Metodología para el desarrollo de modelos basados en redes Neuro-Fuzzy para la predicción de la potencia real generada a corto plazo por paneles solares ubicados en la ciudad de Loja.

This research work concerns a methodology for the development of models based on neural networks and neurofuzzy systems to predict the actual power generated by solar panels short term located in the Loja city, which aims to develop models based temporary neurofuzzy networks able to determine from radiation and temperature variables of voltage, current and power output in real operating conditions of the solar panel. This research work is distributed between several literals which are detailed below: In the literal d is a review of the literature needed to understand basic concepts and rules used for this research work. The literal e reveals the materials and methods used. The photovoltaic panel EXMORK mark used is 100W, the current and voltage values are monitored at the output of the panel, and also acquires the variables temperature and solar panel with suitable sensors for this type of solar panel. The sampling time was every minute. In the literal f is exhibited the results of this research work. This section describes the processing of data from Matlab, methodologies is performed to generate models with neural networks and ANFIS type neurofuzzy networks; procedures for applications with Matlab Builder Ex and how to deploy applications Matlab Builder Ex to run models in Excel spreadsheet also view a report of calibration models using a html document. The discussion on the results of the thesis is presented in subsection g. Formulated in the literal h and i , the research findings for the same approach that led to recommendations respectively, in order to improve the methodology for the development of neurofuzzy models performed. Keys words: Solar panel, ANFIS, ANN'S, nntool, neurofuzzy models, Matlab Builder Ex

• Autor principal: Paredes Luzón, José Luis (author)
• Formato: bachelorThesis
• Idioma: spa
• Publicado em: 2012
• Assuntos: <REDES NEURO FUZZY><PANELES SOLARES><REDES NEURONALES>
• Acesso em linha: http://dspace.unl.edu.ec/jspui/handle/123456789/12285