基于单机扫描与曲线形状聚类的风电场分群等值方法

doi:10.19595/j.cnki.1000-6753.tces.241759

Abstract
Figure/Table
References
Related Citation (13)

Download: PDF (5632 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract The simplified model of grouping and equivalent of wind turbines, inspired by the coherent equivalent method in power systems, has been widely applied. For wind turbines, the transient responses after faults under different operating conditions exhibit significant differences. To achieve accurate grouping, it should be based on the similarity of transient responses. However, most existing wind turbine grouping methods rely on steady-state conditions represented by wind speed, state variables, etc., which struggle to meet simulation demands in terms of grouping and equivalent accuracy. While some methods directly cluster transient response curves, they require access to unit-level transient response data, limiting their practicality. To address the challenge of balancing effectiveness and practicality in wind turbine grouping and equivalent, this paper proposes a method based on single-unit scanning and curve shape clustering. Essentially focusing on transient responses, this method only requires steady-state data and fault scenarios for application, ensuring both effectiveness and practicality.
Firstly, by setting different wind speeds, voltages, and fault conditions for individual wind turbines, key influencing factors and dynamic responses are selected through comparative analysis. This single-unit scanning simulation method enables the acquisition of clustering data while establishing a link between steady-state points and dynamic characteristics. Secondly, considering the actual process of aggregation and equivalent, the correlation coefficient is chosen as the similarity evaluation index between transient responses. Thirdly, by comparing various classical clustering methods, the K-Means clustering algorithm, which is most suitable for this problem, is selected and further optimized using a crested porcupine optimizer optimization algorithm to improve clustering effectiveness. Finally, a time series clustering algorithm capable of effectively capturing shape differences is implemented to precisely categorize transient responses, combining the link between steady-state points and dynamic responses established in the first step to achieve practical and high-precision wind turbine grouping.
Applying the proposed grouping method to a typical direct-drive wind turbine model, the grouping results under full fault conditions are in complete agreement with the distinct grouping boundaries derived from theoretical analysis. In contrast, traditional methods based on steady-state quantities yield significantly different grouping boundaries, demonstrating that the proposed grouping index enables precise grouping based on transient responses. Regarding clustering algorithms, the K-Means algorithm outperforms six other algorithms across three evaluation metrics. By repeatedly clustering and calculating the evaluation metrics, the proposed improved method shows slight improvements in the mean values of the three metrics compared to the original method, with the variance reduced to within 1% of the original method, indicating the effectiveness of the modified K-Means algorithm. Finally, PSCAD simulation cases for three types of faults verify the equivalence. Compared to traditional wind speed clustering methods, the proposed method reduces the average relative error by approximately 60% to 90% and the maximum absolute error by approximately 50% to 90%, indicating its superior performance in grouping and equivalent effectiveness.
In conclusion, this paper finds that: (1) The primary influencing factors of the low-voltage ride-through dynamic characteristics of direct-drive wind turbines are wind speed and fault conditions, with active power being the primary manifestation of dynamic characteristic differences, while reactive power differences are minimal. (2) Measuring time series similarity using the linear correlation coefficient effectively captures the linear relationships between curves, making it suitable for wind turbine grouping and equivalent problems. The crested porcupine optimizer algorithm improved K-Means algorithm achieves optimal clustering results. (3) By acquiring transient time series through single-unit scanning and using them as clustering data, clustering dynamic characteristics to obtain steady-state point grouping boundaries, and then performing wind farm grouping and equivalent, this method significantly improves equivalent accuracy while ensuring practicality.

Key words： Dynamic equivalence direct-drive wind turbine generator wind farm equivalence time series clustering

Received: 08 October 2024

PACS:

TM71

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors

Cite this article:

URL:

https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.241759 OR https://dgjsxb.ces-transaction.com/EN/Y2025/V40/I19/6072

