Pronóstico de generación de una mini central hidroeléctrica mediante aprendizaje automático utilizando software de código abierto.
In this research Project, it was present the challange to see the energy generator in the Catacazon Mini Hydroelectric power plant by learning technique automatic and open source software. the main objective was to compare forescast models to determine the accuracy about generation predictions. In t...
Bewaard in:
| Hoofdauteur: | |
|---|---|
| Andere auteurs: | |
| Formaat: | bachelorThesis |
| Taal: | spa |
| Gepubliceerd in: |
2024
|
| Onderwerpen: | |
| Online toegang: | http://repositorio.utc.edu.ec/handle/27000/11984 |
| Tags: |
Voeg label toe
Geen labels, Wees de eerste die dit record labelt!
|
Voeg label toe | Samenvatting: | In this research Project, it was present the challange to see the energy generator in the Catacazon Mini Hydroelectric power plant by learning technique automatic and open source software. the main objective was to compare forescast models to determine the accuracy about generation predictions. In this research Project, it was taken 87647 data collected over 5 years which were divide into 80% data in order to of model training and 20% data for testing. This datas were used for the porpuse to add learning technique automatic, between Simple Linear Regression, GRU Closed Recurrent Units and LSTM Neural Networks, which were added in the open source software Python. The results were showed that the applied of these models give useful predictions and focus to inform decision, helping significantly to the planification and efficient gestion about energy sources. Besides, it was taken the evaluation of the prediction results in differents temporal horizon. It was focused in the GRU Closed Recurrent Units Model showing a great close to the real Power curve. During this evaluation proces, it was analized several metrics as result a mistake of Absolute Percentage (MAPE) of 1.42% daily case, 1.61% weekly case and 1.82% monthly case. |
|---|