Optimization and Machine Learning in Modeling Approaches to Hybrid Energy Balance to Improve Ports Efficiency
[EN] This research presents a comprehensive review and application of hybrid energy solutions and optimization models for ports and marine environments. It introduces new methodologies, including a strategic energy management framework and a machine learning (ML) tool for predicting energy surpluses...
| Autores: | , , , , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Universitat Politècnica de València (UPV) |
| Repositorio: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia |
| Idioma: | inglés |
| OAI Identifier: | oai:riunet.upv.es:10251/230926 |
| Acceso en línea: | https://riunet.upv.es/handle/10251/230926 |
| Access Level: | acceso abierto |
| Palabra clave: | Ports HY4RES Hybrid energy systems Port efficiency Carbon neutrality Microgrid optimization 06.- Garantizar la disponibilidad y la gestión sostenible del agua y el saneamiento para todos 07.- Asegurar el acceso a energías asequibles, fiables, sostenibles y modernas para todos 17.- Fortalecer los medios de ejecución y reavivar la alianza mundial para el desarrollo sostenible |
| Sumario: | [EN] This research presents a comprehensive review and application of hybrid energy solutions and optimization models for ports and marine environments. It introduces new methodologies, including a strategic energy management framework and a machine learning (ML) tool for predicting energy surpluses and deficits. The hybrid energy module developed for the Port of Avilés is further refined to assess the performance of three tools in optimizing renewable generation and storage: the in-house Energy Management Tool (EMTv1), the HYbrid for Renewable Energy Solutions model (HY4RES), and a commercial model (Hybrid Optimization of Multiple Energy Resources, HOMER). Seven scenarios combining different renewable sources and storage technologies are evaluated. With EMTv1, Scenario 1 shows high surplus energy, while Scenario 2 achieves near grid independence using Pump-as-Turbine (PAT) storage. HY4RES analysis of Scenario 3 indicates a positive grid balance with net energy export, whereas Scenario 4 reveals the limitations of the PAT system due to low installed power. Scenario 5 incorporates a 15 kWh battery, enabling efficient energy storage and discharge, reduced grid dependency, and full coverage of demand. Using HOMER, Scenario 6 requires 546 kWh of grid energy but exports 2385 kWh, and Scenario 7 generates 3450 kWh/year, covering demand with 1834 kWh/year of surplus and a small capacity shortage (1.41 kWh/year. An AI-based ML approach is applied to five scenarios with available numerical data, accurately predicting energy balances and optimizing grid interactions. A neural network time-series (NNTS) model trained on an average-year dataset achieves high accuracy (R² = 0.9253–0.9695). A second case study with an 80% increase in demand confirms the robustness of the model, with Scenario 3 showing the highest mean squared error (0.0166 kWh), Scenario 2 the lowest R² (0.9289), and Scenario 5 the highest R² (0.9693) during validation. Overall, the study demonstrates that AI-driven forecasting is a valuable tool for ports to optimize energy management, reduce grid dependence, and improve operational efficiency. |
|---|