Photovoltaic power forecasting model based on probabilistic TCN-Transformer

doi:10.3969/j.issn.2097-0706.2024.11.002

Abstract

Abstract:

A short-term PV power prediction method based on a temporal convolutional network （TCN） and a Transformer structure is proposed. Firstly， the main factors affecting PV power generation，such as wind speed， rainfall， light intensity and cloudiness， are analysed. Then， TCN is used to extract the global spatial features of the sequence， and Transformer is used to extract the temporal features of long-term dependencies in the sequence， so that a TCN-Transformer composite model with a high prediction precision is applied to PV power deterministic and probabilistic prediction. Simulation analyses are performed on the dataset from DKASC（Australia）， and the results show that the improved TCN-Transformer model exhibits excellent prediction performance under different weather conditions， improving the short-term prediction accuracy on PV power.

Key words: photovoltaic power generation forecast, temporal convolutional network, Transformer, deterministic prediction, probabilistic prediction

CLC Number:

TM615

SHENG Ruixiang, ZHANG Xiaoyu. Photovoltaic power forecasting model based on probabilistic TCN-Transformer[J]. Integrated Intelligent Energy, 2024, 46(11): 10-18.

Figures/Tables 10

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Table 1

Fig.6

Table 2

Evaluation indicators of models

模型	MAE/kW	MSE/(kW)²	RMSE/kW	$R 2$	MAPE/%
CNN	0.818	3.432	1.853	0.943	56.72
TCN	1.093	3.792	1.868	0.899	32.66
LSTM	0.576	2.138	1.462	0.943	25.73
CNN-LSTM	0.547	2.132	1.460	0.941	17.85
Transformer	0.529	2.165	1.471	0.942	16.55
TCN-Transformer	0.447	1.814	1.347	0.969	15.02

Table 2

Table 3

Indicators for evaluating different weather

模型	天气	MAE/kW	MSE/(kW)²	RMSE/kW	$R 2$	MAPE/%
CNN	阴天	1.078	4.774	2.185	0.915	60.57
CNN	晴天	0.912	3.587	1.894	0.935	56.87
TCN	阴天	1.209	4.530	2.128	0.906	35.42
TCN	晴天	0.993	3.531	1.879	0.920	30.21
LSTM	阴天	0.759	2.554	1.598	0.917	27.47
LSTM	晴天	0.633	1.266	1.125	0.976	25.58
Transformer	阴天	0.668	2.452	1.566	0.921	23.48
Transformer	晴天	0.485	1.173	1.083	0.979	16.05
TCN-Transformer	阴天	0.732	1.889	1.374	0.931	19.98
TCN-Transformer	晴天	0.447	1.014	1.006	0.980	15.02

