基于复杂特征提取和Sinkhorn距离的风光荷多阶段场景树生成方法

doi:10.11930/j.issn.1004-9649.202402033

中国电力 ›› 2024, Vol. 57 ›› Issue (12): 30-40.DOI: 10.11930/j.issn.1004-9649.202402033

• 面向新型电力系统的源荷预测技术 • 上一篇下一篇

基于复杂特征提取和Sinkhorn距离的风光荷多阶段场景树生成方法

王蕊(), 傅质馨(), 王健(), 刘皓明()

河海大学电气与动力工程学院，江苏南京　211100

收稿日期:2024-02-07 出版日期:2024-12-28 发布日期:2024-12-27
作者简介:王蕊（2000—），女，硕士研究生，从事电力系统规划研究，E-mail：221606030095@hhu.edu.cn
傅质馨（1983—），女，博士，副教授，从事可再生能源发电技术和物联网技术研究，E-mail：zhixinfu@hhu.edu.cn
王健（1993—），男，通信作者，博士，讲师，从事配电网运行与规划研究，E-mail：eewangjian@hhu.edu.cn
刘皓明（1977—），男，博士，教授，从事智能电网、电力系统优化运行和电力市场研究，E-mail：liuhaom@hhu.edu.cn
基金资助:
国家自然科学基金资助项目（面向规模化异质弹性资源接入的配电系统可行域表征及分层协同优化运行研究，52207091）；江苏省自然科学基金资助项目（BK20220977）。

A Multi-stage Scenario Tree Generation Method for Wind-Solar Load Based on Complex Feature Extraction and Sinkhorn Distance

Rui WANG(), Zhixin FU(), Jian WANG(), Haoming LIU()

School of Electrical and Power Engineering, Hohai University, Nanjing 211100, China

Received:2024-02-07 Online:2024-12-28 Published:2024-12-27
Supported by:
This work is supported by National Natural Science Foundation of China (Research on Feasible Region Characterization and Hierarchically Coordinated Optimal Operation of Distribution System with Large-Scale Heterogeneous Elastic Resource, No.52207091), Natural Science Foundation of Jiangsu Province (No.BK20220977).

摘要/Abstract

摘要：

新能源发电出力和负荷长期增长的不确定性增加了电网规划复杂性，开展新能源出力和负荷长时间尺度上的不确定性分析，对电网的规划与建设具有重要意义。提出了一种基于复杂特征提取和Sinkhorn距离的风光荷多阶段场景树生成方法。首先，为提高风光荷场景的聚类效率，提出基于堆叠稀疏自编码器的风光荷场景特征提取方法，并采用基于密度峰值改进的近邻传播算法对风光荷场景特征集合进行聚类，获得风光荷典型曲线，作为场景树的根节点；然后，考虑负荷不同增长率，逐年生成风光荷场景树，并提出基于Sinkhorn距离的场景树削减方法以降低场景树的规模；最后，算例仿真结果表明，所提方法计算效率高，生成的风光荷多阶段场景树可反映风光出力和负荷增长的不确定性。

关键词: 场景树, 堆叠稀疏自编码器, 改进近邻传播算法, Sinkhorn距离, 负荷不确定性

Abstract:

The uncertainty in the long-term growth of renewable energy generation output and load has heightened the complexity of power grid planning. Conducting an uncertainty analysis on the long-term scale of renewable energy output and load is of significant importance for the planning and construction of the power grid. To address this issue, a multi-stage scenario tree generation method for wind-solar load based on complex feature extraction and Sinkhorn distance was proposed. Firstly, to enhance the clustering efficiency of wind-solar load scenarios, a method based on stacked sparse autoencoders for feature extraction of wind-solar load scenarios was introduced. The feature set of wind-solar load scenarios was clustered by using an improved affinity propagation algorithm based on density peak, and typical curves of wind-solar load were obtained as the root nodes of the scenario tree. Subsequently, by considering different growth rates in load, a yearly generation of wind-solar load scenario trees was performed, and a scenario tree reduction method based on Sinkhorn distance was proposed to reduce the size of the scenario tree. Finally, a simulation example showed that the proposed method had high calculation efficiency, and the generated multi-stage scenario tree for wind-solar load can reflect the uncertainty of wind-solar output and load growth.

