针对传统气象学天气分型的局限性,同时考虑到优化算法的场景匹配度,提出基于广义场景划分（GSS）的RIME-CNN-LSTM-Attention光伏出力组合预测模型。首先,以科学性、典型性为导向,提出基于广义场景划分指标的场景划分方法,为解决传统分类聚合方法存在的日气象相似性和耦合性问题提供解决思路,有效支撑高精度预测模型的构建;其次,结合长短期记忆网络（LSTM）时序特征提取能力以及卷积神经网络（CNN）的空间特征提取优势,引入注意力机制,构建光伏出力基础预测模型,在增强模型的关键信息捕捉能力的同时有效提升预测的整体精度;此外,考虑到强不确定性条件下光伏出力的非线性特点,引入一种基于霜冰物理现象的优化算法,有效提升优化算法对复杂场景的适应能力。仿真结果显示,基于GSS的RIME-CNN-LSTM-Attention组合预测模型能有效提高光伏出力预测的准确性,在光伏出力波动性较强的场景下展现出较高的预测精度和实用性。
Abstract
In response to the limitations of traditional meteorological weather typing and considering the scenario-matching degree of optimization algorithms, RIME algorithm convolutional neural network long short term memory attention（RIME-CNN-LSTM-Attention）photovoltaic output combination prediction model based on generalized scene segmentation（GSS） is proposed. Firstly, guided by scientific and typical principles, a generalized scenario segmentation method based on a generalized scenario segmentation indicator is introduced, providing solution ideas for solving the daily weather similarity and coupling problems existing in traditional classification and aggregation methods, effectively supporting the construction of high-precision prediction models. Secondly, a foundational photovoltaic power output prediction model is constructed by combining the temporal feature extraction capabilities of long short-term memory （LSTM） networks with the spatial feature extraction advantages of convolutional neural networks （CNN） and introducing an attention mechanism. This approach enhances the model’s ability to capture critical information and effectively improves the overall accuracy of the predictions. Additionally, considering the nonlinear characteristics of photovoltaic output under conditions of strong uncertainty, an optimization algorithm based on the physical phenomenon of frost ice is introduced, significantly enhancing the optimization algorithm’s adaptability to complex scenarios. Simulation results show that the GSS-based RIME-CNN-LSTM-Attention combination prediction model can effectively improve the accuracy of photovoltaic output predictions, demonstrating high predictive accuracy and practicality in scenarios where photovoltaic output is highly volatile.
关键词
光伏发电 /
功率预测 /
卷积神经网络 /
长短期记忆网络 /
场景划分 /
优化算法
Key words
photovoltaic power generation /
power forecasting /
convolutional neural network /
long short-term memory network /
scene segmentation /
optimization algorithm
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] KABIR E, KUMAR P, KUMAR S, et al.Solar energy: potential and future prospects[J]. Renewable and sustainable energy reviews, 2018, 82: 894-900.
[2] 李争, 张杰, 徐若思, 等. 基于相似日聚类和PCC-VMD-SSA-KELM模型的短期光伏功率预测[J]. 太阳能学报, 2024, 45(2): 460-468.
LI Z, ZHANG J, XU R S, et al.Short term photovoltaic power prediction based on similar day clustering and PCC-VMD-SSA-KELM model[J]. Acta energiae solaris sinica, 2024, 45(2): 460-468.
[3] ZHANG H, LIU Y Q, YAN J, et al.Improved deep mixture density network for regional wind power probabilistic forecasting[J]. IEEE transactions on power systems, 2020, 35(4): 2549-2560.
[4] 董存, 王铮, 白捷予, 等. 光伏发电功率超短期预测方法综述[J]. 高电压技术, 2023, 49(7): 2938-2951.
DONG C, WANG Z, BAI J Y, et al.Review of ultra-short-term forecasting methods for photovoltaic power generation[J]. High voltage engineering, 2023, 49(7): 2938-2951.
[5] HUANG X Q, LI Q, TAI Y H, et al.Hybrid deep neural model for hourly solar irradiance forecasting[J]. Renewable energy, 2021, 171: 1041-1060.
[6] LIU X H, GENG C, XIE S H, et al.Day ahead thermal-photovoltaic economic dispatch considering uncertainty of photovoltaic power generation[J]. Journal of system simulation, 2022, 34(8): 1874-1884.