[1] 黄萌, 舒思睿, 李锡林, 等. 面向同步稳定性的电力电子并网变流器分析与控制研究综述[J]. 电工技术学报, 2024, 39(19): 5978-5994.
Huang Meng, Shu Sirui, Li Xilin, et al.A review of synchronization-stability-oriented analysis and control of power electronic grid-connected converters[J]. Trans-actions of China Electrotechnical Society, 2024, 39(19): 5978-5994.
[2] 程昊, 吕亚洲, 李威, 等. 大扰动下计及风电场低电压穿越恢复差异的频率响应估计方法[J]. 电力系统自动化, 2024, 48(16): 68-78.
Cheng Hao, Lyu Yazhou, Li Wei, et al.Frequency response estimation method considering low voltage ride-through recovery difference of wind farms under large disturbance[J]. Automation of Electric Power Systems, 2024, 48(16): 68-78.
[3] 潘学萍, 王卫康, 陈海东, 等. 计及低电压穿越影响的感应电动机动态分群[J]. 电工技术学报, 2024, 39(7): 2001-2016, 2032.
Pan Xueping, Wang Weikang, Chen Haidong, et al.Study of dynamic clustering method for induction motors considering low voltage ride through effects[J]. Transactions of China Electrotechnical Society, 2024, 39(7): 2001-2016, 2032.
[4] Wu Feng, Qian Junxia, Ju Ping, et al.Transfer function based equivalent modeling method for wind farm[J]. Journal of Modern Power Systems and Clean Energy, 2019, 7(3): 549-557.
[5] 李龙源, 付瑞清, 吕晓琴, 等. 接入弱电网的同型机直驱风电场单机等值建模[J]. 电工技术学报, 2023, 38(3): 712-725.
Li Longyuan, Fu Ruiqing, Lü Xiaoqin, et al.Single machine equivalent modeling of weak grid connected wind farm with same type PMSGs[J]. Transactions of ChinaElectrotechnical Society, 2023, 38(3): 712-725.
[6] Zhu Lin, Zhang Jian, Zhong Danting, et al.A study of dynamic equivalence using the similarity degree of the equivalent power angle in doubly fed induction generator wind farms[J]. IEEE Access, 2020, 8: 88584-88593.
[7] 徐玉琴, 王娜. 基于聚类分析的双馈机组风电场动态等值模型的研究[J]. 华北电力大学学报, 2013, 40(3): 1-5.
Xu Yuqin, Wang Na.Study on dynamic equivalence of wind farms with DFIG based on clustering analysis[J]. Journal of North China Electric Power University, 2013, 40(3): 1-5.
[8] 林俐, 陈迎. 基于扩散映射理论的谱聚类算法的风电场机群划分[J]. 电力自动化设备, 2013, 33(6): 113-118.
Lin Li, Chen Ying.Wind turbine grouping with spectral clustering algorithm based on diffusion mapping theory[J]. Electric Power Automation Equipment, 2013, 33(6): 113-118.
[9] 张星, 李龙源, 胡晓波, 等. 基于风电机组输出时间序列数据分群的风电场动态等值[J]. 电网技术, 2015, 39(10): 2787-2793.
Zhang Xing, Li Longyuan, Hu Xiaobo, et al.Wind farm dynamic equivalence based on clustering by output time series data of wind turbine generators[J]. Power System Technology, 2015, 39(10): 2787-2793.
[10] 欧阳金鑫, 刁艳波, 郑迪, 等. 基于电流轨迹相似度的双馈风电机群电磁暂态同调分群方法[J]. 中国电机工程学报, 2017, 37(10): 2896-2904.
Ouyang Jinxin, Diao Yanbo, Zheng Di, et al.A clustering method of coherent generators during electro-magnetic transient process based on similar degrees of current trajectories for doubly fed wind farms[J]. Proceedings of the CSEE, 2017, 37(10): 2896-2904.
[11] 米增强, 苏勋文, 余洋, 等. 双馈机组风电场动态等效模型研究[J]. 电力系统自动化, 2010, 34(17): 72-77.
Mi Zengqiang, Su Xunwen, Yu Yang, et al.Study on dynamic equivalence model of wind farms with wind turbine driven doubly fed induction generators[J]. Automation of Electric Power Systems, 2010, 34(17): 72-77.
[12] 付盼, 胡庆林. 基于聚类分析的风电场多机等效建模方法研究[J]. 高压电器, 2019, 55(04): 198-204.
Fu Pan, Hu Qinglin.Research on multi-machine equivalent modeling method of wind farm based on cluster analysis[J]. High Voltage Apparatus, 2019, 55(04): 198-204.
[13] 陈树勇, 王聪, 申洪, 等. 基于聚类算法的风电场动态等值[J]. 中国电机工程学报, 2012, 32(4): 11-19.
Chen Shuyong, Wang Cong, Shen Hong, et al.dynamic equivalence for wind farms based on clustering algori-thm[J]. Proceedings of the CSEE, 2012, 32(4): 11-19.
[14] 颜湘武, 李君岩. 基于主成分分析法的直驱式风电场分群方法[J]. 电力系统保护与控制, 2020, 48(5): 127-133.
