Implementación de un modelo de predicción de la demanda eléctrica mediante redes neuronales artificiales
The objective of this research is to implement a prediction model of electrical demand through artificial neural networks developed by a computational tool with the MATLAB program, using real data of electrical demand based on the power of feeder C of Substation 37 Santa Rosa belonging to to the nei...
Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | bachelorThesis |
| Sprache: | spa |
| Veröffentlicht: |
2022
|
| Schlagworte: | |
| Online Zugang: | http://repositorio.utc.edu.ec/handle/27000/9333 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Tag hinzufügen | Zusammenfassung: | The objective of this research is to implement a prediction model of electrical demand through artificial neural networks developed by a computational tool with the MATLAB program, using real data of electrical demand based on the power of feeder C of Substation 37 Santa Rosa belonging to to the neighborhood of San José de Cutuglahua. Carrying out a type of descriptive, diagnostic and applied research, with a bibliographic, inductive - deductive method, and the application of techniques and instruments such as observation, simulation and programming. Within the programming part, the program has a graphical interface (APP DESIGNER) that consists of three modules for its operation: Electricity Demand, Properties of the neural network and Predictions, with particular characteristics that allowed the correct operation, as well as a screen graph that allows you to see the days required; Within this program, MATLAB neural network libraries were implemented, for time series predictions, the NARNET network was used, using the Levenberg-Marquart training algorithm through 5 neurons with 70% of the data entered, 15% of data for validation and 15% of test data, obtaining a percentage error of 4.06%, concluding that the demand forecast is optimal and reliable.. |
|---|