Digital Twin for Self-Learning, Adaptive Facades (DTAF)

This research project develops an AI-driven digital twin for adaptive building façades to significantly improve energy efficiency and indoor comfort. Building façades play a critical role in regulating heat exchange, yet conventional static systems are limited in their ability to respond to dynamic environmental and user conditions.

The project introduces a self-learning façade system that combines IoT-based sensor networks with deep reinforcement learning (DRL). A digital twin continuously learns from real-time and simulated data, including weather conditions, indoor climate, and occupant behavior, enabling predictive and autonomous control strategies.

As a case study, the system is implemented in a novel, lightweight Trombe wall demonstrator produced through 3D-printing and bio-based phase change materials. The façade integrates actively controlled ventilation openings that regulate heat storage, release, and airflow.

Through long-term training and validation in a cyber-physical test environment, the system optimizes ventilation and thermal performance to reduce energy demand, CO₂ emissions, and peak loads. By enabling proactive rather than reactive control, the project advances adaptive façade technology toward scalable, intelligent building systems for a climate-neutral built environment.

Institute of Building Structures and Structural Design (ITKE), University of Stuttgart
Claudia Valverde, Fabian Eidner, Prof. Jan Knippers

Institute for Computational Design and Construction, Department of Computing in Architecture (ICD/CA)
Luisa Claus, Prof. Thomas Wortmann

Principal Investigator
Fabian Eidner

Klimaschutzstiftung Baden-Württemberg (2026 - 2027)

*More Info

Eidner, F., Dai, A., Gonzalez, E. A., Körner, A., Sahin, E. S., Knippers, J., Menges, A., & Wortmann, T. (2026). Responsive morphology: Co-design of a responsive 3D-printed facade system for lightweight buildings. In M. Makki et al. (Eds.), Humanistic computation and intelligence: Proceedings of CAADRIA 2026 (Vol. 2, pp. 81–90). CAADRIA.

*Link to paper