| Description: |
The accelerating urbanization shifts energy demand towards cities, which account already for ~75% of global energy consumption and ~70% of greenhouse gas emissions. Implementing effective building retrofit strategies is therefore key to reduce energy consumption and to decarbonize building operation, a priority reflected in existing policies and directives. In contrast to research efforts focused primarily on environmental and economic indicators, we introduce a building retrofit planning framework that combines socio-economic and environmental data within an AI computational pipeline for assessment, evaluation and decision-making.Our framework identifies effective building envelope retrofit solutions considering building characteristics, climate projections, and socio-economic information collected from neighborhoods and residents. Besides single-building analysis, the optimization objectives and constraints additionally capture systemic effects, enabling coordinated decision-making at neighbourhood level. Energy efficiency and savings potential are quantified through physics-based models that use publicly available geospatial databases to extract building-specific geometric information, while information about the thermal properties of materials is derived from archetype classification data. To support large-scale decision-making, datasets generated by physics-based models are used to train a surrogate model for energy prediction, facilitating efficient evaluation of multiple retrofit scenarios.To showcase the effectiveness of our building retrofit planning framework, we conduct a city-wide case study in Rotterdam. With the case study we introduce policymakers to the retrofit planning framework allowing them to designing equitable, actionable and sustainable building retrofit strategies at multiple scales. Our ultimate goal is to promote the adoption of sustainable retrofit packages at household and neighborhood levels, accelerating carbon footprint reduction in urban environments. |