Solar Irradiance Prediction Algorithm Based on WOA-TCN
School of Information Engineering, Tianjin University of Commerce, Tianjin 300134, China
| Abstract: | The prediction of solar irradiance plays a crucial role in the optimization operation of multi energy collaborative heating systems. As a key factor affecting the efficiency of the heating system, the accuracy of solar irradiance prediction is directly related to the operational efficiency and energy consumption of the heating system. However, solar irradiance has uncertain characteristics such as intermittency and nonlinearity in time and space, which poses challenges for existing machine learning prediction algorithms in hyperparameter optimization and makes it difficult to meet optimization and regulation requirements. To address this issue, we propose a solar irradiance prediction model based on Whale Algorithm Optimization Time Convolutional Network (WOA-TCN). This model fully utilizes the parallel processing capability of TCN time convolutional networks and the temporal modeling function of recurrent neural networks. The hyperparameters of the TCN are optimized by a swarm intelligence optimization algorithm that simulates the behavior of whales, which improves the prediction accuracy of the model, enables the temporal and spatial evolution patterns of solar irradiance to be captured more accurately, and also provides strong technical support for the optimal operation of the multi-energy cooperative heating system. The prediction indexes of CNN, RNN and other algorithms are also compared through experiments, and the results show that the prediction accuracy of the solar irradiance prediction model based on WOA-TCN is higher. |
| Keywords: | Solar Irradiance Prediction; Whale Optimization Algorithm; Time Convolutional Network; Deep Learning |
| DOI: | 10.57237/j.jest.2024.04.001 |
View:
Downloads:
Received: 21 June 2024
Accepted: 13 July 2024
Published Online: 18 July 2024
Funding
| 1. | 大学生创新创业训练计划资助项目 (202310069029) |
References
| [1] | 陈怡然, 李进喜, 陈思洁, 等 新能源环境下的太阳辐射预测系统 [J]. 计算机仿真, 2023, 40(05): 140-4. |
| [2] | 李津, 史加荣, 张琰妮, 等. 基于最大信息系数的短期太阳辐射协同估计 [J]. 太阳能学报, 2023, 44(09): 286-94. |
| [3] | 普智勇, 夏攀, 张璐, 等 机器学习与统计方法在太阳能预报中的比较性分析 [J]. 太阳能学报, 2023, 44(07): 162-7. |
| [4] | 张晓英, 常正云, 罗童, 等. 基于聚类EEMD -PCA -LSTM与误差补偿的光热电站短期太阳直接法向辐射预测 [J]. 电气工程学报: 1-9. |
| [5] | 尹青, 张华, 何金海. 近48年华东地区地面太阳总辐射变化特征和影响因子分析 [J]. 大气与环境光学学报, 2011, 6(01): 37-46. |
| [6] | 达选芳, 李照荣, 王小勇, 等. 有云条件下太阳辐射短临预报订正技术研究 [J]. 干旱气象, 2021, 39(06): 1006-16. |
| [7] | 鲁玉军, 周世豪, 胡小勇. 基于BP神经网络和小波神经网络的太阳辐射强度预测 [J]. 软件工程, 2023, 26(01): 5-8+4. |
| [8] | 陈垒, 李杨露西. 基于实际天空图像和CNN卷积神经网络的太阳辐射强度预测研究 [J]. 建筑科学, 2022, 38(08): 177-83. |
| [9] | 臧海祥, 张越, 程礼临, 等. 基于ICEEMDAN-LSTM和残差注意力的短期太阳辐照度预测 [J]. 太阳能学报, 2023, 44(12): 175-81. |
| [10] | 鹿晨东, 许英朝, 张帆, et al. 基于生成对抗网络的太阳辐照度预测模型 [J]. 厦门理工学院学报, 2023, 31(05): 17-24. |
| [11] | 官松泽, 唐钰本, 蔡争, et al. 基于Kmeans++-Bi-LSTM的太阳辐照度超短期预测 [J]. 太阳能学报, 2023, 44(12): 170-4. |
| [12] | LIU J, HUANG X Q, LI Q, et al. Hourly stepwise forecasting for solar irradiance using integrated hybrid models CNN-LSTM-MLP combined with error correction and VMD [J]. Energy Conv Manag, 2023, 280: 21. |
| [13] | JALALI S M J, AHMADIAN S, KAVOUSI-FARD A, et al. Automated Deep CNN-LSTM Architecture Design for Solar Irradiance Forecasting [J]. IEEE Trans Syst Man Cybern -Syst, 2022, 52(1): 54-65. |
| [14] | GU Q, LI S J, GONG W Y, et al. L-SHADE with parameter decomposition for photovoltaic modules parameter identification under different temperature and irradiance [J]. Appl Soft Comput, 2023, 143: 15. |
| [15] | WANG P, XUE B, LIANG J, et al. Feature clustering-Assisted feature selection with differential evolution [J]. Pattern Recognit, 2023, 140: 13. |
| [16] | MIRJALILI S, LEWIS A. The Whale Optimization Algorithm [J]. Adv Eng Softw, 2016, 95(C): 51–67. |
Join the Editorial Board
We invite active, qualified and high profile scientists and researchers to join as Editorial Board Members.
Join UsBecome a Reviewer
Scholars with a strong interest in reviewing are invited to join the reviewer panel to ensure the quality of the research to be published.
Join Us