Yan Xiangwu, Li Junyan.Grouping method of direct drive wind farm based on principal component analysis[J]. Power System Protection and Control, 2020,48(5): 127-133.
[15] Jiang Wenbo, Zhong Mingyue.Dynamic equivalent modeling of wind farm based on dominant variable hierarchical clustering algorithm[J]. Mathematical Problems in Engineering, 2021, 2021: 7629414.
[16] Wang Xiaohui, Yu Hao, Lin Yong, et al.Dynamic equivalent modeling for wind farms with DFIGs using the artificial bee colony with K-means algorithm[J]. IEEE Access, 2020, 8: 173723-173731.
[17] 孙元存, 刘三明, 王致杰, 等. 基于IGWO-K-means的风电场动态等值建模[J]. 现代电力, 2018, 35(5): 49-55.
Sun Yuancun, Liu Sanming, Wang Zhijie, et al.dynamic equivalent modeling of wind farm based on IGWO-K-means method[J]. Modern Electric Power, 2018, 35(5): 49-55.
[18] 王增平, 杨国生, 汤涌, 等. 基于特征影响因子和改进BP算法的直驱风机风电场建模方法[J]. 中国电机工程学报, 2019, 39(9): 2604-2615.
Wang Zengping, Yang Guosheng, Tang Yong, et al.modeling method of direct-driven wind generators wind farm based on feature influence factors and improved BP algorithm[J]. Proceedings of the CSEE, 2019, 39(9): 2604-2615.
[19] 吴志鹏, 曹铭凯, 李银红. 计及Crowbar状态改进识别的双馈风电场等值建模方法[J]. 中国电机工程学报, 2022, 42(2): 603-614.
Wu Zhipeng, Cao Mingkai, Li Yinhong.An equivalent modeling method of DFIG-based wind farm considering improved identification of crowbar status[J]. Proceedings of the CSEE, 2022, 42(2): 603-614.
[20] 王磊, 盖春阳, 王恒. 基于改进D-K聚类算法的直驱型风电场动态等值建模[J]. 太阳能学报, 2021, 42(3): 48-55.
Wang Lei, Gai Chunyang, Wang Heng.Dynamic equivalence method of PMSG wind farms based on improved D-K clustering algorithm[J]. Acta Energiae Solaris Sinica, 2021, 42(3): 48-55.
[21] 张扬帆, 李奕霖, 叶林, 等. 低温天气下考虑风机运行状态聚类的短期风电功率预测方法[J]. 发电技术, 2025, 46(2): 326-335.
Zhang Yangfan, Li Yilin, Ye Lin, et al.Short-term wind power prediction method considering wind turbine operation status clustering under low-temperature conditions[J]. Power Generation Technology, 2025, 46(2): 326-335.
[22] 邓俊, 张阳, 李怡然, 等. 基于高斯混合模型聚类的双馈风电场动态等值建模方法[J]. 太阳能学报, 2024, 45(1): 342-350.
Deng Jun, Zhang Yang, Li Yiran, et al.Dynamic equivalence modeling of doubly-fed wind farm based on gaussian mixture model clustering algorithm[J]. Acta Energiae Solaris Sinica, 2024, 45(1): 342-350.
[23] Li Dongsheng, Shen Chen, Liu Ye, et al.A dynamic equivalent method for PMSG-WTG based wind farms considering wind speeds and fault severities[J]. IEEE Transactions on Power Systems, 2024, 39(2): 3738-3751.
[24] 王耀函, 张扬帆, 赵庆旭, 等. 低电压穿越过程中风电机组载荷特性联合仿真研究[J]. 发电技术, 2024, 45(4): 705-715.
Wang Yaohan, Zhang Yangfan, Zhao Qingxu, et al.Joint simulation study on load characteristics of wind turbines in low voltage ride through process[J]. Power Generation Technology, 2024, 45(4): 705-715.
[25] Rousseeuw P J.Silhouettes: a graphical aid to the interpretation and validation of cluster analysis[J]. Journal of Computational and Applied Mathematics, 1987, 20: 53-65.
[26] Calinski T, Harabasz J.A dendrite method for cluster analysis[J]. Communications in Statistics-Theory and Methods, 1974, 3(1): 780133830.
[27] David L. Davies, Donald W.Bouldin. A cluster separation measure. IEEE Transactions on Pattern Analysis and Machine Intelligence[J], 1(2): 224-227.
[28] Abdel-Basset M, Mohamed R, Abouhawwash M.Crested Porcupine Optimizer: a new nature-inspired metaheuristic[J]. Knowledge-Based Systems, 2024, 284: 111257.
[29] 李东晟, 沈沉, 吴林林, 等. 考虑初始风速与机端故障稳态电压跌落程度的直驱风机故障响应特性分类及其判别方法研究[J]. 中国电机工程学报, 2024, 44(4): 1247-1260.
Li Dongsheng, Shen Chen, Wu Linlin, et al.Study on fault response characteristics classification and discrimi-nant method of PMSG considering initial wind speed and drop degree of terminal fault steady-state voltage[J]. Proceedings of the CSEE, 2024, 44(4): 1247-1260.