Key words: scenario tree, stacked sparse autoencoder, improved affinity propagation algorithm, Sinkhorn distance, load uncertainty

王蕊, 傅质馨, 王健, 刘皓明. 基于复杂特征提取和Sinkhorn距离的风光荷多阶段场景树生成方法[J]. 中国电力, 2024, 57(12): 30-40.

Rui WANG, Zhixin FU, Jian WANG, Haoming LIU. A Multi-stage Scenario Tree Generation Method for Wind-Solar Load Based on Complex Feature Extraction and Sinkhorn Distance[J]. Electric Power, 2024, 57(12): 30-40.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202402033

https://www.electricpower.com.cn/CN/Y2024/V57/I12/30

图/表 11

图 1 堆叠稀疏自编码器模型

Fig.1 Stacked sparse autoencoder model

图 2 场景树

Fig.2 Scenario tree

图 3 3阶段场景树约简示例

Fig.3 Example of three-stage scenario tree generation

图 4 风光荷全时序场景集

Fig.4 Full temporal scenario set of wind-solar load

图 5 基于SSAE和改进AP聚类效果

Fig.5 Clustering effect based on SSAE and improved AP

表 1 典型日概率

Table 1 Typical day probabilities

序号		概率
1		0.2329
2		0.2658
3		0.5013

表 2 聚类结果

Table 2 Clustering results

聚类方法	簇数	轮廓系数	CH指标	DB指标	迭代次数/次	求解时间/s
原始AP	23	0.20	31.66	0.97	75	28.1429
改进AP	3	0.30	108.46	0.97	69	19.2524
k-means	3	0.23	104.79	0.88
PCA+改进AP	3	0.35	133.17	0.71	69	18.0893
SSAE+改进AP	3	0.42	151.71	0.65	66	17.2524

图 6 3阶段风光荷场景树

Fig.6 Three-stage wind-solar load scenario tree

图 7 节点1、2、3代表的风光荷场景

Fig.7 Wind-solar load scenarios represented by nodes 1, 2 and 3

图 8 节点1、4、10代表的风光荷场景

Fig.8 Wind-solar load scenarios represented by nodes 1, 4 and 10

表 3 3种方法计算的过程距离和时间

Table 3 Process distance and time calculated by three methods

方法	过程距离	总计算时间/s
基于Kantorovich距离和线性规划结合方法^[21,23]	0.1279	0.2785
基于k-means和线性规划结合方法^[21]	0.0855	0.2952
基于传统最优传输理论的线性规划方法^[22]	0.1479	0.2139
本文方法	0.0448	0.1947

参考文献 28

1	李华, 程子月, 李旭东, 等. 考虑出力与电价不确定性的光伏集群有功功率分配方法[J]. 中国电力, 2023, 56 (10): 211- 218.
	LI Hua, CHENG Ziyue, LI Xudong, et al. Active power allocation method for photovoltaic cluster considering output and electricity price uncertainty[J]. Electric Power, 2023, 56 (10): 211- 218.
2	陈国平, 董昱, 梁志峰. 能源转型中的中国特色新能源高质量发展分析与思考[J]. 中国电机工程学报, 2020, 40 (17): 5493- 5506.
	CHEN Guoping, DONG Yu, LIANG Zhifeng. Analysis and reflection on high-quality development of new energy with Chinese characteristics in energy transition[J]. Proceedings of the CSEE, 2020, 40 (17): 5493- 5506.
3	袁铁江, 杨洋, 李瑞, 等. 考虑源荷不确定性的氢能微网容量优化配置[J]. 中国电力, 2023, 56 (7): 21- 32.
	YUAN Tiejiang, YANG Yang, LI Rui, et al. Optimized configuration of hydrogen-energy microgrid capacity considering source charge uncertainties[J]. Electric Power, 2023, 56 (7): 21- 32.
4	汪宁渤, 马明, 强同波, 等. 高比例新能源电力系统的发展机遇、挑战及对策[J]. 中国电力, 2018, 51 (1): 29- 35, 50.
	WANG Ningbo, MA Ming, QIANG Tongbo, et al. High-penetration new energy power system development: challenges, opportunities and countermeasures[J]. Electric Power, 2018, 51 (1): 29- 35, 50.
5	张辰毓, 许刚. 分布式多元随机动态场景生成及快速精准场景降维算法[J]. 电网技术, 2022, 46 (2): 671- 679.
	ZHANG Chenyu, XU Gang. Distributed multivariate random dynamic scenario generation and fast & accurate scenario simplified algorithm[J]. Power System Technology, 2022, 46 (2): 671- 679.
6	DING T, HU Y, BIE Z H. Multi-stage stochastic programming with nonanticipativity constraints for expansion of combined power and natural gas systems[J]. IEEE Transactions on Power Systems, 2018, 33 (1): 317- 328. DOI
7	赵雪楠, 段凯悦, 李萌, 等. 考虑调度决策非预期性的多阶段随机规划调度策略[J]. 电力建设, 2022, 43 (10): 87- 97. DOI
	ZHAO Xuenan, DUAN Kaiyue, LI Meng, et al. Scheduling strategy of multi-stage stochastic programming considering non-anticipativity of scheduling decision[J]. Electric Power Construction, 2022, 43 (10): 87- 97. DOI
8	KAUT M, STEIN W. Evaluation of scenario-generation methods for stochastic programming[M]. Humboldt-Universität Zu Berlin, Mathematisch-Naturwissenschaftliche Fakultät II, Institut für Mathematik, 2003.
9	GÜLPıNAR N, RUSTEM B, SETTERGREN R. Simulation and optimization approaches to scenario tree generation[J]. Journal of Economic Dynamics and Control, 2004, 28 (7): 1291- 1315. DOI
10	RUBASHEUSKI U, OPPEN J, WOODRUFF D L. Multi-stage scenario generation by the combined moment matching and scenario reduction method[J]. Operations Research Letters, 2014, 42 (5): 374- 377. DOI
11	LATORRE J M, CERISOLA S, RAMOS A. Clustering algorithms for scenario tree generation: application to natural hydro inflows[J]. European Journal of Operational Research, 2007, 181 (3): 1339- 1353. DOI
12	LEI Y, WANG D, JIA H J, et al. Multi-stage stochastic planning of regional integrated energy system based on scenario tree path optimization under long-term multiple uncertainties[J]. Applied Energy, 2021, 300, 117224. DOI
13	张颖, 韩风. 考虑时间一致性的电力系统风险规避多阶段随机规划建模与求解[J]. 电力系统及其自动化学报, 2022, 34 (3): 28- 36.
	ZHANG Ying, HAN Feng. Risk aversion multi-stage stochastic programming for the modeling and solving of power system considering time consistency[J]. Proceedings of the CSU-EPSA, 2022, 34 (3): 28- 36.
14	余轶, 陈峰, 曾杨, 等. 计及新能源发展目标的电力系统灵活资源多阶段优化配置[J]. 武汉大学学报(工学版), 2024, 57 (10): 1394- 1405.
	YU Yi, CHEN Feng, ZENG Yang, et al. Multi-stage optimal planning of flexible resources in power system considering renewable energy developing target[J]. Engineering Journal of Wuhan University, 2024, 57 (10): 1394- 1405.
15	TANG L, SUN L L, GUO C H, et al. Adaptive spectral affinity propagation clustering[J]. Journal of Systems Engineering and Electronics, 2022, 33 (3): 647- 664. DOI
