Elementos vegetais na simulação digital da luz natural: contribuição ao desenvolvimento de modelos tridimensionais virtuais

Abstract

Several factors influence the availability of daylighting such as the obstruction caused by the buildings or natural environment surroundings. Natural elements as arboreous vegetation can cause considerable obstruction resulting in changes in the amount of daylighting available at indoor environments. On the hand, vegetation can cast shadows and improve the quality of the light passing through the openings, reducing the intensity of the direct sunlight and glare from the sky. However, the vegetation is a difficult element to be incorporated in daylighting design. Regarding computer simulation, for example, all the elements involved in the reproduction of the light should be modeled in three dimensions. Ideally, variables such as height, canopy geometry, quantity, size and spacing of leaves, size of the trunk, branches and twigs, reflectance index, opacity index, among others, should be reproduced in a threedimensional model. However, the modeling is unfeasible, not only because of the complexity of all the variables involved in the process, but also because it overloads the model. Taking these aspects into considerations and based on previous studies on the subject, this research takes the daylighting intercepted by arboreous vegetation as its object of study. Accordingly, the overall goal is to characterize a method for representation and modeling of arboreous vegetation in computer simulation for daylighting. To reach this aim, the hemispherical photography method was used to determine the tree canopies gap fraction. These data helped to develop three different configurations of 3D virtual models of canopies, which were used to simulate the effect of the daylight interception by the tree. Using graphical and statistical methods, we assessed the degree of adjustment of these models with real trees. The results obtained with the MBE and RMSE statistical indicators and considering the variability of light and the scatter levels represented by the median showed that all three models tested in this study may be representative appropriate for real trees. However, it is noteworthy that, possibly, the type of representative model of each tree species is intrinsically linked to the characteristics of each tree.