李昂,周雷金,闫群民,贺海育

• 1:陕西理工大学电气工程学院

摘要(Abstract)：

准确预测太阳辐射强度（SI）对电力调度和光伏选址至关重要。随着高性能计算机和大容量存储设备的发展，基于数据驱动的深度学习模型在SI预测领域获得广泛关注，然而，深度学习模型的“黑箱”特性在物理解释性上的缺失，限制了其在特定场合的应用可信度。为了在保持预测精度和模型结构不变、不增加计算复杂度的前提下，提升模型的可解释性，构建了一个基于长短时记忆（LSTM）神经网络的模型。其性能比传统神经网络提高了8.07%，并展示出更优的离群点处理能力。通过采用分层相关传播（LRP）算法，从时间和空间2个维度对影响模型输出的因素进行了评分，增强了模型的可解释性。研究结果表明：该模型在确保性能的前提下，具备良好的可解释性，其中历史辐射强度、时间相关特征（如时日周月）、太阳高度角信息（如日出和日落时刻）、云层覆盖度、辐射时长、温度和露点温度等因素是影响太阳辐射强度预测的主要因素。

关键词(KeyWords)： 太阳辐射强度预测;深度学习;可解释性;LRP算法;LSTM

Abstract:

Keywords:

基金项目(Foundation): 陕西省教育厅重点科学研究计划项目（20JS018）;陕西省教育厅专项科研计划（5JK1125）~~

作者(Author): 李昂,周雷金,闫群民,贺海育

DOI: 10.19666/j.rlfd.202312188

参考文献(References)：

• [1] ACIKGOZ H. A novel approach based on integration of convolutional neural networks and deep feature selection for short-term solar radiation forecasting[J]. Applied Energy, 2022, 305:117912.
• [2] NARVAEZ G, GIRALDO L F, BRESSAN M, et al.Machine learning for site-adaptation and solar radiation forecasting[J]. Renewable Energy, 2021, 167:333-342.
• [3] ESPINOSA A R, BRESSAN M, GIRALDO L F. Failure signature classification in solar photovoltaic plants using RGB images and convolutional neural networks[J].Renewable Energy, 2020, 162:249-256.
• [4] GAO B, HUANG X, SHI J, et al. Hourly forecasting of solar irradiance based on CEEMDAN and multi-strategy CNN-LSTM neural networks[J]. Renewable Energy,2020, 162:1665-1683.
• [5] ZHOU Y, LIU Y, WANG D, et al. A review on global solar radiation prediction with machine learning models in a comprehensive perspective[J]. Energy Conversion and Management, 2021, 235:113960.
• [6] KHATIB T, MOHAMED A, MAHMOUD M, et al.Modeling of daily solar energy on a horizontal surface for five main sites in Malaysia[J]. International Journal of Green Energy, 2011, 8(8):795-819.
• [7] BAMISILE O, OLUWASANMI A, EJIYI C, et al.Comparison of machine learning and deep learning algorithms for hourly global/diffuse solar radiation predictions[J]. International Journal of Energy Research,2022, 46(8):10052-10073.
• [8] BEHRANG M A, ASSAREH E, GHANBARZADEH A,et al. The potential of different artificial neural network(ANN)techniques in daily global solar radiation modeling based on meteorological data[J]. Solar Energy,2010, 84(8):1468-1480.
• [9] CAO J C, CAO S H. Study of forecasting solar irradiance using neural networks with preprocessing sample data by wavelet analysis[J]. Energy, 2006,31(15):3435-3445.
• [10] AHMAD A, ANDERSON T N, LIE T T. Hourly global solar irradiation forecasting for New Zealand[J]. Solar Energy, 2015, 122:1398-1408.
• [11] SHARMA V, YANG D, WALSH W, et al. Short term solar irradiance forecasting using a mixed wavelet neural network[J]. Renewable Energy, 2016, 90:481-492.
• [12] PANG Z, NIU F, O’NEILL Z. Solar radiation prediction using recurrent neural network and artificial neural network:a case study with comparisons[J]. Renewable Energy, 2020, 156:279-289.
• [13] ASLAM M, SEUNG K H, LEE S J, et al. Long-term solar radiation forecasting using a deep learning approach-GRUs[C]//2019 IEEE 8th International Conference on Advanced Power System Automation and Protection(APAP). IEEE, 2019:917-920.
• [14] SRIVASTAVA S, LESSMANN S. A comparative study of LSTM neural networks in forecasting day-ahead global horizontal irradiance with satellite data[J]. Solar Energy, 2018, 162:232-247.
• [15] OBIORA C N, ALI A, HASAN A N. Forecasting hourly solar irradiance using long short-term memory(LSTM)network[C]//2020 11th International Renewable Energy Congress(IREC). IEEE, 2020:1-6.
• [16] GUNNARSSON B R, VANDEN BROUCKE S,BAESENS B, et al. Deep learning for credit scoring:do or don’t?[J]. European Journal of Operational Research,2021, 295(1):292-305.
• [17] MIGNAN A, BROCCARDO M. Neural network applications in earthquake prediction(1994—2019):Meta-analytic and statistical insights on their limitations[J]. Seismological Research Letters, 2020,91(4):2330-2342.
