电网故障下双馈风力发电机控制改进与仿真[J]. 系统仿真学报, 2020, 32(2): 299-305.
[2]	李冰, 刘石, 龙腾, 段琦玮. 无刷双馈发电机在电网故障时的动态响应分析[J]. 系统仿真学报, 2018, 30(2): 605-613.
[3]	蒋莹莹, 纪志成. 基于混合量子粒子群算法的DFIG参数辨识[J]. 系统仿真学报, 2016, 28(5): 1054-1062.