Passive Solar Building Design
Using Genetic Programming

by Mohammad Mahdi Oraei Gholami
Supervisor: Brian Ross

This research uses genetic programming to design passive solar buildings. Passive solar building design is the process of designing a building while considering sunlight exposure for receiving heat in winter and rejecting heat in summer. To evaluate energy costs, the EnergyPlus system is used, which uses a simulation to consider heating and cooling costs with respect to a number of parameters, for example, building size, window placement, materials, geographic location, weather factors, time of day, and time of year. Genetic programming also considers other design characteristics such as window placement, building shape, size, and complexity. GP uses a split grammar for generating 3D structures. To examine the impact of environment on designs, experiments are run on five different geographic locations. Also, both single floor models and multi-floor models are examined in this research. Results show that solutions are consistent with respect to materials, sizes, and appearance, and satisfied problem constraints in all instances. Experiments also show that smaller buildings with few or no windows, and no skylights, are the most energy efficient.



Images copyright (c) 2013 Mahdi Oraei.

Back up: