基于自适应多任务扩散模型的风光荷场景生成方法

doi:10.3969/j.issn.2097-0706.2025.11.002

综合智慧能源 ›› 2025, Vol. 47 ›› Issue (11): 14-23.doi: 10.3969/j.issn.2097-0706.2025.11.002

• 灵活资源调控与协同优化 • 上一篇下一篇

基于自适应多任务扩散模型的风光荷场景生成方法

张志洪¹(), 胡旭光¹^,^*(), 吴恩凯²(), 张弛¹(), 周诚浩¹()

1.东北大学信息科学与工程学院，沈阳 110819
2.辽宁省电子信息产品监督检验院，沈阳 110053

收稿日期:2025-06-16 修回日期:2025-08-04 出版日期:2025-09-30
通讯作者: *胡旭光（1992），男，副教授，博士，从事数模混合驱动的能源系统智能化建模、综合高效利用与优化调控等方面的研究，huxuguang@mail.neu.edu.cn。
作者简介:张志洪（2003），男，硕士生，从事电力系统数据挖掘与分析方面的研究，zhihongxuexi@163.com；
吴恩凯（1977），男，工程师，从事电子产品检测方面的研究，wuenkai1977@163.com；
张弛（2001），男，硕士生，从事非侵入式电力负荷监测方面的研究，2400739@stu.neu.edu.cn；
周诚浩（2002），男，硕士生，从事电力系统数据挖掘和电力系统场景生成等方面的研究，2371042@stu.neu.edu.cn。
基金资助:
国家自然科学基金项目(62303103);国家自然科学基金项目(62373089);中央高校基本科研业务费专项资金项目(N25ZJL020);辽宁省自然科学基金项目(2023-BSBA-140);辽宁省教育厅科研项目(JYTQN2023161)

Wind-solar-load scenario generation method based on adaptive multi-task diffusion model

ZHANG Zhihong¹(), HU Xuguang¹^,^*(), WU Enkai²(), ZHANG Chi¹(), ZHOU Chenghao¹()

1. College of Information Science and Engineering，Northeastern University，Shenyang 110819，China
2. Liaoning Electronic Product Supervision and Inspection Institute，Shenyang 110053，China

Received:2025-06-16 Revised:2025-08-04 Published:2025-09-30
Supported by:
National Natural Science Foundation of China(62303103);National Natural Science Foundation of China(62373089);Fundamental Research Funds for the Central Universities in ChinaFundamental Research Funds for the Central Universities in China(N25ZJL020);Natural Science Foundation of Liaoning Province(2023-BSBA-140);Scientific Research Project of Liaoning Provincial Department of Education(JYTQN2023161)

摘要/Abstract

摘要：

高比例可再生能源并网背景下，精准构建能捕捉各变量动态特性及复杂关联的风光荷联合场景，是电力系统调度与控制的核心需求，为此提出了基于自适应多任务扩散模型的风光荷场景生成方法。建立了基于联合去噪网络的多任务扩散学习架构，通过联合处理多变量状态向量并融合时间信息，生成兼具物理耦合关系及时间依赖模式的真实联合场景；在此基础上，提出了基于异构数据动态特征引导的自适应扩散策略模块，通过提取生成数据的动态统计特征并据此动态调整扩散过程的噪声调度，有效表征数据的非平稳和时变动态特性；同时，引入结构化一致性引导的训练准则，在训练目标中约束边际分布与联合依赖两种数据的结构特性，实现了对模型生成过程的有效引导，提高风光荷场景的生成质量。基于IES-134标准数据集的验证表明，所提模型能够有效生成物理特性真实且统计规律合理的风光荷联合场景，可为电力系统的优化调度与风险评估提供实用工具。

关键词: 可再生能源, 风光荷联合场景, 自适应多任务扩散模型, 去噪扩散概率模型, 多任务学习, 动态特性自适应

Abstract:

In the context of high-penetration renewable energy integration， accurately constructing wind-solar-load joint scenarios that can capture the dynamic characteristics and complex correlations of various variables has emerged as a core requirement for power system scheduling and control. To this end， a generation method for wind-solar-load scenarios based on an adaptive multi-task diffusion model was proposed. A multi-task diffusion model learning architecture based on a joint denoising network was established. By jointly processing multivariate state vectors and integrating temporal information， realistic joint scenarios that integrated physical coupling relationships and temporal dependency patterns were generated. On this basis， an adaptive diffusion strategy module guided by the dynamic features of heterogeneous data was proposed. Dynamic statistical features of the generated data were extracted， and the noise scheduling of the diffusion process was dynamically adjusted accordingly， thereby effectively characterizing the non-stationary and time-varying dynamic characteristics of the data. Meanwhile， a training criterion guided by structured consistency was introduced， where the marginal distribution and joint dependency characteristics of the data structure were constrained within the training objectives. It effectively guided the model generation process and improved the quality of wind-solar-load scenario generation. Validation based on the IES-134 standard dataset demonstrated that the proposed model could effectively generate wind-solar-load joint scenarios with realistic physical characteristics and reasonable statistical patterns， providing a practical tool for optimal scheduling and risk assessment in power systems.