Table 3

Fig.7

References 28

[1]	姚芳, 杨晓娜, 葛磊蛟, 等. 风-光-氢能源系统容量优化配置研究[J]. 综合智慧能源, 2022, 44(5):56-63. doi: 10.3969/j.issn.2097-0706.2022.05.006
	YAO Fang, YANG Xiaona, GE Leijiao, et al. Optimization of capacity allocation scheme for wind-solar-hydrogen energy system[J]. Integrated Intelligent Energy, 2022, 44(5): 56-63. doi: 10.3969/j.issn.2097-0706.2022.05.006
[2]	蔡源, 吴浩, 唐丹. 光伏发电功率预测方法综述[J]. 四川电力技术, 2024, 47(2):25-31.
	CAI Yuan, WU Hao, TANG Dan. Reviews of photovoltaic power prediction methods[J]. Sichuan Electric Power Technology, 2024, 47(2):25-31.
[3]	KUMARI P, TOSHNIWAL D. Deep learning models for solar irradiance forecasting: A comprehensive review[J]. Journal of Cleaner Production, 2021, 318: 128566.
[4]	倪超, 王聪, 朱婷婷, 等. 基于CNN-Bi-LSTM的太阳辐照度超短期预测[J]. 太阳能学报, 2022, 43(3):197-202. doi: 10.19912/j.0254-0096.tynxb.2020-0581
	NI Chao, WANG Cong, ZHU Tingting, et al. Super-short-term forecast of solar irradiance based on CNN-Bi-LSTM[J]. Acta Energiae Solaris Sinica, 2022, 43(3):197-202. doi: 10.19912/j.0254-0096.tynxb.2020-0581
[5]	AHMED R, SREERAM V, MISHRA Y, et al. A review and evaluation of the state-of-the-art in PV solar power forecasting: Techniques and optimization[J]. Renewable and Sustainable Energy Reviews, 2020, 124: 109792.
[6]	杨帆, 申亚, 李东东, 等. 基于GA-GNNM的极地光伏发电功率预测方法[J]. 太阳能学报, 2022, 43(4):167-174. doi: 10.19912/j.0254-0096.tynxb.2020-0768
	YANG Fan, SHEN Ya, LI Dongdong, et al. Polar photovoltaic power forecasting method based on GA-GNNM[J]. Acta Energiae Solaris Sinica, 2022, 43(4):167-174. doi: 10.19912/j.0254-0096.tynxb.2020-0768
[7]	LIU C C, LI M. YU Y J, et al. A review of multitemporal and multispatial scales photovoltaic forecasting methods[J]. IEEE Access, 2022, 10: 35073-35093.
[8]	吴峰, 王飞, 顾康慧, 等. 基于MEEMD-ARIMA模型的波浪能发电系统输出功率预测[J]. 电力系统自动化, 2021, 45(1):65-70.
	WU Feng, WANG Fei, GU Kanghui, et al. Output power prediction of wave energy generation system based on modified ensemble empirical mode decomposition-autoregressive integrated moving average model[J]. Automation of Electric Power Systems, 2021, 45(1):65-70.
[9]	丁坤, 刘增泉, 张经炜, 等. 基于图像奇异值分解的局部遮挡光伏阵列输出特性建模研究[J]. 综合智慧能源, 2023, 45(2): 53-60. doi: 10.3969/j.issn.2097-0706.2023.02.007
	DING Kun, LIU Zengquan, ZHANG Jingwei, et al. Modeling for output characteristics of the partially occluded photovoltaic array based on image singular value decomposition[J]. Integrated Intelligent Energy, 2023, 45(2):53-60. doi: 10.3969/j.issn.2097-0706.2023.02.007
[10]	缪月森, 夏红军, 黄宁洁, 等. 基于Informer的负荷及光伏出力系数预测[J]. 综合智慧能源, 2024, 46(4): 60-67. doi: 10.3969/j.issn.2097-0706.2024.04.008
	MIAO Yuesen, XIA Hongjun, HUANG Ningjie, et al. Prediction on loads and photovoltaic output coefficients based on Informer[J]. Integrated Intelligent Energy, 2024, 46(4): 60-67. doi: 10.3969/j.issn.2097-0706.2024.04.008
[11]	LINDBERG O, LINGFORS D, ARNQVIST J, et al. Day-ahead probabilistic forecasting at a co-located wind and solar power park in Sweden: trading and forecast verification[J]. Advances in Applied Energy, 2023, 9: 100120.
[12]	万灿, 崔文康, 宋永华. 新能源电力系统概率预测:基本概念与数学原理[J]. 中国电机工程学报, 2021, 41(19):6493-6509.
	WAN Can, CUI Wenkang, SONG Yonghua. Probabilistic forecasting for power systems with renewable energy sources: basic concepts and mathematical principles[J]. Proceedings of the CSEE, 2021, 41(19): 6493-6509.
[13]	GOLESTANEH F, PINSON P, GOOI H B. Very Short-term nonparametric probabilistic forecasting of renewable energy generation—With application to solar energy[J]. IEEE Transactions on Power Systems, 2016, 31(5): 3850-3863.
