Photovoltaic Power Prediction with Optimized Transformer Integrating Pyramid Attention Module and Temporal Convolutional Network

doi:10.19678/j.issn.1000-3428.0069489

Abstract

Abstract:

Accurate photovoltaic power prediction is crucial for enhancing grid stability and improving energy utilization efficiency. To address the limitations of existing methods, which struggle to simultaneously consider both long-term dependencies and short-term variation patterns of photovoltaic power, this study proposes a novel photovoltaic power prediction method named Solarformer. This method integrates a Pyramid Attention Module (PAM) with a Temporal Convolutional Network (TCN) to optimize the Transformer architecture. First, multiple feature selection mechanisms are employed to screen the input features, to enhance the model′s ability to characterize photovoltaic data features. Second, a coarse-grained construction module and PAM are utilized to optimize the Transformer encoder, capturing the long-term temporal dependency features of photovoltaic power at multiple scales. Third, a constraint mechanism based on the sunrise-sunset effect of photovoltaic power and the TCN are employed to optimize the Transformer decoder, strengthening the model′s ability to capture short-term variation features of photovoltaic power and better model its short-term variation patterns. Experimental results on the Sanyo dataset from Australia demonstrate that Solarformer can effectively improve photovoltaic power forecasting accuracy. Compared with the DLinear model, it reduces the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Symmetric Mean Absolute Percentage Error (SMAPE) by approximately 7.45%, 6.99%, and 14.10%, respectively.

Key words: photovoltaic power prediction, Transformer model, Pyramidal Attention Module (PAM), binding mechanism, Temporal Convolutional Network (TCN

摘要：

准确的光伏功率预测对于提高电网稳定性和用电效率至关重要。针对现有研究难以同时考虑光伏功率长期依赖性和短期变化模式的缺陷，提出一种金字塔注意力模块(PAM)结合时间卷积网络(TCN)优化Transformer的光伏功率预测方法Solarformer。基于多种特征选择机制筛选输入特征，增强对光伏数据特征的表征能力；利用粗粒度构造模块和PAM优化Transformer编码器，在多尺度上捕获光伏功率的长期时间依赖特征；利用光伏功率日出日落效应约束机制和TCN优化Transformer解码器，增强光伏功率的短期变化特征，以更好地捕捉其短期变化模式。在澳大利亚Sanyo数据集上进行实验，结果表明，Solarformer能够有效提高光伏功率的预测精度，相比DLinear模型，其均方根误差(RMSE)、平均绝对误差(MAE)和对称平均绝对百分比误差(SMAPE)分别降低了约7.45%、6.99%和14.10%。

关键词: 光伏功率预测, Transformer模型, 金字塔注意力模块, 约束机制, 时间卷积网络

ZHANG Hong, LI Feng, MA Yanhong, JI Wenxuan, ZHENG Qipeng. Photovoltaic Power Prediction with Optimized Transformer Integrating Pyramid Attention Module and Temporal Convolutional Network[J]. Computer Engineering, 2025, 51(10): 140-149.

张红, 李峰, 马彦宏, 姬文宣, 郑启鹏. PAM结合TCN优化Transformer的光伏功率预测研究[J]. 计算机工程, 2025, 51(10): 140-149.

