Sensitivity analysis of urban morphology factors regarding solar energy potential of buildings in a Brazilian tropical context
CHALLENGE - Planning the urban morphology of the built environment is a key issue in shifting to a climate adapted urban environment. This paper addresses an important threefold energy challenge of tropical cities: the major potential of harnessing solar energy as renewable resource for local electricity production and the energy-saving paradox of reducing the undesirable solar heat gains in buildings while providing satisfactory levels of daylight. It aims at measuring the effect size of urban form factors regarding these energy goals.
SOLUTION - Firstly, a set of known energy-related urban morphological parameters is identified within the specific literature. A statistical hypothesis is established, which consists of assuming that all variables considered in this study are relevant to the response-variables under investigation: solar energy potential, daylight potential and thermal radiation emission from building surfaces. Through a Design of Experiments (DOE) methodology performed in modeFRONTIER, all urban morphological parameters are brought together on a simplified urban block base model, composing a set of diverse and contrasted urban scenarios. For this purpose, a fractional factorial design methodology is applied, which allows statistically representative and non-redundant scenarios to be elaborated. The urban scenarios are then assessed regarding the solar irradiation and daylight availability on their surfaces using a Simplified Radiosity Algorithm (SRA) provided by CitySim software, with an application for the tropical context of the city of Maceió, located in the northeast of Brazil.
BENEFITS - Results indicate a significant impact of the aspect ratio, the distance between buildings and the surface equivalent albedo. Establishing high values of street aspect ratio may cut solar irradiation on roofs by 130 kW h/m2 year, while increasing the plot ratio may only yield 26 kW h/m2 year. The results also point out important first order interaction effects between certain variables.