[14]	HUANG Q, WEI S Y. Improved quantile convolutional neural network with two-stage training for daily-ahead probabilistic forecasting of photovoltaic power[J]. Energy Conversion and Management, 2020, 220: 113085.
[15]	GU B, SHEN H Q, LEI X H, et al. Forecasting and uncertainty analysis of day-ahead photovoltaic power using a novel forecasting method[J]. Applied Energy, 2021, 299: 117291.
[16]	ZHANG X Y, WATKINS C, KUENZEL S. Multi-quantile recurrent neural network for feeder-level probabilistic energy disaggregation considering roof-top solar energy[J]. Engineering Applications of Artificial Intelligence, 2022, 110:104707.
[17]	吴永洪, 张智斌. 基于贝叶斯优化的CNN-GRU短期电力负荷预测[J]. 现代电子技术, 2023, 46(20):125-129.
	WU Yonghong, ZHANG Zhibin. CNN-GRU short-term power load forecasting based on Bayesian optimization[J]. Modern Electronics Technique, 2023, 46(20):125-129.
[18]	LIMOUNI T, YAAGOUBI R, BOUZIANE K, et al. Accurate one step and multistep forecasting of very short-term PV power using LSTM-TCN model[J]. Renewable Energy, 2023, 205:1010-1024.
[19]	ZHANG M Y, HAN Y, WANG C Y et al. Ultra-short-term photovoltaic power prediction based on similar day clustering and temporal convolutional network with bidirectional long short-term memory model: A case study using DKASC data[J]. Applied Energy, 2024, 375:124085.
[20]	庞昊, 高金峰, 杜耀恒. 基于时间卷积网络分位数回归的短期负荷概率密度预测方法[J]. 电网技术, 2020, 44(4):1343-1350.
	PANG Hao, GAO Jinfeng,DU Yaoheng. A short-term load probability density prediction based on quantile regression of time convolution network[J]. Power System Technology, 2020, 44 (4): 1343-1350.
[21]	陈禹帆, 温蜜, 张凯, 等. 基于相似日匹配及TCN-Attention的短期光伏出力预测[J]. 电测与仪表, 2022, 59(10):108-116.
	CHEN Yufan, WEN Mi, ZHANG Kai, et al. Short-term photovoltaic output forecasting based on similar day matching and TCN-Attention[J]. Electrical Measurement & Instrumentation, 2022, 59(10): 108- 116.
[22]	XIANG X L, LI X Y, ZHANG Y L, et al. A short-term forecasting method for photovoltaic power generation based on the TCN-ECANet-GRU hybrid model[J]. Scientific Reports, 2024, 14(1):6744. doi: 10.1038/s41598-024-56751-6 pmid: 38509109
[23]	LIN Y, KOPRINSKA I, RANA M. Temporal convolutional neural networks for solar power forecasting[C]// Proceedings of 2020 International Joint Conference on Neural Networks (IJCNN), IEEE, 2020: 1-8.
[24]	崔京港, 王芳, 叶泽甫, 等. 基于QD和因果注意力TCN的光伏功率区间预测[J]. 太阳能学报, 2024, 45(3): 488-495.
	CUI Jinggang, WANG Fang, YE Zefu, et al. Photovoltaic power interval prediction based on QD and causal attention TCN[J]. Acta Energiae Solaris Sinica, 2024, 45(3): 488-495.
[25]	KIM J, OBREGON J, PARK H, et al. Multi-step photovoltaic power forecasting using transformer and recurrent neural networks[J]. Renewable and Sustainable Energy Reviews, 2024, 200: 114479.
[26]	TAO K J, ZHAO J H, TAO Y, et al. Operational day-ahead photovoltaic power forecasting based on transformer variant[J]. Applied Energy, 2024, 373: 123825.
[27]	王腾飞, 杨力, 孙龙, 等. 基于C-Transformer的光伏负荷预测方法[J/OL]. 微电子学与计算机,1-9(2024-07-22)[2024-08-01]. http://kns.cnki.net/kcms/detail/61.1123.TN.20240719.1609.006.html.
	WANG Tengfei, YANG Li, SUN Long, et al. Transformer based photovoltaic load prediction method[J/OL]. Microelectronics & Computer,1-9(2024-07-22)[2024-08-01]. http://kns.cnki.net/kcms/detail/61.1123.TN.20240719.1609.006.html.
[28]	黄莉, 甘恒玉, 刘兴举, 等. 基于Transformer编码器的超短期光伏发电功率预测[J]. 智慧电力, 2024, 52(5):16-22,59.
	HUANG Li, GAN Hengyu, LIU Xingju, et al. Ultra-short-term photovoltaic power generation prediction based on trans-formerencoder[J]. Smart Power, 2024, 52(5):16-22,59.

参数	数值
输入通道数	96
输出通道数	36
隐藏层数	80
LSTM层数	1
注意力头数	3
失活率	0.3
批量大小	80
学习率	0.000 2