针对传统基于相似日聚类方法无法实现短时变化气象类型的低成本准确识别,从而影响光伏功率预测精度的问题,该文融合理想GHI与实测GHI的差分数据特征,利用XGBoost算法进行超短时气象场景分类计算,然后采用VMD-CNN-LSTM组合预测模型,捕捉时序数据局部和全局的特征,提高预测准确性。实验结果表明,该方法在实际应用时无需额外高成本的气象观测方式,即可实现超短时光伏发电更高精度的预测。相比传统预测算法,晴天理想天气下,30 min预测的指标R2上升3.4%,指标MSE、MAE和RMSE分别下降52.6%、33.0%和29.0%。
Abstract
To address the limitations of traditional similar-day clustering methods, which fail to achieve low-cost and accurate identification of short-term variations in weather types and thus affect photovoltaic （PV） power forecasting accuracy, this study integrates the differential features between ideal global horizontal irradiance （GHI） and measured GHI, employing the XGBoost algorithm for ultra-short-term weather scenario classification. Subsequently, a combined VMD-CNN-LSTM prediction model is adopted to capture both local and global characteristics of time-series data, thereby improving forecasting accuracy. Experimental results demonstrate that the proposed method can achieve higher accuracy in ultra-short-term PV power forecasting without requiring additional high-cost meteorological observations. Compared with traditional forecasting algorithms, under ideal sunny conditions, the 30-minute-ahead forecasting performance achieves an R² increase of 3.4%, while the MSE, MAE, and RMSE are reduced by 52.6%, 33%, and 29%, respectively.
关键词
光伏发电 /
分类分析 /
XGBoost /
功率预测
Key words
photovoltaic power generation /
classification analysis /
XGBoost /
power prediction
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] MIRZA A F, SHU Z K, USMAN M, et al.Quantile-transformed multi-attention residual framework (QT-MARF) for medium-term PV and wind power prediction[J]. Renewable energy, 2024, 220: 119604.
[2] 高寒旭, 袁祖晴, 张淑婷, 等. 基于LSTM模型的短期光伏功率预测[J]. 太阳能学报, 2024, 45(6): 376-381.
GAO H X, YUAN Z Q, ZHANG S T, et al.Short-term photovoltaic power prediction based on LSTM model[J]. Acta energiae solaris sinica, 2024, 45(6): 376-381.
[3] 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.
[4] MAYER M J.Benefits of physical and machine learning hybridization for photovoltaic power forecasting[J]. Renewable and sustainable energy reviews, 2022, 168: 112772.
[5] YAN S, HU M Q.A multi-stage planning method for distribution networks based on ARIMA with error gradient sampling for source-load prediction[J]. Sensors, 2022, 22(21): 8403.
[6] 殷豪, 李奕甸, 谢智锋, 等. 混合图神经网络和门控循环网络的短期光伏功率预测[J]. 太阳能学报, 2024, 45(3): 523-532.
YIN H, LI Y D, XIE Z F, et al.Short-term photovoltaic power prediction method based on mixed graph neural network and gated recurrent unit network[J]. Acta energiae solaris sinica, 2024, 45(3): 523-532.
[7] WANG Z, WANG Y S, CAO S L, et al.A robust spatial-temporal prediction model for photovoltaic power generation based on deep learning[J]. Computers and electrical engineering, 2023, 110: 108784.
[8] 周育才, 肖添, 谢七月, 等. 基于聚类的HPO-BILSTM光伏功率短期预测[J]. 太阳能学报, 2024, 45(4): 512-518.
ZHOU Y C, XIAO T, XIE Q Y, et al.Clustering-based HPO-BILSTM short-term prediction of PV power[J]. Acta energiae solaris sinica, 2024, 45(4): 512-518.
[9] 欧阳卫年, 赵紫昱, 陈渊睿. 自样本特征构造的1DCNN-BiLSTM网侧光伏功率预测[J]. 电力系统及其自动化学报, 2023, 36(3): 151-158.
OUYANG W N, ZHAO Z Y, CHEN Y R.1DCNN - BiLSTM method of grid-side photovoltaic power prediction with self-sampled feature construction[J]. Proceedings of the CSU-EPSA, 2023, 36(3): 151-158.
[10] JALALI S M J, AHMADIAN S, KAVOUSI-FARD A, et al. Automated deep CNN-LSTM architecture design for solar irradiance forecasting[J]. IEEE transactions on systems, man, and cybernetics: systems, 2022, 52(1): 54-65.
[11] SABRI N M, EL HASSOUNI M.Accurate photovoltaic power prediction models based on deep convolutional neural networks and gated recurrent units[J]. Energy sources, part A: recovery, utilization, and environmental effects, 2022, 44(3): 6303-6320.
[12] GUO X F, WANG X L, AO Y S, et al.Short-term photovoltaic power forecasting with adaptive stochastic configuration network ensemble[J]. WIREs data mining and knowledge discovery, 2022, 12(6): e1477.
[13] LIU H, ZHOU Y Q, LUO Q F, et al.Prediction of photovoltaic power output based on similar day analysis using RBF neural network with adaptive black widow optimization algorithm and K-means clustering[J]. Frontiers in energy research, 2022, 10: 990018.
[14] 龙小慧, 秦际赟, 张青雷, 等. 基于相似日聚类及模态分解的短期光伏发电功率组合预测研究[J]. 电网技术, 2024, 48(7): 2948-2957.
LONG X H, QIN J Y, ZHANG Q L, et al.Short-term photovoltaic power prediction study based on similar day clustering and modal decomposition[J]. Power system technology, 2024, 48(7): 2948-2957.
[15] PENG S R, GUO L J, LI B, et al.Photovoltaic power-stealing identification method based on similar-day clustering and QRLSTM interval prediction[J]. Applied sciences, 2023, 13(6): 3506.
[16] WU Z, PAN F F, LI D D, et al.Prediction of photovoltaic power by the informer model based on convolutional neural network[J]. Sustainability, 2022, 14(20): 13022.
[17] ZHOU H X, WANG J, OUYANG F L, et al.A two-stage method for ultra-short-term PV power forecasting based on data-driven[J]. IEEE access, 2023, 11: 41175-41189.
PDF(1240 KB)