Key words: renewable energy, wind-solar-load joint scenarios, adaptive multi-task diffusion model, denoising diffusion probabilistic model, multi-task learning, dynamic feature adaptation

中图分类号:

TK01

张志洪, 胡旭光, 吴恩凯, 张弛, 周诚浩. 基于自适应多任务扩散模型的风光荷场景生成方法[J]. 综合智慧能源, 2025, 47(11): 14-23.

ZHANG Zhihong, HU Xuguang, WU Enkai, ZHANG Chi, ZHOU Chenghao. Wind-solar-load scenario generation method based on adaptive multi-task diffusion model[J]. Integrated Intelligent Energy, 2025, 47(11): 14-23.

图/表 11

图1

图2

表1

图3

图4

图5

图6

表2

图7

图8

表3

参考文献 22

[1]	LI Z, YE L, SONG X R, et al. Heterogeneous spatiotemporal graph convolution network for multi-modal wind-PV power collaborative prediction[J]. IEEE Transactions on Power Systems, 2024, 39(4): 5591-5608.
[2]	张冬冬, 单琳珂, 刘天皓. 人工智能技术在风力与光伏发电数据挖掘及功率预测中的应用综述[J]. 综合智慧能源, 2025, 47(3): 32-46.
	ZHANG Dongdong, SHAN Linke, LIU Tianhao. Review on the application of artificial intelligence in data mining and wind and photovoltaic power forecasting[J]. Integrated Intelligent Energy, 2025, 47(3): 32-46.
[3]	孙秋野, 于潇寒, 王靖傲. “双高” 配电系统的挑战与应对措施探讨[J]. 中国电机工程学报, 2024, 44(18): 7115-7136.
	SUN Qiuye, YU Xiaohan, WANG Jing'ao. Discussion on challenges and countermeasures of "double high" power distribution system[J]. Proceedings of the CSEE, 2024, 44(18): 7115-7136.
[4]	翁志鹏, 周京华, 李津, 等. 含风光接入的微电网可靠性影响分析[J]. 综合智慧能源, 2023, 45(1): 67-74.
	WENG Zhipeng, ZHOU Jinghua, LI Jin, et al. Impact of wind and solar power grid connection on microgrid reliability[J]. Integrated Intelligent Energy, 2023, 45(1): 67-74.
[5]	欧阳婷, 蔡晔, 王炜宇, 等. 计及风电、光伏预测不确定性的抽水蓄能日前全调度优化[J]. 综合智慧能源, 2022, 44(11): 20-27.
	OUYANG Ting, CAI Ye, WANG Weiyu, et al. Overall day-ahead scheduling optimization for pumped-storage power stations considering the uncertainty of wind and photovoltaic power prediction[J]. Integrated Intelligent Energy, 2022, 44(11): 20-27.
[6]	王利利, 王皓, 任洲洋, 等. 计及灵活资源调节潜力的高压配电网新能源接纳能力评估[J]. 中国电力, 2022, 55(10): 124-131.
	WANG Lili, WANG Hao, REN Zhouyang, et al. Evaluation of renewable energy accommodation capacity of high voltage distribution networks considering regulation potential of flexible resources[J]. Electric Power, 2022, 55(10): 124-131.
[7]	ZHENG K, SUN Z Y, SONG Y, et al. Stochastic scenario generation methods for uncertainty in wind and photovoltaic power outputs: A comprehensive review[J]. Energies, 2025, 18(3): 503.
[8]	RAYATI M, BOZORG M, CARPITA M, et al. Stochastic optimization and Markov chain-based scenario generation for exploiting the underlying flexibilities of an active distribution network[J]. Sustainable Energy, Grids and Networks, 2023, 34: 100999.
[9]	BECKER R. Generation of time-coupled wind power infeed scenarios using pair-copula construction[J]. IEEE Transactions on Sustainable Energy, 2018, 9(3): 1298-1306.
[10]	王玲玲, 王昕, 郑益慧, 等. 计及多个风电机组出力相关性的配电网无功优化[J]. 电网技术, 2017, 41(11): 3463-3469.
	WANG Lingling, WANG Xin, ZHENG Yihui, et al. Reactive power optimization of distribution network considering output correlation of multiple wind turbines[J]. Power System Technology, 2017, 41(11): 3463-3469.