/ Recommend / Download Citations

URL: https://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0069489

https://www.ecice06.com/EN/Y2025/V51/I10/140

Figures/Tables 9

Fig.1 Solarformer model structure

Fig.2 Coarser-scale construction module

Fig.3 PAM structure

Fig.4 Decoder structure

Fig.5 Importance of features

Fig.6 Performance comparison of different models on Solibro dataset

Fig.7 Results of ablation experiment

References 26

1	BÖÖK H , LINDFORS A V . Site-specific adjustment of a NWP-based photovoltaic production forecast. Solar Energy, 2020, 211, 779- 788. doi: 10.1016/j.solener.2020.10.024
2	FARA L , DIACONU A , CRACIUNESCU D , et al. Forecasting of energy production for photovoltaic systems based on ARIMA and ANN advanced models. International Journal of Photoenergy, 2021, 2021, 6777488.
3	CUI C Y , LI Z X , ZHANG J J . Building a prediction model of solar power generation based on improved grey Markov chain. International Journal of Global Energy Issues, 2022, 44(2/3): 139. doi: 10.1504/IJGEI.2022.121396
4	李春来, 张海宁, 杨立滨, 等. 基于无迹卡尔曼滤波神经网络的光伏发电预测. 重庆大学学报, 2017, 40(4): 54- 61.
	LI C L , ZHANG H N , YANG L B , et al. Photovoltaic power generation prediction based on unscented Kalman filter neural network. Journal of Chongqing University, 2017, 40(4): 54- 61.
5	ALSHAFEEY M , CSÁKI C . Evaluating neural network and linear regression photovoltaic power forecasting models based on different input methods. Energy Reports, 2021, 7, 7601- 7614. doi: 10.1016/j.egyr.2021.10.125
6	ZAZOUM B . Solar photovoltaic power prediction using different machine learning methods. Energy Reports, 2022, 8, 19- 25.
7	GUPTA P , SINGH R . PV power forecasting based on data-driven models: a review. International Journal of Sustainable Engineering, 2021, 14(6): 1733- 1755. doi: 10.1080/19397038.2021.1986590
8	MEENAL R , BINU D , RAMYA K C , et al. Weather forecasting for renewable energy system: a review. Archives of Computational Methods in Engineering, 2022, 29(5): 2875- 2891. doi: 10.1007/s11831-021-09695-3
9	SINGLA P , DUHAN M , SAROHA S . A comprehensive review and analysis of solar forecasting techniques. Frontiers in Energy, 2022, 16(2): 187- 223. doi: 10.1007/s11708-021-0722-7
10	葛浩然, 夏宇, 邹文进, 等. 基于RF-XGBoost的光伏发电功率预测. 电气自动化, 2022, 44(5): 12- 15.
	GE H R , XIA Y , ZOU W J , et al. Photovoltaic power generation prediction based on RF-XGBoost. Electrical Automation, 2022, 44(5): 12- 15.
11	SAHU R K , SHAW B , NAYAK J R , et al. Short/medium term solar power forecasting of Chhattisgarh state of India using modified TLBO optimized ELM. Engineering Science and Technology, an International Journal, 2021, 24(5): 1180- 1200. doi: 10.1016/j.jestch.2021.02.016
12	SHARMA J , SONI S , PALIWAL P , et al. A novel long term solar photovoltaic power forecasting approach using LSTM with Nadam optimizer: a case study of India. Energy Science & Engineering, 2022, 10(8): 2909- 2929.
13	ZHANG Y , KONG L Q . Photovoltaic power prediction based on hybrid modeling of neural network and stochastic differential equation. ISA Transactions, 2022, 128, 181- 206. doi: 10.1016/j.isatra.2021.11.008
14	赵博超, 马嘉骏, 崔磊, 等. 基于改进VMD-XGBoost-BiLSTM组合模型的光伏发电异常检测. 计算机工程, 2024, 50(3): 306- 316. doi: 10.19678/j.issn.1000-3428.0067583
	ZHAO B C , MA J J , CUI L , et al. Anomaly detection for photovoltaic based on improved VMD-XGBoost-BiLSTM combination model. Computer Engineering, 2024, 50(3): 306- 316. doi: 10.19678/j.issn.1000-3428.0067583
15	TIAN F Y , FAN X X , WANG R T , et al. A power forecasting method for ultra-short-term photovoltaic power generation using Transformer model. Mathematical Problems in Engineering, 2022, 2022, 9421400.
16	ZHOU Y , LI Y Z , WANG D J , et al. A multi-step ahead global solar radiation prediction method using an attention-based Transformer model with an interpretable mechanism. International Journal of Hydrogen Energy, 2023, 48(40): 15317- 15330. doi: 10.1016/j.ijhydene.2023.01.068
17	WU H X, XU J H, WANG J M, et al. Autoformer: decomposition transformers with auto-correlation for long-term series forecasting[EB/OL]. [2023-10-05]. https://arxiv.org/abs/2106.13008v5.
18	ZHOU H Y , ZHANG S H , PENG J Q , et al. Informer: beyond efficient Transformer for long sequence time-series forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(12): 11106- 11115. doi: 10.1609/aaai.v35i12.17325
19	JIANG Y Q , GAO T L , DAI Y X , et al. Very short-term residential load forecasting based on deep-autoformer. Applied Energy, 2022, 328, 120120. doi: 10.1016/j.apenergy.2022.120120
20	CAO Y S , LIU G , LUO D H , et al. Multi-timescale photovoltaic power forecasting using an improved stacking ensemble algorithm based LSTM-Informer model. Energy, 2023, 283, 128669. doi: 10.1016/j.energy.2023.128669
21	LIMOUNI T , YAAGOUBI R , BOUZIANE K , et al. Accurate one step and multistep forecasting of very short-term PV power using LSTM-TCN model. Renewable Energy, 2023, 205, 1010- 1024. doi: 10.1016/j.renene.2023.01.118
22	WANG N , ZHAO X L . Time-series prediction based on double pyramid bidirectional feature fusion mechanism. IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences, 2023, 106(6): 886- 895.
23	蔡华锋, 唐艺豪. 基于Pyraformer网络的短期电力负荷预测. 武汉大学学报(工学版), 2023, 56(9): 1105- 1113.
	CAI H F , TANG Y H . Short-term electricity load forecasting based on Pyraformer network. Journal of Wuhan University (Engineering Edition), 2023, 56(9): 1105- 1113.
24	LUO X , ZHANG D X , ZHU X . Deep learning based forecasting of photovoltaic power generation by incorporating domain knowledge. Energy, 2021, 225, 120240. doi: 10.1016/j.energy.2021.120240
25	吴珺玥, 赵二刚, 郭增良, 等. 基于Spearman系数和TCN的光伏出力超短期多步预测. 太阳能学报, 2023, 44(9): 180- 186.
	WU J Y , ZHAO E G , GUO Z L , et al. Ultra-short-term multi-step prediction of photovoltaic output based on Spearman coefficient and TCN. Acta Energiae Solaris Sinica, 2023, 44(9): 180- 186.
26	ZENG A L , CHEN M X , ZHANG L , et al. Are transformers effective for time series forecasting?. Proceedings of the AAAI Conference on Artificial Intelligence, 2023, 37(9): 11121- 11128. doi: 10.1609/aaai.v37i9.26317

