OPTIMIZATION OF RESIDENTIAL BUILDING HEIGHT COMBINATION BASED ON PHOTOVOLTAIC POWER GENERATION PREDICTION
This study focuses on residential communities and begins by using Ladybug to simulate photovoltaic generation on rooftops and walls of buildings within the community.Subsequently,a machine learning model is trained to predict the total photovoltaic generation of the entire community with different height combinations within 0.5 seconds.Finally,a genetic algorithm is utilized to optimize the height combinations of the buildings in the community with the objective of maximizing the total photovoltaic generation.The findings indicate that a reasonable combination of building heights can effectively enhance the photovoltaic potential of the community.Specifically,when the building height control surface tilts towards the west-southwest direction and exhibits a concave shape centered on the southwest side,the parameterized community in Changsha achieves the highest annual photovoltaic generation of 838297 kWh.For regions across the country with varying climatic conditions and latitudes,the method described in this study can be used to calculate optimal height combinations for residential communities.
solar energyphotovoltaic powermachine learninggenetic algorithmsarchitectural design
NETL
NSTL
万方数据
