Metodología basada en redes neuronales para caracterización de perturbaciones de tensión en sistemas de distribución con gd
This work develops a methodology for the detection and identification of voltage disturbances in distribution systems with distributed generation from the discrete Wavelet transform and a convolutional neural network. First, a characterization of the voltage sags and swells that occur in each type o...
- Autores:
-
Turizo Prieto, Sergio Ricardo
- Tipo de recurso:
- Trabajo de grado de pregrado
- Fecha de publicación:
- 2021
- Institución:
- Universidad de los Andes
- Repositorio:
- Séneca: repositorio Uniandes
- Idioma:
- spa
- OAI Identifier:
- oai:repositorio.uniandes.edu.co:1992/51520
- Acceso en línea:
- http://hdl.handle.net/1992/51520
- Palabra clave:
- Sistemas de energía eléctrica
Generación de energía eléctrica distribuida
Fallas de sistemas (Ingeniería)
Redes neurales (Computadores)
Localización de fallas eléctricas
Ingeniería
- Rights
- openAccess
- License
- https://repositorio.uniandes.edu.co/static/pdf/aceptacion_uso_es.pdf
| Summary: | This work develops a methodology for the detection and identification of voltage disturbances in distribution systems with distributed generation from the discrete Wavelet transform and a convolutional neural network. First, a characterization of the voltage sags and swells that occur in each type of fault is performed by applying the discrete Wavelet transform. Second, with the levels of energy decomposition extracted from the transform, a neural network training is performed, which will provide information on fault type. Finally, the magnitude of the voltage, the phases involved, and an approximation of the fault's origin are determined. The functional tool can reach 97.5% accuracy for the IEEE 13 Node Test Feeder case study. |
|---|