• [18] SHAMSHIRBAND S, FATHI M, DEHZANGI A, et al.A review on deep learning approaches in healthcare systems:Taxonomies, challenges, and open issues[J].Journal of Biomedical Informatics, 2021, 113:103627.
• [19] DAS A, RAD P. Opportunities and challenges in explainable artificial intelligence(XAI):a survey[J/OL].arXiv:2006.11371, 2020:1-22. https://doi.org/10.48550/arXiv.2006.11371
• [20] LI A, XIAO F, ZHANG C, et al. Attention-based interpretable neural network for building cooling load prediction[J]. Applied Energy, 2021, 299:117238.
• [21] LóPEZ SANTOS M, GARCíA-SANTIAGO X,ECHEVARRíA CAMARERO F, et al. Application of temporal fusion transformer for day-ahead PV power forecasting[J]. Energies, 2022, 15(14):5232.
• [22] GAO Y, MIYATA S, AKASHI Y. Interpretable deep learning models for hourly solar radiation prediction based on graph neural network and attention[J]. Applied Energy, 2022, 321:119288.
• [23] GAO Y, MIYATA S, MATSUNAMI Y, et al.Spatio-temporal interpretable neural network for solar irradiation prediction using transformer[J]. Energy and Buildings, 2023, 297:113461.
• [24] BACH S, BINDER A, MONTAVON G, et al. On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation[J]. PloS One, 2015,10(7):e0130140.
• [25] YANG Y, TRESP V, WUNDERLE M, et al. Explaining therapy predictions with layer-wise relevance propagation in neural networks[C]//2018 IEEE International Conference on Healthcare Informatics(ICHI). IEEE, 2018:152-162.
• [26] GREZMAK J, ZHANG J, WANG P, et al. Interpretable convolutional neural network through layer-wise relevance propagation for machine fault diagnosis[J].IEEE Sensors Journal, 2019, 20(6):3172-3181.
• [27] KIM D, HO C H, PARK I, et al. Untangling the contribution of input parameters to an artificial intelligence PM2.5 forecast model using the layer-wise relevance propagation method[J]. Atmospheric Environment, 2022, 276:119034.
• [28]王琛,王颖,郑涛,等.基于ResNet-LSTM网络和注意力机制的综合能源系统多元负荷预测[J].电工技术学报, 2022, 37(7):1789-1799.WANG Chen, WANG Ying, ZHENG Tao, et al.Multi-energy load forecasting in integrated energy system based on resNet-LSTM network and attention mechanism[J]. Transactions of China Electrotechnical Society, 2022, 37(7):1789-1799.
• [29] CHO K, VAN MERRI?NBOER B, GULCEHRE C,et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[J/OL]. arXiv preprint arXiv:1406.1078, 2014:1-10.
• [30]殷孝雎,都治良,卲国策,等.基于GWO-SVR的风电机组柔性塔架振动建模与仿真分析[J].太阳能学报,2023, 44(8):404-411.YIN Xiaoju, DU Zhiliang, SHAO Guoce, et al. Vibration modeling and simulation analysis of flexible tower of wind turbines based on GWO-SVR[J]. Acta Energiae Solaris Sinica, 2023, 44(8):404-411.
• [31]雷江龙,余娟,向明旭,等.基于深度神经网络的数据驱动潮流计算异常误差改进策略[J].电力系统自动化,2022, 46(1):76-84.LEI Jianglong, YU Juan, XIANG Mingxu, et al.Improvement strategy for abnormal error of data-driven power flow calculation based on deep neural network[J].Automation of Electric Power Systems, 2022, 46(1):76-84.
• [32] CHICCO D, WARRENS M J, JURMAN G. The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation[J]. Peer J Computer Science, 2021, 7:e623.
• [33] GAO Y, LI P, YANG H, et al. A solar radiation intelligent forecasting framework based on feature selection and multivariable fuzzy time series[J]. Engineering Applications of Artificial Intelligence, 2023, 126:106986.
• [34] CHAKCHAK J, CETIN N S. Investigating the impact of weather parameters selection on the prediction of solar radiation under different genera of cloud cover:a case-study in a subtropical location[J]. Measurement,2021, 176:109159.
• [35] BRISTOW K L, CAMPBELL G S. On the relationship between incoming solar radiation and daily maximum and minimum temperature[J]. Agricultural and Forest Meteorology, 1984, 31(2):159-166.
• [36] SHRESTHA A K, THAPA A, GAUTAM H. Solar radiation, air temperature, relative humidity, and dew point study:Damak, Jhapa, Nepal[J]. International Journal of Photoenergy, 2019, 2019:1-7.
• [37] MEDVIGY D, BEAULIEU C. Trends in daily solar radiation and precipitation coefficients of variation since1984[J]. Journal of Climate, 2012, 25(4):1330-1339.