Publications

Computer  simulation  programs  are  currently  considered  the  main  tools  for  performing predictions of building performance. To do so, these tools use descriptions about the building and  the  environment  as  input  data,  for  example  the  city's  weather  information,  which  is inserted  in  the  programs  through  weather  files.  However,  although  the  high  level  of technological development of these programs, they also have its limitations. The simulation is a reality emulation process made up of theoretical mathematical methods, therefore, it is possible to make predictions with theses tools, but with a certain degree of uncertainty, since it  deals  with  a  reality  simplifications.  And  the  selection  of  the  methods  to  weather  files development is one of these uncertainties, so the research objectives were: (1) to develop weather file for the city of Viçosa-MG, presenting all the procedures, from the weather data collection in meteorological stations, to the application of the representative climatic year selection  methods;  (2)  to  evaluate  the  uncertainties  of  the  representative  climatic  year selection  methods,  pointing  the  most  reliable.  Weather  data,  from  November  2005  to December 2015, from the city of Viçosa-MG were collected. Those of this data with gaps and errors  were  treated,  while  the  climatic  elements  not  found  in  the  data  collection  were calculated.  In  a  second  moment,  a  statistical  characterization  of  the  city  's  weather  was carried out and, finally, differences in the results of computational simulations caused by the use of different types of climate files were analyzed. About the last step mentioned, a weather file was developed for each of the 10 years of weather data collected from Viçosa-MG, besides others developed with traditional selection methods, used nationally and internationally, and also files developed with innovative selection methods. The results of the analyzes pointed to the Multianual_3anos method as the most reliable, with special emphasis also to the TMY3, which was the traditional method with the best results.