Green Infrastructure to reduce cooling loads and heat stress in Mediterranean Climates. A building simulation and machine learning approach

Climate change impact on cities and urban warming due to anthropogenic effects are urgent problems to be solved. Among the most beneficious strategies to reduce those impacts we can account the development of green infrastructures in cities, a kind of intervention that assure both mitigation of global warming by reducing greenhouse gases emissions , and adaptation to warmer urban environments. This work presents a building simulation and machine learning methodology to estimate the energy and comfort -related benefits that can be obtained by using a green infrastructure to shadow buildings' façades and roofs. We used previously developed simulation models to test the energy saving s provided by different types of trees planted to produce shadows on buildings. Then, we tested different algorithms to predict using a machine learning approach the saving that can be obtained in different buildings -trees contexts for the cities of Catania, Rome, Santiago de Chile and Viña del Mar. Results show that the saving obtained is in the range 5 - 60%, mainly depending on the number of façade shadowed and on the specie of trees; and the prediction accuracy of machine learning process is over 90% for a binary classification (energy saving > 15% or <15%).