[11]	胡旭光, 张化光, 孙秋野, 等. 基于异构数据模型的电-气耦合网络状态感知方法研究[J]. 中国科学: 技术科学, 2020, 50(3): 346-360.
	HU Xuguang, ZHANG Huaguang, SUN Qiuye, et al. Research on situation awareness of electricity-gas coupling network based on a heterogeneous data model[J]. Scientia Sinica (Technologica), 2020, 50(3): 346-360.
[12]	WANG C G, SHARIFNIA E, GAO Z, et al. Generating multivariate load states using a conditional variational autoencoder[J]. Electric Power Systems Research, 2022, 213: 108603.
[13]	胡旭光, 马大中, 郑君, 等. 基于关联信息对抗学习的综合能源系统运行状态分析方法[J]. 自动化学报, 2020, 46(9): 1783-1797.
	HU Xuguang, MA Dazhong, ZHENG Jun, et al. An operation state analysis method for integrated energy system based on correlation information adversarial learning[J]. Acta Automatica Sinica, 2020, 46(9): 1783-1797.
[14]	HO J, JAIN A, ABBEEL P. Denoising diffusion probabilistic models[J]. Advances in Neural Information Processing Systems, 2020, 33: 6840-6851.
[15]	黄越辉, 孙亚南, 李驰, 等. 基于条件生成对抗网络的多区域风电短期出力场景生成方法[J]. 电网技术, 2023, 47(1): 63-77.
	HUANG Yuehui, SUN Yanan, LI Chi, et al. Constructing method of short-term output scenarios for multi-regional wind power based on conditional generative adversarial network[J]. Power System Technology, 2023, 47(1): 63-77.
[16]	詹兆康, 胡旭光, 赵浩然, 等. 基于多变量时空融合网络的风机数据缺失值插补研究[J]. 自动化学报, 2024, 50(6): 1171-1184.
	ZHAN Zhaokang, HU Xuguang, ZHAO Haoran, et al. Study of missing value imputation in wind turbine data based on multivariate spatiotemporal integration network[J]. Acta Automatica Sinica, 2024, 50(6): 1171-1184.
[17]	戴宇欣, 张俊, 乔骥, 等. 基于改进去噪扩散概率模型和模型迁移的新能源场站超短期出力场景生成[J]. 电网技术, 2025, 49(2): 511-521.
	DAI Yuxin, ZHANG Jun, QIAO Ji, et al. Ultra-short-term output scenario generation for renewable energy plants based on improved denoising diffusion probabilistic models and model-based transfer learning[J]. Power System Technology, 2025, 49(2): 511-521.
[18]	王长刚, 刘伟, 曹宇, 等. 基于改进条件生成扩散模型的新能源日前场景生成方法[J]. 电网技术, 2025, 49(4): 1358-1368.
	WANG Changgang, LIU Wei, CAO Yu, et al. The day-ahead scenario generation method for new energy based on an improved conditional generative diffusion model[J]. Power System Technology, 2025, 49(4): 1358-1368.
[19]	DONG X C, MAO Z H, SUN Y Y, et al. Short-term wind power scenario generation based on conditional latent diffusion models[J]. IEEE Transactions on Sustainable Energy, 2024, 15(2): 1074-1085.
[20]	SILVA W N, BANDÓRIA L H T, DIAS B H, et al. Generating realistic load profiles in smart grids: An approach based on nonlinear independent component estimation (NICE) and convolutional layers[J]. Applied Energy, 2023, 351: 121902.
[21]	宋晓维. 基于变分自编码高斯混合模型的海量新能源出力场景生成方法[J]. 电气自动化, 2024, 46(3): 1-3.
	SONG Xiaowei. Method for generating massive new energy output scenarios based on variational autocoding and hidden variable modeling[J]. Electrical Automation, 2024, 46(3): 1-3.
[22]	许琴, 刘一鸣, 陈志刚, 等. 负荷水平约束条件下新能源出力场景生成新方法[J]. 南方电网技术, 2025, 19(1): 51-62.
	XU Qin, LIU Yiming, CHEN Zhigang, et al. A new method for generating new energy output scenarios with load level constraints[J]. Southern Power System Technology, 2025, 19(1): 51-62.

