Evaluation of ground motion and optimum seismic monitoring at an inner-city deep geothermal power plant

Keil, Sabrina (2023): Evaluation of ground motion and optimum seismic monitoring at an inner-city deep geothermal power plant. Dissertation, LMU München: Fakultät für Geowissenschaften

Abstract

Experience shows that even non-pressure-stimulated geothermal power plants can generate induced seismicity even in low seismic hazard settings. Due to the shallow hypocentral depth of a few kilometers, already relatively small earthquakes can lead to ground shaking, which is felt by the population. As most of the geothermal power plants are located within or close to densely inhabited areas, monitoring and estimating the maximum ground motion and its spatial distribution are of crucial importance for the authorities and the public. However, this is challenging as in most cases the monitoring network is weak, due to high noise levels and limited station numbers, and only few information about local site effects are available. Therefore, several open questions for the development of proper risk governance strategies remain: How does the shallow subsurface structure look like and is there the risk of seismic wave amplification? How large is the expected ground motion and what areas are affected? How can the seismic monitoring within urban environments be improved? These questions are going to be answered by the example of Munich, Germany, where Eu-rope’s largest inner-city geothermal project was carried out with a total of six deep wells, which increases the risk of induced seismicity in the area. To answer the question of local site effects a microzonation study is conducted in Munich´s inner city. The common approach of array measurements is challenging in urban environments due to the space requirements. Therefore, the recently developed approach of single-station six-component (6C) measurements is applied, combining three translational and three rotational motions. This new method is much simpler in terms of logistics and therefore allows an easier and faster estimation of the local velocity structure. Another problem that is encountered in microzonation studies is the existing ambiguity of the inversion results. Since conventional inversion methods suffer from different limitations, a machine learning algorithm is trained, which selects the appropriate number of subsurface layers and returns a complete probability distribution of the S-wave velocity structure. The shallow velocity structure is important, as it can amplify seismic waves. However, the maximum induced ground motion is often not recorded because the station network is sparse at most geothermal sites. A dense network is neither feasible nor required by (Bavarian) law, which impedes a spatial interpretation of the ground motion after an earthquake. Because of that, large uncertainties remain in the determination of affected areas according to the German mining laws. This is a major problem when it comes to compensating for damage claims brought forward by local population and companies. I show that 3D numerical simulations are able to close observational gaps and can be used to estimate the maximum ground motion and its spatial distribution. This way public authorities can make fast and precise decisions in case of damaging events. Not only the sparse number of stations, but also the high noise levels pose a big problem for microseismic monitoring within urban environments, which make it difficult to reach the required magnitude threshold and location accuracy. Therefore, a network optimization method is applied, which calculates the optimum number and location of seismic stations even in environments with heterogeneous noise conditions. The dissertation results have implications for future inner-city geothermal projects, as they will facilitate the seismic risk assessment during the planning stage, the seismic monitoring during operation and the evaluation of shaking effects after an event.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 550 Geowissenschaften
• Fakultäten: Fakultät für Geowissenschaften
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 11. Juli 2023
• 1. Bericht­erstatter:in: Igel, Heiner