Climate change presents threatening challenges to the built environment. Global warming is making traditional design practice inadequate. For example, in mild climates, as building design follows an overall passive solar heating strategy complemented with night ventilation, most studies focused on reducing the heating losses during winter. Therefore, highly insulated envelopes were proposed in Madrid, Milan and Lisbon as climate mitigation measures. These results were expected, as buildings in these regions are known for their low indoor air quality due to being poorly built. Thus, the benefits of high insulation levels were expected.
However, these studies implicitly assume that occupants will just adapt to the future climate without resorting to cooling-based air-conditioning systems. However, this assumption is questioned by recent studies that argue that occupants will purchase those systems, thus shifting from today’s winter heating paradigm towards cooling-based needs. As future climate scenarios indicate longer hours of solar availability, the electricity consumed by cooling systems may be compensated using photovoltaic and thermal collector systems. However, these systems’ implementation viability depends strongly on surroundings, building shape, and available installation surface area.
It is, therefore, essential to analyze the building’s performance holistically and on an annual basis. Whenever incorporated in the analysis, studies show that cooling needs are a significant part of the total energy consumed. Thus, highly insulated buildings will increase overheating, which was estimated for several countries. For example, retirement villages in the UK, designed to meet the net-zero energy target, expose the elderly to grave health risks from overheating. In Cyprus, the country’s annual electricity demand will rise by 6 %. In Switzerland, apartment buildings’ cooling demand will increase to 244 %. Similar concerns on overheating were also found in Belgium.
Natural ventilation and higher temperature setpoints are recommended to prevent both overheating risk and the increase in cooling energy consumption in Turkey. In the Iberian Peninsula, the use of passive measures to prevent overheating was found to be beneficial, such as direct ventilation and evaporative cooling, window shades, low infiltration, and natural ventilation. However, studies show that natural ventilation will be less effective in a temperate oceanic climate. In some United States cities, night ventilation will simply not work in several cities. Night ventilation will be only useful when combined with thermal mass to prevent daytime overheating peaks in the United Kingdom.
The number of studies that use climate change-based generated weather is still small, and the analyzed locations are dispersed worldwide. These analyze a single case study or a small number of simplified reference buildings, in which uncertainties from both climate data and building’s operation are usually overlooked. In this sense, the results may not yet be generalized in confidence to provide firm conclusions. Furthermore, the ‘one-dimensional’ objective of some of those studies presents a limited assessment. Besides, these do not explain how to design new buildings but rather explain how to improve the existing building stock.
Thus, to determine new building guidelines under climate change, this research plan aims to develop a novel statistical approach, which takes into account the future climate and building’s operation uncertainties, and a radically new performance-based generative approach, which will be based on the latest algorithmic advances in generative adversarial networks. The PI and team members have successfully used an evolutionary algorithm to created residential buildings from their indoor space arrangements. Besides producing a statistically significant number of different building types, such a generative design method creates solutions based on robust and rigorous design conditions and performance criteria. Despite being extensively tested with residential buildings (multi-story apartment buildings and single-family residential buildings), the used algorithm has not proved generative efficacy for large service buildings, such as schools and hospitals. Therefore, innovative and radically new developments are expected in this matter.