[7] 王涛, 李薇, 许野, 等. 计及相似日的VMD-FE-LSTM光伏出力组合预测模型研究[J]. 太阳能学报, 2024, 45(5): 490-499.
WANG T, LI W, XU Y, et al.Study on photovoltaic power prediction of VMD-FE-LSTM considering similar days[J]. Acta energiae solaris sinica, 2024, 45(5): 490-499.
[8] 陈昌松, 段善旭, 殷进军. 基于神经网络的光伏阵列发电预测模型的设计[J]. 电工技术学报, 2009, 24(9): 153-158.
CHEN C S, DUAN S X, YIN J J.Design of photovoltaic array power forecasting model based on neutral network[J]. Transactions of China Electrotechnical Society, 2009, 24(9): 153-158.
[9] 李燕青, 杜莹莹. 基于双维度顺序填补框架与改进Kohonen天气聚类的光伏发电短期预测[J]. 电力自动化设备, 2019, 39(1): 60-65.
LI Y Q, DU Y Y.Short-term photovoltaic power forecasting based on double-dimensional sequential imputation framework and improved Kohonen clustering[J]. Electric power automation equipment, 2019, 39(1): 60-65.
[10] 叶林, 裴铭, 路朋, 等. 基于天气分型的短期光伏功率组合预测方法[J]. 电力系统自动化, 2021, 45(1): 44-54.
YE L, PEI M, LU P, et al.Combination forecasting method of short-term photovoltaic power based on weather classification[J]. Automation of electric power systems, 2021, 45(1): 44-54.
[11] 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.
[12] 田翠霞, 黄敏, 朱启兵. 基于EMD-LMD-LSSVM联合模型的逐时太阳辐照度预测[J]. 太阳能学报, 2018, 39(2): 504-512.
TIAN C X, HUANG M, ZHU Q B.Hourly solar irradiance forecast based on EMD-LMD-LSSVM joint model[J]. Acta energiae solaris sinica, 2018, 39(2): 504-512.
[13] WANG K J, QI X X, LIU H D.A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network[J]. Applied energy, 2019, 251: 113315.
[14] WANG K J, QI X X, LIU H D.Photovoltaic power forecasting based LSTM-convolutional network[J]. Energy, 2019, 189: 116225.
[15] 雷柯松, 吐松江·卡日, 伊力哈木·亚尔买买提, 等. 基于WGAN-GP和CNN-LSTM-Attention的短期光伏功率预测[J]. 电力系统保护与控制, 2023, 51(9): 108-118.
LEI K S, TUSONGJIANG K, YILIHAMUY, et al. Prediction of short-term photovoltaic power based on WGAN-GP and CNN-LSTM-Attention[J]. Power system protection and control, 2023, 51(9): 108-118.
[16] CAO B, FAN S S, ZHAO J W, et al.Large-scale many-objective deployment optimization of edge servers[J]. IEEE transactions on intelligent transportation systems, 2021, 22(6): 3841-3849.
[17] CAO B, ZHANG W Z, WANG X S, et al.A memetic algorithm based on two_Arch2 for multi-depot heterogeneous-vehicle capacitated arc routing problem[J]. Swarm and evolutionary computation, 2021, 63: 100864.
[18] LIU Y, HEIDARI A A, CAI Z N, et al.Simulated annealing-based dynamic step shuffled frog leaping algorithm: optimal performance design and feature selection[J]. Neurocomputing, 2022, 503: 325-362.
[19] ZENG N Y, WANG Z D, LIU W B, et al.A dynamic neighborhood-based switching particle swarm optimization algorithm[J]. IEEE transactions on cybernetics, 2022, 52(9): 9290-9301.
[20] CAO B, ZHAO J W, LYU Z H, et al.Diversified personalized recommendation optimization based on mobile data[J]. IEEE transactions on intelligent transportation systems, 2021, 22(4): 2133-2139.
[21] DONG R Y, CHEN H L, HEIDARI A A, et al.Boosted kernel search: framework, analysis and case studies on the economic emission dispatch problem[J]. Knowledge-based systems, 2021, 233: 107529.
[22] SU H, ZHAO D, HEIDARI A A, et al.RIME: a physics-based optimization[J]. Neurocomputing, 2023, 532: 183-214.
基金
中央高校基本科研业务费专项资金（B240205007）
PDF(1542 KB)
