Performance Comparison of ANN-Based PV Generation Prediction Models Using Dropout Technique

Sung-Hyup Hong; Kwang Ho Lee

doi:10.21086/ksles.2024.10.31.5.353

All Issue

2024 Vol.31, Issue 5 Preview Page Next Page

Research Article

Performance Comparison of ANN-Based PV Generation Prediction Models Using Dropout Technique Dropout 기법을 활용한 ANN 기반 태양광 발전량 예측 모델의 성능 비교

31 October 2024. pp. 353-362

PDF XML

Abstract

This study investigates the effect of the Dropout technique on the generalizability of PV generation prediction models utilizing Artificial Neural Networks (ANNs). Using solar energy data from Daejeon, South Korea, covering 2017 to 2021, we developed three ANN-based predictive models with Dropout rates of 0%, 10%, and 20%. The Dropout method was implemented during the training phase to mitigate overfitting by randomly deactivating neurons in the network. The model with a 20% Dropout rate (Case 3) demonstrated the most consistent and reliable performance across various metrics. Although all models met the performance criteria based on CV(RMSE), those utilizing appropriate Dropout rates exhibited enhanced stability, particularly during specific timeframes. The results indicate that while Dropout effectively prevents overfitting and enhances prediction stability, variations in dropout rates do not necessarily lead to significant changes in overall model performance. This study highlights the importance of Dropout as a regularization technique within ANN frameworks and suggests further exploration of time-series analysis and other methods to improve solar power forecasting models.

Keywords

Dropout

Machine learning

Confidence interval

References

A.S.H.R.A.E. (2014). Measurement of energy, demand, and water savings. ASHRAE Guideline, 4, 1-150.

Compare the effect of different scalers on data with outliers. (2022). In Scikit-learn. Retrieved September 10, 2024 from https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html

Gensler, A., Henze, J., Sick, B., & Raabe, N. (2016). Deep learning for solar power forecasting-An approach using autoencoder and LSTM neural networks. In 2016 IEEE International Conference on Systems, Man, and Cybernetics, 2858-2865

10.1109/SMC.2016.7844673

Hinton, G. E., Srivastava, N., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. R. (2012). Improving neural networks by preventing co-adaptation of feature detectors. arXiv Preprint arXiv:1207.0580.

IEA. (2023). Renewable report.

Lee, Y, T., Kim, D, H., Sin, W, S., Kim, C,K., Kim, H, G., & Han, S. W. (2021). A Comparison of machine learning models in photovoltaic power generation forecasting. Journal of the Korean Institute of Industrial Engineers, 47(5), 444-458.

10.7232/JKIIE.2021.47.5.444

Lin, L., Wang, M., Cheng, H., Liu, R., & Chen, F. (2023). OptionNet: A multiscale residual deep learning model with confidence interval to predict option price. The Journal of Finance and Data Science, 9, 100105.

10.1016/j.jfds.2023.100105

Ministry of Trade, Industry and Energy (MOTIE). (2023). The 10th Basic Plan for Electricity Supply and Demand.

Nasiri, V., Hawryło, P., Janiec, P., & Socha, J. (2023). Comparing object-based and pixel-based machine learning models for tree-cutting detection with Planet Scope satellite images: Exploring model generalization. International Journal of Applied Earth Observation and Geoinformation, 125, 103555.

10.1016/j.jag.2023.103555

Park, J., Hong, S. H., Yeon, S. H., Seo, B. M., & Lee, K. H. (2023). Predictive model for solar insolation using the deep learning technique. International Journal of Energy Research, 1, 3525651.

10.1155/2023/3525651

Riffenburgh, R. H. (2012). Statistics in medicine. Academic Press.

Saxena, N., Kumar, R., Rao, Y. K., Mondloe, D. S., Dhapekar, N. K., Sharma, A., & Yadav, A. S. (2024). Hybrid KNN-SVM machine learning approach for solar power forecasting. Environmental Challenges, 14, 100838.

10.1016/j.envc.2024.100838

Son, W. B. (2023). Current status of solar power generation forecasting technology in preparation for the expansion of renewable energy deployment. Information and Communications Magazine, 40(12), 20-26.

Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. (2014). Dropout: A simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research, 15(1), 1929-1958.

Tahir, M. F., Yousaf, M. Z., Tzes, A., El Moursi, M. S., & El-Fouly, T. H. (2024). Enhanced solar photovoltaic power prediction using diverse machine learning algorithms with hyperparameter optimization. Renewable and Sustainable Energy Reviews, 200, 114581.

10.1016/j.rser.2024.114581

Tripathi, A. K., Aruna, M., Elumalai, P. V., Karthik, K., Khan, S. A., Asif, M., & Rao, K. S. (2024). Advancing solar PV panel power prediction: A comparative machine learning approach in fluctuating environmental conditions. Case Studies in Thermal Engineering, 59, 104459.

10.1016/j.csite.2024.104459

Woo, T. K., So, M. S., Kang, S. Y., & Sin, J. H. (2024). Development of a solar power generation prediction model utilizing photovoltaics and climate data. Society for Computational Design and Engineering, 29(1), 33-41.

10.7315/CDE.2024.033

Ying, X. (2019). An overview of overfitting and its solutions. In Journal of Physics: Conference Series, 1168, 022022

10.1088/1742-6596/1168/2/022022

Information

Publisher :The Korean Society of Living Environmental System
Publisher(Ko) :한국생활환경학회
Journal Title :Journal of The Korean Society of Living Environmental System
Journal Title(Ko) :한국생활환경학회
Volume : 31
No :5
Pages :353-362
Received Date : 2024-09-11
Revised Date : 2024-10-16
Accepted Date : 2024-10-31
DOI :https://doi.org/10.21086/ksles.2024.10.31.5.353

Journal of The Korean Society of Living Environmental System ISSN:1226-1289(Print) 한국생활환경학회

All Issue