[1]	ZHANG Rui, ZHANG Xueying, CHEN Guijun, HUANG Lixia. Emotion Recognition in EEG Based on Granger Causality and Brain Regions Frequency Bands Transformer Model [J]. Computer Engineering, 2025, 51(6): 311-319.
[2]	LIU Zhong, TANG Hong, WANG Ningzhe, ZHU Chuanrun. Text Summarization Method Incorporating RNN and Sparse Self-Attention [J]. Computer Engineering, 2025, 51(1): 312-320.
[3]	XIAO Chaoen, LI Zifan, ZHANG Lei, WANG Jianxin, QIAN Siyuan. Differential Cryptanalysis Based on Transformer Model and Attention Mechanism [J]. Computer Engineering, 2025, 51(1): 156-163.
[4]	ZHU Kai, LI Li, ZHANG Tong, JIANG Sheng, BIE Yiming. Multi-Stage Motion Blur Image Restoration Network Based on Transformer [J]. Computer Engineering, 2024, 50(9): 276-285.
[5]	WANG Yanguo, LÜ Pengyuan, LAN Jinjiang, LIU Mingzhe, QIN Guanjun, ZHANG Shuohua, ZHOU Yu. Wind Turbine Fault Classification Method Based on Adversarial Training and Transformer [J]. Computer Engineering, 2024, 50(9): 377-384.
[6]	Hong ZHAO, Xiao WANG. Study on Lesion Segmentation of Melanoma Images Based on Swin-Transformer [J]. Computer Engineering, 2024, 50(8): 249-258.
[7]	Yazhou ZHANG, Yu HE, Lu RONG, Xiangkai WANG. Depression Detection Based on Contextual Knowledge Enhanced Transformer Network [J]. Computer Engineering, 2024, 50(8): 75-85.
[8]	Xuebing BAI, Jin CHE, Jinman WU, Yumin CHEN. Image Captioning Method Based on Transformer Visual Features Fusion [J]. Computer Engineering, 2024, 50(8): 229-238.
[9]	Ying HOU, Lin YANG, Xin HU, Shun HE, Wanying SONG, Qian ZHAO. Automatic Escalator Pedestrian Safety Detection Algorithm Based on SwinT-YOLOX Model [J]. Computer Engineering, 2024, 50(3): 277-289.
[10]	Chenmin GAN, Hong TANG, Haolan YANG, Xiaojie LIU, Jie LIU. Abstractive Text Summarization Method Incorporating Convolutional Shrinkage Gating [J]. Computer Engineering, 2024, 50(2): 98-104.
[11]	Shuaiwei LIU, Zhi LI, Guomei WANG, Li ZHANG. Adversarial Example Generation Algorithm Based on Transformer and GAN [J]. Computer Engineering, 2024, 50(2): 180-187.
[12]	SHEN Haiyun, YU Peng, WANG Haichuan. Long-term UAV Vision Tracking with Time Domain Siamese Network Fusion Transformer [J]. Computer Engineering, 2024, 50(11): 107-118.
[13]	ZHAO Jian, CUI Qian, SHI Jia, LIU Yue. Bimodal Fused Depressive Tendency Recognition Algorithm Based on Textual and Acoustic Features [J]. Computer Engineering, 2024, 50(11): 49-58.
[14]	Yaping CHI, Ziyan YUE, Yuheng LIN. Working Mode Recognition for SM4 Algorithm Based on Transformer [J]. Computer Engineering, 2023, 49(9): 109-117.
[15]	WANG Tao, LIU Chaohui, ZHENG Qingqing, HUANG Jiaxi. Multi-turn Task-oriented Dialogue Technology Based on Unidirectional Transformer and Siamese Network [J]. Computer Engineering, 2021, 47(7): 55-58,66.

Please choose a citation manager

Content to export