16	金伟超, 张旭, 刘晟源, 等. 基于剪枝策略和密度峰值聚类的行业典型负荷曲线辨识[J]. 电力系统自动化, 2021, 45 (4): 20- 28. DOI
	JIN Weichao, ZHANG Xu, LIU Shengyuan, et al. Identification of typical industrial power load curves based on pruning strategy and density peak clustering[J]. Automation of Electric Power Systems, 2021, 45 (4): 20- 28. DOI
17	徐来烽, 张沈习, 叶琳浩, 等. 考虑动态重构和智能软开关接入的配电网源网荷储联合规划[J]. 南方电网技术, 2024, 18 (4): 130- 140.
	XU Laifeng, ZHANG Shenxi, YE Linhao, et al. Joint planning of source-network-load-storage in distribution network considering dynamic reconfiguration and intelligent soft open point[J]. Southern Power System Technology, 2024, 18 (4): 130- 140.
18	XIONG Y H, LU Y H. Deep feature extraction from the vocal vectors using sparse autoencoders for Parkinson’s classification[J]. IEEE Access, 2020, 8, 27821- 27830. DOI
19	SHI C M, LUO B, HE S P, et al. Tool wear prediction via multidimensional stacked sparse autoencoders with feature fusion[J]. IEEE Transactions on Industrial Informatics, 2020, 16 (8): 5150- 5159. DOI
20	丁涛, 李澄, 胡源, 等. 考虑非预期条件的电力系统多阶段随机规划建模理论与方法[J]. 电网技术, 2017, 41 (11): 3566- 3573.
	DING Tao, LI Cheng, HU Yuan, et al. Multi-stage stochastic programming for power system planning considering nonanticipative constraints[J]. Power System Technology, 2017, 41 (11): 3566- 3573.
21	XU D B, CHEN Z P, YANG L. Scenario tree generation approaches using K-means and LP moment matching methods[J]. Journal of Computational and Applied Mathematics, 2012, 236 (17): 4561- 4579. DOI
22	MEIRA L A A, COELHO G P, SANTOS A A S, et al. Selection of representative models for decision analysis under uncertainty[J]. Computers & Geosciences, 2016, 88, 67- 82.
23	WANG C, GAO R, WEI W, et al. Risk-based distributionally robust optimal gas-power flow with Wasserstein distance[J]. IEEE Transactions on Power Systems, 2019, 34 (3): 2190- 2204. DOI
24	李晖, 刘栋, 梁涵卿, 等. 基于条件生成对抗网络随机场景的电力系统日前多阶段优化调度[J]. 新型电力系统, 2023, 1 (3): 272- 282.
	LI Hui, LIU Dong, LIANG Hanqing, et al. Conditional-generative-adversarial-network-based day-ahead multi-stage stochastic scheduling of power systems[J]. New Type Power Systems, 2023, 1 (3): 272- 282.
25	聂晓音, 谢刚, 李洋, 等. 基于栈式相关性稀疏自编码的电力通信网故障诊断[J]. 电力系统保护与控制, 2019, 47 (19): 158- 163.
	NIE Xiaoyin, XIE Gang, LI Yang, et al. Fault diagnosis of power communication network based on stacked relational sparse autoencoder[J]. Power System Protection and Control, 2019, 47 (19): 158- 163.
26	杨悦, 陈宇航, 成龙, 等. 考虑节点功率储备与中心性的主动配电网动态集群电压控制[J]. 电网技术, 2024, 48 (2): 618- 629.
	YANG Yue, CHEN Yuhang, CHENG Long, et al. Power reserve and GIN centrality of buses considered dynamic cluster voltage control of active distribution networks[J]. Power System Technology, 2024, 48 (2): 618- 629.
27	PEYRÉ G, CUTURI M. Computational optimal transport: with applications to data science[J]. Foundations and Trends® in Machine Learning, 2019, 11 (5/6): 355- 607.
28	高崇, 周鹏, 罗强, 等. 考虑实际负荷增长模式下的配电网供电能力评估方法[J]. 南方电网技术, 2023, 17 (7): 115- 124.
	GAO Chong, ZHOU Peng, LUO Qiang, et al. Evaluation method of load supply capability of distribution network considering actual load growth mode[J]. Southern Power System Technology, 2023, 17 (7): 115- 124.

基于复杂特征提取和Sinkhorn距离的风光荷多阶段场景树生成方法

A Multi-stage Scenario Tree Generation Method for Wind-Solar Load Based on Complex Feature Extraction and Sinkhorn Distance

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 28

相关文章 1

编辑推荐

Metrics

模态框（Modal）标题

基于复杂特征提取和Sinkhorn距离的风光荷多阶段场景树生成方法

A Multi-stage Scenario Tree Generation Method for Wind-Solar Load Based on Complex Feature Extraction and Sinkhorn Distance

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 28

相关文章 1

编辑推荐

Metrics