The current building practice resulted from a long construction tradition when the climate was steady. As discussed in the previous section, these design strategies will become ineffective as temperatures continue to rise and heat waves will get longer and more severe. Therefore, the challenge is now determining how to design new buildings with high-performance considering the changing climate. By pursuing this objective, this research will also answer how design mitigates overheating, how to prevent oversizing of the cooling systems, and how to maximize the renewable energy systems’ integration. Therefore, the aims of the research are:
- to characterize building types and their operation according to IPCC energy use and greenhouse gas emission scenarios
- to select future weather from climate models from the IPCC latest report
- to develop a new weather ‘morphing’ technique
- to develop a radically new space allocation generative design algorithm to produce alternative buildings
- to develop a novel ensemble neural network as a surrogate method to dynamic simulation
- to develop a new statistical model that quantifies the uncertainties associated with climate predictions and building operation
- to generate future weather and buildings datasets using the developed algorithms
- to analyze the datasets using the developed statistical model
- to realize new building guidelines from the analysis results of the dataset
The project team will reach these aims in five work packages (WP). Aims (1) and (2) are met in the first work package, aims (3), (4), and (5) are concluded in the second, aims (6), (7), and (8) are completed in the third, and lastly, aim (9) is met on the fourth work package. The remaining work package is dedicated to disseminating research findings and the management of the project.
In ‘WP1 – Future scenarios‘, the building types are characterized, buildings’ operation is determined, and future climate scenarios are selected. The WP starts by identifying and choosing different locations worldwide that represent the different types of climate zones. Meanwhile, alternative building types are identified and characterized in their function, geometry, and construction. The buildings’ operation data will be from known validated models of equivalent programmatic buildings. After describing the energy use and occupancy patterns, the CESAM team will select IPCC energy use and greenhouse gas emissions scenarios. This data will be used to develop the algorithms in WP2 and assess the buildings’ robustness and resilience in WP3.
In ‘WP2 – Algorithmic development‘, we will develop three new algorithmic approaches. In the first development, the team will develop a novel ‘morphing’ technique to match climate models’ mid and long-term predictions for the needed weather data in the building thermal simulation.
The second development will be a radically new generative design approach for creating buildings. The approach will take advantage of generative adversarial networks’ latest advances in rapidly producing realistic indoor arrangements and combine them with the geometric precision of the solutions generated by the project team’s hybrid evolution strategy algorithm. The method will provide the overall building’s shape and structure for the evolutionary algorithm to allocate indoor spaces.
In the last development, an ensemble of small neural networks will be developed to fast estimate hourly indoor thermal and energy variables. This needed development relates to the fact that complex buildings’ multi-zone dynamic simulation is very time-consuming to compute. Each neural network will predict the next timestep value of a specific building phenomenon. The output of these neural networks will be the input of a single large neural network that will predict the energy and thermal performance in each thermal zone. The developed algorithms are used in WP3 to create the future weather and generated the synthetic datasets.
In ‘WP3 – Case studies’, we will develop a new statistical model. The model will identify key performance aspects of the building design, determine the interrelations between different variables, measure the design features’ significance, and quantify the uncertainties related to the climate models and buildings’ operation.
The team will generate future climate data based on two different future climate scenarios from IPCC with the resolution needed for the building thermal simulation using the developed ‘morphing’ technique. With the climate data, synthetic datasets with the buildings’ geometry, construction, and energy systems information are generated using the algorithms developed in WP2. As the time to estimate the performance using dynamic simulation grows significantly with the number of thermal zones in the building, the more complex edifices will be evaluated using the ensemble neural network. During thermal and energy evaluation, the generated future weather and sampled extreme weather events produced in WP1 will be used to define each location’s outdoor environment conditions.
After producing the synthetic datasets, the new statistical model is used to identify key design variables, explore their interrelated roles, and quantify the uncertainties associated with each climate scenario and building type operation. Key performance indicators are correlated with the geometry, construction solutions, and energy systems variables to understand the buildings’ behavior to their change. In this step, a sensitivity analysis will evaluate the influence of every input design variable. The most impactful aspects of the building design are identified in both current and future climate scenarios. Comparing those two scenarios will determine which design practices must change to guarantee the occupants’ thermal comfort. The results of the study are used to draw the building design guidelines in WP4.
In ‘WP4 – Design guidelines’, the results from WP3 are inferred into guidelines as packages of alternative design paths. For each type of building, location, and time horizon, those packages will provide recommendations on the building’s geometry, construction, HVAC, and renewable energy systems and their interrelated roles. The most impactful variables will be highlighted and explained their covariance in the design strategy. By comparing the most representative solutions of the best and worst performances, the team determines the robustness of design strategies in the uncertainty of the different climate scenarios and the resilience against extreme weather events on a scale of compliance. The results are disseminated through the project in outreach activities organized in WP5. In this sense, this project presents a synergetic value by promoting the exchange of knowledge that pushes forward other disciplines’ understanding of climate change impacts. For instance, climatologists help engineers imagine how buildings will operate in the future outdoor environment, and engineers will explain the path to architects on how to reach high-performance buildings.
‘WP5 – Dissemination, and Management’, which will run throughout the entire project, it will coordinate the project’s human, material, and financial resources. The project team will implement three orientation plans: the dissemination and communication plan (DCP), the transfer-of-knowledge plan (TKP), and the data management plan (DMP).
The outcomes are expected to improve the professionals’ quality of the design work, the building stock performance, and the nations’ ability to fight climate change, supporting the development of new policies on energy efficiency and sustainability on a national level in several world regions. In this regard, this research plan is in line with two Sustainable Development Goals in the UN 2030 Agenda.
Lastly, by understanding how future high-performance buildings are designed, this research project opens development lines for new and more efficient climate scenario-oriented technologies. This knowledge contributes to having safer and more sustainable buildings and identifies development directions of novel tools, construction solutions, and innovative materials.