| Title | Toward calibrating an automatic detection system to monitor micro-seismic activity induced by geothermal projects in the Upper Rhine Graben |
|---|---|
| Authors | Maurer, V; Grunberg, M; Cuenot N; Richard, A. |
| Year | 2016 |
| Conference | European Geothermal Congress |
| Keywords | seismology, automatic detection, calibration, seismic monitoring, enhanced geothermal system, Alsace |
| Abstract | Designed to produce 24 MWth (170 °C, 70 l/s) with a doublet, the Rittershoffen EGS project is located 6 km eastwards of well-known Soultz-sous-Forêts EGS project, in Northern Alsace, France. Those 2 projects share a seismic network in order to monitor the induced seismic activity during drilling operations or geothermal exploitation. At Rittershoffen, two deep wells have been drilled between 2012 and 2014 to 2500 m TVD (True Vertical Depth) for targeting local normal-faults located at the interface between the clastic Triassic sediments and the top crystalline basement. Before, during and after drilling operations, induced seismicity is monitored by a series of sensors. Since 2012, the micro-seismicity of both Soultz-sous-Forêts and ECOGI geothermal projects are monitored by two joined permanent seismic networks composed of 12 surface stations in total. As some induced seismic activity was expected during the reservoir development of the first Rittershoffen well, GRT-1, the permanent seismic monitoring network has been reinforced with a temporary short-period network. In addition to this permanent network, the project management decided in June 2013 to deploy a temporary network composed of 16 real-time telemetered surface stations in order to monitor the stimulation operations of the first well GRT-1. This network was reinforced in March 2014, leading to a network composed of 31 short-period stations. All the signals were sent to the Strasbourg University where they were processed and archived by an automatic detection system (SeisComp3). This system is based on a STA/LTA method to detect induced seismicity. During the stimulation of GRT-1, 217 induced events were automatically detected, however a much larger number of events were visible. Hence, manual processing of all visible events was initiated to build an exhaustive catalogue of the seismicity induced during these operations. At this stage, about 870 events were manually picked and localised but this analysis is still ongoing. A first estimate indicates that the rate of the automatic detection system was less than 10% of the total number of visible events, leaving room for further improvements of the real-time seismic monitoring of geothermal projects. Special attention has been paid to configure the auto-detection and location modules of the automatic system in order to allow it operating at a very local scale. An intensive work on tuning detection and location parameters, based on waveforms playbacks based on the manually picked catalogue. The main changes consisted in adapting the system to work at a very local scale adapted to geothermal projects. This eventually led to increase the detection rate by a factor of two and to improve the location capabilities of the system, leading to a magnitude of completeness of the system of about 0.5. Finally, the corresponding parameters were used to replay the signal recorded during the stimulation of the first well. |