Enhancing Seismic Risk Assessment in Munich's Geothermal Energy Sector: Data-driven Model Updating for Building Vibrations under Induced Seismicity

Project Description

This project is embedded in the Geothermal Alliance Bavaria, specifically the subproject ‘Seismic Risk Minimization’. It enhances seismic risk assessment in Munich’s geothermal energy sector by advancing the prediction of building vibrations induced by geothermal micro‑earthquakes. It combines region‑specific ground‑motion prediction equations with dense in‑building sensor data collected at multiple floor levels near active geothermal power plants. Statistical analysis of recorded seismic data is complemented by physics‑based simulations to address data scarcity at low magnitudes and short distances. Data‑driven and Bayesian surrogate models are developed to update and replace detailed structural models, explicitly accounting for uncertainty in ground motions and structural parameters. By distinguishing horizontal and vertical vibration components and emphasizing short‑period spectral content critical for serviceability and human comfort, the project enables improved vibration prediction, uncertainty‑aware risk assessment, and the foundation for future warning systems. The results support safe, reliable geothermal operation and informed decision‑making for urban environments.

Results

• Development of new semi‑empirical ground motion prediction equations tailored to geothermal‑induced micro‑earthquakes in Southern Germany
• Integration of recorded seismic data with physics‑based simulations to address sparse observations at small magnitudes and distances
• Separate prediction of horizontal and vertical ground‑motion components, including peak values and response spectra
• Identification of elevated short‑period spectral amplitudes (<0.1 s) relevant for building serviceability and vibration assessment
• Demonstration of practical application through serviceability verification of a residential building near a geothermal site
• Improved basis for seismic risk assessment, mitigation strategies, and public reassurance in geothermal energy projects

Follow-up

In the next step, the developed ground‑motion and building response models could be extended by integrating longer‑term monitoring data, additional geothermal sites, and real‑time operational parameters such as injection rates and temperatures. This would enable calibration of early‑warning systems that dynamically link geothermal operation to vibration risk. Further research could expand Bayesian surrogate models to multi‑building and district‑scale risk assessment, incorporate soil–structure interaction in greater detail, and support Monte‑Carlo‑based urban resilience studies for enhanced geothermal systems in Bavaria and beyond.