项目	节点数	峰值运行功率/MW	工作模式
用户负荷	58	81.4	PQ
光伏出力	15	15.3	PQ
风机出力	5	37.0	PV,PQ

模型	变量	KL散度	JS散度	MAE/MW	RMSE/MW
VAE	负荷	0.426	0.146	0.402	0.751
	风电	0.534	0.234	0.550	1.024
	光伏	0.521	0.177	0.523	0.951
T-Copula	负荷	0.433	0.126	0.386	0.705
	风电	0.558	0.193	0.522	0.985
	光伏	0.473	0.187	0.485	0.895
GAN	负荷	0.379	0.098	0.347	0.663
	风电	0.484	0.176	0.489	0.894
	光伏	0.427	0.154	0.438	0.823
IDDPM	负荷	0.365	0.088	0.341	0.650
	风电	0.450	0.162	0.475	0.868
	光伏	0.408	0.135	0.433	0.805
ICGDM	负荷	0.371	0.091	0.344	0.658
	风电	0.462	0.168	0.480	0.877
	光伏	0.415	0.141	0.436	0.811
CLDM	负荷	0.360	0.085	0.338	0.644
	风电	0.441	0.159	0.470	0.859
	光伏	0.401	0.131	0.431	0.799
NICE	负荷	0.375	0.094	0.351	0.670
	风电	0.478	0.172	0.495	0.901
	光伏	0.420	0.148	0.442	0.836
AMDM	负荷	0.352	0.080	0.335	0.635
	风电	0.425	0.154	0.464	0.847
	光伏	0.394	0.127	0.429	0.792

模型	训练时间/min	场景推理时间/s
VAE	27.26	0.032
T-Copula	4.87	0.007
GAN	56.32	0.019
IDDPM	147.45	1.312
ICGDM	137.35	1.782
CLDM	112.24	1.145
NICE	40.52	0.025
AMDM	117.12	1.082

基于自适应多任务扩散模型的风光荷场景生成方法

Wind-solar-load scenario generation method based on adaptive multi-task diffusion model

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 22

相关文章 15

编辑推荐

Metrics

本文评价

[1]	李晓宁, 孙娜, 黄阿敏, 董海鹰. 基于蛇优化算法的PEMFC空气供给系统模糊自抗扰控制[J]. 综合智慧能源, 2025, 47(6): 57-73.
[2]	叶幸, 邱辰, 艾琳, 辛颂旭, 郭雁珩. “双碳”目标下可再生能源绿证与碳交易衔接机制的思考[J]. 综合智慧能源, 2025, 47(5): 12-20.
[3]	陈浩, 马刚, 钱达, 马健, 彭乐瑶. 绿证-碳交易机制下考虑阶梯需求响应的区域综合能源系统优化调度[J]. 综合智慧能源, 2025, 47(5): 21-30.
[4]	张冬冬, 单琳珂, 刘天皓. 人工智能技术在风力与光伏发电数据挖掘及功率预测中的应用综述[J]. 综合智慧能源, 2025, 47(3): 32-46.
[5]	张敏, 王雁河, 孙周, 刘斌, 姜之未, 曹灿, 高峰. 交能融合应用场景与技术体系研究[J]. 综合智慧能源, 2025, 47(2): 13-28.
[6]	庄富豪, 吴军, 朱瑞金, 吴红梅, 田润, 薛雨润. 绿电制氢产业关键技术现状及展望[J]. 综合智慧能源, 2025, 47(10): 1-9.
[7]	苏锐, 王希龙, 江阎, 宋晨辉. 基于改进成本分配法的虚拟电厂调度规划[J]. 综合智慧能源, 2025, 47(1): 10-17.
[8]	徐强. 基于NSGA-Ⅲ算法的分布式光伏储能系统优化配置方法[J]. 综合智慧能源, 2025, 47(1): 26-33.
[9]	聂雪颖, 程懋松, 左献迪, 戴志敏. 考虑风光消纳的风光核储混合系统容量优化[J]. 综合智慧能源, 2025, 47(1): 51-61.
[10]	王晓燕, 吴书泉. 基于改进粒子群优化算法的源网荷储系统容量配置研究[J]. 综合智慧能源, 2024, 46(9): 28-36.
[11]	樊颜搏, 熊亚选, 李想, 田曦, 杨洋. 基于遗传算法的建筑用能多目标优化应用进展[J]. 综合智慧能源, 2024, 46(9): 69-85.
[12]	邓振宇, 汪茹康, 徐钢, 云昆, 王颖. 综合能源系统中热电联产机组故障预警现状[J]. 综合智慧能源, 2024, 46(8): 67-76.
[13]	殷林飞, 蒙雨洁. 基于DenseNet卷积神经网络的短期风电预测方法[J]. 综合智慧能源, 2024, 46(7): 12-20.
[14]	郑庆明, 井延伟, 梁涛, 柴露露, 吕梁年. 基于DDPG算法的离网型可再生能源大规模制氢系统优化调度[J]. 综合智慧能源, 2024, 46(6): 35-43.
[15]	董强, 徐君, 方东平, 方丽娟, 陈妍琼. 基于光伏出力特性的分布式光储系统优化调度策略[J]. 综合智慧能源, 2024, 46(4): 17-23.