Photovoltaic Power Forecast Using Deep Learning Techniques with Hyperparameters Based on Bayesian Optimization: A Case Study in the Galapagos Islands
-
Published:2023-08-09
Issue:16
Volume:15
Page:12151
-
ISSN:2071-1050
-
Container-title:Sustainability
-
language:en
-
Short-container-title:Sustainability
Author:
Guanoluisa Richard1, Arcos-Aviles Diego1ORCID, Flores-Calero Marco1ORCID, Martinez Wilmar23ORCID, Guinjoan Francesc4ORCID
Affiliation:
1. Grupo de Investigación en Propagación, Control Electrónico y Networking (PROCONET), Departamento de Eléctrica, Electrónica y Telecomunicaciones, Universidad de las Fuerzas Armadas ESPE, Av. General Rumiñahui s/n, Sangolquí 171-5-231B, Ecuador 2. Department of Electrical Engineering, ESAT, KU Leuven, Agoralaan gebouw B bus 8, 3590 Diepenbeek, Belgium 3. EnergyVille-Thor Park 8310, 3600 Genk, Belgium 4. Department of Electronics Engineering, Escuela Técnica Superior de Ingenieros de Telecomunicación de Barcelona, Universitat Politècnica de Catalunya, 08034 Barcelona, Spain
Abstract
Hydropower systems are the basis of electricity power generation in Ecuador. However, some isolated areas in the Amazon and Galapagos Islands are not connected to the National Interconnected System. Therefore, isolated generation systems based on renewable energy sources (RES) emerge as a solution to increase electricity coverage in these areas. An extraordinary case occurs in the Galapagos Islands due to their biodiversity in flora and fauna, where the primary energy source comes from fossil fuels despite their significant amount of solar resources. Therefore, RES use, especially photovoltaic (PV) and wind power, is essential to cover the required load demand without negatively affecting the islands’ biodiversity. In this regard, the design and installation planning of PV systems require perfect knowledge of the amount of energy available at a given location, where power forecasting plays a fundamental role. Therefore, this paper presents the design and comparison of different deep learning techniques: long-short-term memory (LSTM), LSTM Projected, Bidirectional LSTM, Gated Recurrent Units, Convolutional Neural Networks, and hybrid models to forecast photovoltaic power generation in the Galapagos Islands of Ecuador. The proposed approach uses an optimized hyperparameter-based Bayesian optimization algorithm to reduce the forecast error and training time. The results demonstrate the accurate performance of all the methods by achieving a low-error short-term prediction, an excellent correlation of over 99%, and minimizing the training time.
Funder
Research Group of Propagation, Electronic Control, and Networking (PROCONET) of Universidad de las Fuerzas Armadas ESPE and KU Leuven VLIR-UOS and the Belgian Development Cooperation
Subject
Management, Monitoring, Policy and Law,Renewable Energy, Sustainability and the Environment,Geography, Planning and Development,Building and Construction
Reference47 articles.
1. An adaptive deep learning framework for day-ahead forecasting of photovoltaic power generation;Luo;Sustain. Energy Technol. Assess.,2022 2. An analysis of different deep learning neural networks for intra-hour solar irradiation forecasting to compute solar photovoltaic generators’ energy production;Etxegarai;Energy Sustain. Dev.,2022 3. Forecasting intra-hour solar photovoltaic energy by assembling wavelet based time-frequency analysis with deep learning neural networks;Vadillo;Int. J. Electr. Power Energy Syst.,2022 4. Guanoluisa-Pineda, R., Ibarra, A., Arcos-Aviles, D., Martinez, W., Motoasca, E., and Guinjoan, F. (2022, January 18–21). Short-Term forecasting of photovoltaic power in an isolated area of Ecuador using deep learning techniques. Proceedings of the 11th International Conference on Renewable Energy Research and Application, Istanbul, Turkey. 5. Aribia, H.B., El-Rifaie, A.M., Tolba, M.A., Shaheen, A., Moustafa, G., Elsayed, F., and Elshahed, M. (2023). Growth Optimizer for Parameter Identification of Solar Photovoltaic Cells and Modules. Sustainability, 15.
Cited by
2 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
|
|