An approach to urban quarter design using building generative design and thermal performance optimization

摘要

Buildings thermal performance is influenced by the urban context, such as adjacent buildings shadows, block wind paths, or solar radiation reflection. For this reason, in this paper, an automated procedure is used to generate and optimize buildings' thermal performance in a closed O-shape urban quarter with ten building blocks to determine the importance of including the surroundings in estimating the buildings' thermal behavior. The overall shape of the urban quarter is pre-designed, being mainly residential, with several stores in the ground floor and limited to three levels. Each building will have four alternative designs created using a hybrid evolutionary strategy technique that generates building's floor plans according to practitioner's preferences and requirements. Then, a sequential variable optimization procedure coupled with dynamic simulation engine is used to explore the improvement potential of each solution by changing and adding several building elements. The final quarter design is determined by combining the best of all buildings' solutions from thermal performance criteria. The results demonstrate the influence of urban context in the resulting building's thermal performance. Despite the building's shape is similar in the four solutions, these have significant thermal behavior difference due to their interior organization and position in the urban quarter. A comparison analysis is carried out between all building block designs. The buildings, which have exterior walls with openings facing south, have almost half degree-hours of thermal discomfort due to the combination of large openings and shading overhangs. It is also possible to conclude that generative tools, enhanced with optimization procedures, may help practitioners in designing more energy efficient buildings.