基于改进Transformer-CGAN的长时间尺度光伏出力场景生成方法

叶禹江; 轩顺德; 师瑞峰; 贾利民

doi:10.19912/j.0254-0096.tynxb.2025-0529

PDF(1805 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (8) : 169-177. DOI: 10.19912/j.0254-0096.tynxb.2025-0529

基于改进Transformer-CGAN的长时间尺度光伏出力场景生成方法

叶禹江¹, 轩顺德¹, 师瑞峰^1,2, 贾利民^2,3

作者信息 +

LONG-TERM PHOTOVOLTAIC POWER OUTPUT SCENARIO GENERATION METHOD BASED ON IMPROVED TRANSFORMER-CGAN

Ye Yujiang¹, Xuan Shunde¹, Shi Ruifeng^1,2, Jia Limin^2,3

Author information +

文章历史 +

摘要

针对传统方法难以有效生成长时间尺度光伏出力不确定性场景问题,提出一种结合改进Transformer与条件生成对抗网络（CGAN）的长时间尺度光伏出力场景生成方法。首先,基于Transformer及CGAN设计适用于长时间光伏出力场景生成的Transformer-CGAN（T-CGAN）模型;其次,提出一种多时间尺度周期注意力机制,捕捉光伏出力的长短期周期性特征;最后,以中国西北地区某光伏电站历史数据为例,将所提方法与现有3种典型模型进行对比实验,并从统计指标、时间相关性和生成场景有效性3方面验证方法的有效性。实验结果表明,该方法能够有效捕捉光伏出力长短期时间依赖性,准确有效地生成全年光伏出力场景集。

Abstract

To overcome the limitations of traditional methods in effectively generating long-term photovoltaic （PV） power output uncertainty scenarios, this paper proposes a novel approach that integrates an improved Transformer model with a conditional generative adversarial network （CGAN）. Specifically, a Transformer-CGAN （T-CGAN） architecture is developed for generating long-term PV output scenarios. To better capture the inherent periodicity of PV output, a multi-timescale periodic attention mechanism is introduced, enabling the model to learn both short-term and long-term temporal patterns. A case study using historical data from a PV power plant in Northwest China is conducted to compare the proposed method with three benchmark models. The performance is evaluated from three perspectives： statistical accuracy, temporal correlation, and scenario validity. Experimental results demonstrate that the proposed approach significantly improves the accuracy of scenario generation, effectively captures temporal dependencies across multiple time scales, and enables efficient generation of annual PV output scenario sets.

导出引用

叶禹江, 轩顺德, 师瑞峰, 贾利民. 基于改进Transformer-CGAN的长时间尺度光伏出力场景生成方法[J]. 太阳能学报. 2026, 47(8): 169-177 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0529

Ye Yujiang, Xuan Shunde, Shi Ruifeng, Jia Limin. LONG-TERM PHOTOVOLTAIC POWER OUTPUT SCENARIO GENERATION METHOD BASED ON IMPROVED TRANSFORMER-CGAN[J]. Acta Energiae Solaris Sinica. 2026, 47(8): 169-177 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0529

