MCDM framework: empowering sustainable building design and retrofitting through data-driven decision-making

The MCDM (Multi-Criteria Decision-Making) framework developed by METU and R2M aims to improve the decision-making process in building design and retrofitting by integrating artificial intelligence decision-making with expressive graphics and visualization. This approach considers various building solutions to foster energy-positive buildings, balancing energy efficiency and renewable energy generation.

By utilizing data from energy simulations and parametric analysis of building solutions, the MCDM framework predicts energy consumption and helps stakeholders in selecting optimal materials and technologies in buildings. This approach not only supports sustainable building practices but also informs users, equips urban planners, researchers, and construction professionals with actionable insights for environmentally friendly and comfortable buildings.

MCDM Framework

The aim of the MCDM framework is to provide a structured and systematic approach to evaluate, compare, and prioritize multiple solution options, improving the process of building design and retrofitting. It seeks to identify the best-performing construction sets via ML model developed by METU, facilitate the connection of these construction sets to market-available products in T2.4, calculate the financial and environmental impacts of the chosen products within the framework of identified KPIs in T2.5, and visualize the findings to aid the user in the decision-making process with a tool developed by R2M.

The MCDM framework starts with an extensive data generation process to identify various active and passive solutions that improve building energy performance. Using energy simulations, different combinations of active solutions (i.e. heat pumps and solar photovoltaics) and passive solutions (i.e. external surface constructions) are performed.

Machine Learning Model

Machine learning-based predictive models for each pilot developed by METU is trained using the previously generated datasets, which will help determine which passive system combinations of construction sets yield lower ideal loads. As each pilot building and its climate is different, the ML algorithm will suggest different parameter sets that will minimize both heating and cooling thermal loads. These findings serve as a basis for selecting market-available products, directly guiding the choice of materials for building retrofits in following tasks.

Identification of Passive and Active Systems, Key Performance Indicators and Visualization

The output of the ML algorithm informs T2.4 in identifying market-available products that closely match the performance metrics of the chosen solutions. The product data, combined with calculated information on environmental and financial impacts within T2.5—evaluated through KPIs such as cost, operational carbon emissions, and embodied carbon—are used to assess the feasibility and impacts of the solutions. Additionally, the outputs of Task 2.5 feed directly into Task 2.6, where these KPIs are utilized to visualize the impacts and their corresponding solution set. These results are visualized through a generated web report, which presents complex data in an accessible format. Using this report, users can analyze various retrofit options, evaluate their impact on energy usage and carbon emissions, and compare cost-benefit analyses.

The visualization report tool, or web report, is generated after all the different scenarios are simulated and the relevant KPIs are calculated and compiled into a report template. This report generation process is being developed with available open-source libraries, like python, and further languages for the web such as Javascript, CSS and HTML. In this way and after simulations are carried out, the user could see the simulation data and a summary of the results and select the best scenario for their application.

Pilot Projects and Stakeholder Involvement

The MCDM framework is currently being implemented across five pilot projects located in Luxembourg, Turkey, Hungary, the Netherlands, and Spain. Each pilot represents a unique building typology, expanding the algorithm’s applicability across different construction styles and regional climates. By working with varied building types and scales, the algorithm can develop solutions suitable for a broader range of buildings, from residential to commercial buildings, making it adaptable to diverse climates and typologies.

Moreover, with the use of the reporting tool, stakeholders could engage with the results of the simulations and identify any modifications or new set of constructions or design parameters that could be suggested to be considered beforehand.

Long-Term Goals, Integration with Other Tools & Accessibility

Beyond supporting the immediate goals of energy efficiency and comfort goals of LEGOFIT, the MCDM framework is positioned as a long-term tool for sustainable construction. With strong extendibility potential, the algorithm can eventually integrate additional building performance metrics, such as indoor thermal quality, addressing a broader scope of sustainability and comfort concerns. Moreover, the framework has strong integration potential with tools such as BIM and digital twins, which could streamline the implementation of MCDM insights into actual building projects.