Desenvolvimento de modelos matemáticos para a determinação da disponibilidade de luz natural no crepúsculo para a cidade de Belo Horizonte

• Autor(a) principal: Camila Carvalho Ferreira
• Data de Publicação: 2010
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/BUBD-A5BGFF
• topic: Medições de iluminância; Duração da noite; Crepúsculo; Disponibilidade de luz natural; Iluminação natural

This study presents an approach to the determination of the availability of daylight in the twilight to the city of Belo Horizonte, by means of the creation of multiple regression models as a support in the computation of invoicing hours of artificial lighting for power utility. To determine the availability of daylight, through the measurement of the monthly illuminance of the sky, the local climatic variables and solar geometry were considered. The data relating to climatic variables was collected by the Measuring Station Natural Light of Belo Horizonte (EMIN-BH) and by the meteorological station of PUC / CEMIG. Illuminance data was collected within one minute intervals during the period of August 2007 to November 2008 for the twilight. Global and diffuse irradiance and vertical north, south, east and west illuminances and irradiances were also measured. Additionally, photographic records of the sky were made to verify the different sky coverages. The data from global illuminance for the twilight was collected, organized and processed to extract the set of spurious data. From the data processed, models are developed for each condition of sky in Excel ® software through multiple regression, considering as the independent variable the solar altitude, the total precipitable water and brightness, and the dependent variable was the global illuminance. The models that performed better were the polynomial models of 5th degree of global illuminance on solar altitude. Analysis of the duration of the night depending on the model and sky types was made from the application of models developed for each type of typical sky for each month throughout the year. The predicted night duration was found and then faced with the measured night duration, showing satisfactory results. The average measured night duration was of 11 hours and 30 minutes while the average predicted night duration was of 11 hours and 29 minutes. In conclusion, the developed models for determining the availability of daylight in Belo Horizonte denote good adequacy for the definition of time needed for the estimation of artificial public lighting use throughout the year.