O presente trabalho teve por objetivo otimizar e avaliar dois métodos de estimativa da radiação global diária para diferentes localidades brasileiras. Os métodos de Bristow e Campbell (BC) e Hargreaves e Samani (H) foram otimizados tomando como base 32 séries históricas de dados meteorológicos em diferentes localidades do Brasil. O desempenho de cada modelo foi avaliado por meio de índices estatísticos de precisão (r2), acurácia (d) e raiz do erro médio quadrático (RMSE). Após a otimização, o melhor desempenho foi dado pelo modelo BC, em todas as regiões do Brasil, com RMSE médio de 3,74 MJ m²- d-1. A otimização dos parâmetros contribuiu para diminuição do RMSE em 10% para H e 25% para BC. Encontraram-se valores médios de Kt para método H de 0,152 para a região Norte, 0,173 Nordeste, 0,145 Centro Oeste, 0,163 Sul e 0,152 Sudeste. Para os parâmetros A, B e C do método BC obteve-se, respectivamente, para cada região 0,619, 0,026 e 1,845 para a região Norte, 0,694, 0,074 e 1,489 Nordeste, 0,635, 0,029 e 1,697 Centro Oeste, 0,671, 0,044 e 1,580 Sul e 0,702, 0,025 e 1,747 Sudeste.

With the increasing cost associated with energy consumption, climate change and the greater awareness of the population to issues related to energy and environmental efficiency, energy conservation in buildings has been encouraged, along with the development of several solutions based on a more sustainable construction. Building cooling is the most challenging issue in the Mediterranean climate. The roof is one of the main elements of the building's opaque envelope, where the choice of materials and the implementation of appropriate passive technologies determine the thermal performance of a building. The present work aims to assess the impact of natural ventilation of a roof cavity on the thermal environment of a dwelling house under Mediterranean summer conditions. An experimental study was developed in a small-scale prototype of a typical dwelling house, comprising a ceramic tile roof with vented eaves and insulated sub-tile panels according to the construction solution of the Humbelino Monteiro SA company. The thermal performance of this roof solution was assessed under real climatic conditions based on continuous measurements of the air velocity inside the air gap, the temperature of the air and the surface temperature of all roof layers. Weather conditions were also monitored continuously. Connected with the heat transfer mechanisms, the obtained temperature and air velocity profiles data were analysed and discussed.

Calibration and evaluation of models for estimating of global solar radiation for Brazil Global solar radiation (RG) is one of the most important weather variables for understanding the biophysical processes in agricultural support tools. Currently, it can be measured by different low cost sensors. In Brazil, weather stations just recently start to record RG values. In the absence of long term observed data, models for estimating RG are needed, and the objective of this study was to analyse two Bristow and Campbell (1984) (BC) and Hargreaves and Samani (1982) (H) models for different regions in Brazil against a reference database of 32 places collected the National Institute of Space Research (INPE) through the coefficient of determination (R²), correlation coefficient (r), Willmott concordance index (d), modelling efficiency (E), and root mean square error (RMSE). They were also statistically optimized based on an iterative approach. Using the original parameters, the H model presented the best performance for all Brazilian regions, with values with RMSE of 4.24 MJ.m-²d-1 for a North region, 4.55 MJ.m-²d-1 for the Northeast Region, 4.39 MJ.m-²d-1 for a Midwest region, 4.61 MJ.m-²d-1 for a South region and 4.21 MJ.m-²d-1 for a Southeast region. After the optimization process, the best performance was given by the BC model for all Brazilian regions, with RMSE of 3.44 MJ.m-²d-1 for a North Region, 3.70 MJ.m-²d-1 for a Northeast region, 3.62 MJ.m-²d-1 for a Midwest region, 4.43 MJ.m-²d-1 for a South region and 3.50 MJ.m-²d-1 for a Southeast region. After the parameter optimization, mean values for KT for H model were 0.152 °C-0.5 for the North region, 0.173°C-0.5 for the Northeast region, 0.145°C-0.5 for the Midwest region, 0.163 °C-0.5 for the South region and 0.152°C-0.5 for the Southeast region. For the BC´s dimensionless parameters "A", "B" and "C" it was found the following values, respectively: 0.619, 0.026 and 1.845 for the North region, 0.694, 0.074 and 1.489 for the Northeast region, 0.635, 0.029 and 1.697 for the Midwest region, 0.671, 0.044 and 1.580 for the South region, and 0.702, 0.025 and 1.747 for the Southeast region.