中图分类号： TM615

参考文献

[1] 杨轶航, 韩璐, 史华勃, 等. 基于相似日与ISC-BiLSTM的短期光伏功率预测方法[J]. 太阳能学报, 2025, 46(1): 676-685.
Yang Y H, Han L, Shi H B, et al.Short-term photovoltaic power forecast method based on similar days and ISC-BiLSTM[J]. Acta Energiae Solaris Sinica, 2025, 46(1): 676-685.
[2] Li J H, Zhou J S, Chen B.Review of wind power scenario generation methods for optimal operation of renewable energy systems[J]. Applied Energy, 2020, 280: 115992.
[3] 董骁翀, 张姝, 李烨, 等. 电力系统中时序场景生成和约简方法研究综述[J]. 电网技术, 2023, 47(2): 709-721.
Dong X C, Zhang S, Li Y, et al.Review of power system temporal scenario generation and reduction methods[J]. Power System Technology, 2023, 47(2): 709-721.
[4] Chen H P, Gao L, Zhang Y L, et al.Optimal scheduling strategy of a regional integrated energy system considering renewable energy uncertainty and heat network transmission characteristics[J]. Energy Reports, 2022, 8: 7691-7703.
[5] 涂青宇, 苗世洪, 林毓军, 等. 基于动态R藤Copula模型的区域风电集群超短期功率区间预测方法[J]. 高电压技术, 2022, 48(2): 456-466.
Tu Q Y, Miao S H, Lin Y J, et al.Ultra-short-term interval forecasting method for regional wind farms based on dynamic R-vine Copula model[J]. High Voltage Engineering, 2022, 48(2): 456-466.
[6] 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.
[7] 米阳, 卢长坤, 申杰, 等. 基于条件生成对抗网络的风电功率极端场景生成[J]. 高电压技术, 2023, 49(6): 2253-2263.
Mi Y, Lu C K, Shen J, et al.Wind power extreme scenario generation based on conditional generative adversarial network[J]. High Voltage Engineering, 2023, 49(6): 2253-2263.
[8] 吴艳梅, 陈红坤, 陈磊, 等. 基于改进MMD-GAN的可再生能源随机场景生成[J]. 电力系统保护与控制, 2024, 52(19): 85-96.
Wu Y M, Chen H K, Chen L, et al.Stochastic scenario generation for renewable energy based on improved MMD-GAN[J]. Power System Protection and Control, 2024, 52(19): 85-96.
[9] Yilmaz B, Korn R.A Comprehensive guide to generative adversarial networks (GANs) and application to individual electricity demand[J]. Expert Systems with Applications, 2024, 250: 123851.
[10] Goodfellow I J, Pouget-Abadie J, Mirza M, et al.Generative adversarial nets[C]//Proceedings of the 27th International Conference on Neural Information Processing Systems. Montreal: ACM, 2014: 2672-2680.
[11] Mansour S H, Azzam S M, Hasanien H M, et al.Wasserstein generative adversarial networks-based photovoltaic uncertainty in a smart home energy management system including battery storage devices[J]. Energy, 2024, 306: 132412.
[12] 胡石峰, 朱瑞金, 唐波. 含有梯度惩罚的WGAN光伏功率场景生成方法[J]. 电力系统及其自动化学报, 2022, 34(11): 109-115.
Hu S F, Zhu R J, Tang B.PV power scenario generation method based on Wasserstein generative adversarial networks with gradient penalty[J]. Proceedings of the CSU-EPSA, 2022, 34(11): 109-115.
[13] Song Z Y, Huang Y H, Xie H B, et al.Generation method of multi-regional photovoltaic output scenarios-set using conditional generative adversarial networks[J]. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2023, 13(3): 861-870.
[14] 王成磊, 匡熠, 濮永现, 等. 基于改进生成对抗网络场景生成的长-短期储能优化配置[J]. 高电压技术, 2024, 50(3): 1131-1144.
Wang C L, Kuang Y, Pu Y X, et al.Optimal configuration of long-term and short-term energy storage based on improved generative adversarial network scenarios[J]. High Voltage Engineering, 2024, 50(3): 1131-1144.
[15] 程卓, 许仪勋, 李泽霜. 基于改进生成对抗网络生成风光场景的微电网多时间尺度优化调度策略研究[J]. 现代电力, 2024, 41(4): 642-651.
Cheng Z, Xu Y X, Li Z S.Research on multi-time scale optimization scheduling strategy of microgrid based on improved generative adversarial network for wind and PV power scenario generation[J]. Modern Electric Power, 2024, 41(4): 642-651.
[16] Wang C L, Liu C M, Zhou X L, et al.Flexibility-based expansion planning of active distribution networks considering optimal operation of multi-community integrated energy systems[J]. Energy, 2024, 307: 132601.
[17] Chen X Q, Dong W, Yang Q.Robust optimal capacity planning of grid-connected microgrid considering energy management under multi-dimensional uncertainties[J]. Applied Energy, 2022, 323: 119642.
[18] Fu X Q, Zhang C Y, Zhang X R, et al.A novel GAN architecture reconstructed using Bi-LSTM and style transfer for PV temporal dynamics simulation[J]. IEEE Transactions on Sustainable Energy, 2024, 15(4): 2826-2829.
[19] Yuan R, Wang B, Sun Y Q, et al.Conditional style-based generative adversarial networks for renewable scenario generation[J]. IEEE Transactions on Power Systems, 2023, 38(2): 1281-1296.
[20] Liang J K, Tang W Y.Sequence generative adversarial networks for wind power scenario generation[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(1): 110-118.
[21] 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): 63-72.
Huang Y H, Sun Y N, Li C, 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): 63-72.
[22] Ye L, Peng Y S, Li Y L, et al.A novel informer-time-series generative adversarial networks for day-ahead scenario generation of wind power[J]. Applied Energy, 2024, 364: 123182.
[23] Hanif M F, Mi J.Harnessing AI for solar energy: Emergence of transformer models[J]. Applied Energy, 2024, 369: 123541.
[24] Jiang Y F, Chang S Y, Wang Z Y.TransGAN: two pure transformers can make one strong GAN, and that can scale up[J]. Advances in Neural Information Processing Systems, 2021, 34: 14745-14758.
[25] Zhang X, Jiao W T, Wang B, et al.CT-GAN: a conditional Generative Adversarial Network of transformer architecture for text-to-image[J]. Signal Processing: Image Communication, 2023, 115: 116959.
[26] Aslam N, Kolekar M H.TransGANomaly: transformer based generative adversarial network for video anomaly detection[J]. Journal of Visual Communication and Image Representation, 2024, 100: 104108.
[27] 张帅, 刘文霞, 万海洋, 等. 基于改进条件生成对抗网络的可控场景生成方法[J]. 电力自动化设备, 2024, 44(6): 9-17.
Zhang S, Liu W X, Wan H Y, et al.Controllable scenario generation method based on improved conditional generative adversarial network[J]. Electric Power Automation Equipment, 2024, 44(6): 9-17.
[28] 徐武, 范鑫豪, 沈智方, 等. 多尺度特征提取的Transformer短期风电功率预测[J]. 太阳能学报, 2025, 46(2): 640-648.
Xu W, Fan X H, Shen Z F, et al.Short-term wind power prediction using transformer with multi-scale feature extraction[J]. Acta Energiae Solaris sinica, 2025, 46(2): 640-648.
[29] 黄玲, 李林霞, 程瑜, 等. 基于SAM-WGAN-GP的短期风电功率预测[J]. 太阳能学报, 2023, 44(4): 180-188.
Huang L, Li L X, Cheng Y, et al.Short-term wind power prediction based on SAM-WGAN-GP[J]. Acta Energiae Solaris Sinica, 2023, 44(4): 180-188.
[30] 高帆, 包道日娜, 狄彦强, 等. 考虑风光出力动态相关性的场景生成方法[J]. 太阳能学报, 2024, 45(8): 256-264.
Gao F, Bao D R N, Di Y Q, et al. Scene generation method considering dynamic correlation of wind and photovoltaic outputs[J]. Acta Energiae Solaris Sinica, 2024, 45(8): 256-264.