Design and Implementation of a Traffic Light System for Geothermal Stimulation in Finland
Abstract
St1 Deep Heat is developing a geothermal doublet to deliver deep geothermal heat to local district heat networks. As part of the project, two circa 6.5 km deep wells are intended to be drilled as a geothermal doublet in the City of Espoo, just west of Helsinki, Finland. The first well was stimulated in June to July 2018 to improve the rock permeability in contact with the well.
The stimulation took place in a densely populated area with multiple sensitive receptors. The City of Espoo's buildings department therefore required that a seismic Traffic Light System (TLS) be developed and approved before the start of well stimulation activities. The TLS thresholds were based on a combination of estimated seismic magnitudes and measured ground motion in order to ensure events were related to an induced seismic event, and not from other sources. A peak ground velocity (PGV) of 1 mm/s associated with a ML ≥ 1 event triggered an Amber alert, while a PGV of 7.5 mm/s associated with a ML ≥ 2.1 event triggered a Red alert. Specific thresholds based on PGV and peak ground acceleration (PGA) were gathered for sensitive receptors and related to earthquake magnitudes in a probabilistic way. To address data gaps in the surface network, TLS exceedances solely based on magnitude were also adopted. A ML ≥ 1.2 event alone triggered an Amber alert, while the ML 2.1 threshold alone was maintained for a Red alert. The implementation of the TLS relied on two seismic monitoring networks. A surface monitoring network was composed of seventeen 1Hz 3-component geophones located at the surface and in the basement of sensitive receptors in order to quantify PGV and PGA. A satellite network was composed of twelve 4.5Hz 3-component gimballed geophones installed in boreholes at depth ranging from 240 to 1,200m to estimate magnitudes and locations of induced events. From a regulatory point of view, the absence of existing local data prevented the design of a TLS solely based on forward-looking models and the TLS in Otaniemi therefore relied on conservative thresholds and associated hazard mitigation measures. The results from this project indicate that forward-looking models would have overestimated the probability of a TLS Red alert, given the deficit of ML ≥ 1.5 events compared to the magnitude-frequency distribution of ML < 1.5 events.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.H31C..03A
- Keywords:
-
- 1822 Geomechanics;
- HYDROLOGYDE: 1847 Modeling;
- HYDROLOGYDE: 1869 Stochastic hydrology;
- HYDROLOGYDE: 1873 Uncertainty assessment;
- HYDROLOGY