ALBERT

Description: 〈span〉〈div〉ABSTRACT〈/div〉Since 2013 to date more than 1000 seismic events have been recorded by the Servicio Geologico Colombiano (Colombian Geological Survey, SGC) in the municipality of Puerto Gaitán (Colombia). A total of 14 earthquakes are moment magnitude 〈span〉M〈/span〉〈sub〉w〈/sub〉 〉 4.0. The largest event ever recorded in the area occurred in November 2015 with 〈span〉M〈/span〉〈sub〉w〈/sub〉 4.8. It seems like the case of Puerto Gaitán is associated with the deep injection of coproduced wastewater from oil and gas extraction. The data presented in this work suggests a close relationship in space and time between injection operations and seismicity. An analysis of temporality between both data sets resulted in a time lag equivalent to about 218 d. For this paper, we computed the input and output energy during injection operations from 2013 to 2015 in order to estimate the fraction of total input energy that is radiated as seismic waves. Our results suggest that the seismic energy is only a small fraction of the total energy into the system. Although Puerto Gaitan is one of the places with the most significant volume of wastewater injected among the ones reported in the literature, the energy efficiency of the system is the lowest reported to date in comparison with other applied technologies. The low efficiency seems to be associated to the aseismic deformation of the reservoir rocks. The observed clustering of earthquakes is delimited by the basement crystalline depth. From an operational point of view, we determine that, like most cases associated with fluid injection, volume of fluid is the variable that determines change in the seismic moment released. Furthermore, the sequence of events in Puerto Gaitán may not fit into a well-known correlation between the volume of fluid injected and the maximum expected magnitude. The observed magnitudes in Puerto Gaitan are well bellow compared to those reported in the literature for similar volumes of injected fluid.〈/span〉

Description: The occurrence of induced seismic events during hydraulic fracturing of reservoirs to enhance permeability is an unavoidable process. Due to the increased public concern with respect to the risks imposed by induced seismicity, however, the development of a soft stimulation method is needed creating higher permeability with less induced seismicity. We use a discrete element model of naturally fractured rock with pore fluid flow algorithm in order to analyse two scenarios of high-pressure fluid injection (hydraulic fracturing) at depth and associated induced seismicity. The ratio of pumped-in energy to released seismic energy is in agreement with field data. Our results suggest that cyclic reservoir treatment is a safer alternative to conventional hydraulic fracture stimulation as both, the total number of induced events as well as the occurrence of larger magnitude events are lowered. This work is motivated by results of laboratory triaxial indenter tests on granite rock samples where continuous loading leads to a wide fracture process zone while cyclic treatment with frequent starting and stopping of loading fatigues the rock, resulting in smaller damage volume and more persistent fracture growth.

Description: In this paper, an underground experiment at the Äspö Hard Rock Laboratory (HRL) is described. Main goal is optimizing geothermal heat exchange in crystalline rock mass at depth by multistage hydraulic fracturing with minimal impact on the environment, that is, seismic events. For this, three arrays with acoustic emission, microseismicity and electromagnetic sensors are installed mapping hydraulic fracture initiation and growth. Fractures are driven by three different water injection schemes (continuous, progressive and pulse pressurization). After a brief review of hydraulic fracture operations in crystalline rock mass at mine scale, the site geology and the stress conditions at Äspö HRL are described. Then, the continuous, single-flow rate and alternative, multiple-flow rate fracture breakdown tests in a horizontal borehole at depth level 410 m are described together with the monitoring networks and sensitivity. Monitoring results include the primary catalogue of acoustic emission hypocentres obtained from four hydraulic fractures with the in situ trigger and localizing network. The continuous versus alternative water injection schemes are discussed in terms of the fracture breakdown pressure, the fracture pattern from impression packer result and the monitoring at the arrays. An example of multistage hydraulic fracturing with several phases of opening and closing of fracture walls is evaluated using data from acoustic emissions, seismic broad-band recordings and electromagnetic signal response. Based on our limited amount of in situ tests (six) and evaluation of three tests in Ävrö granodiorite, in the multiple-flow rate test with progressively increasing target pressure, the acoustic emission activity starts at a later stage in the fracturing process compared to the conventional fracturing case with continuous water injection. In tendency, also the total number and magnitude of acoustic events are found to be smaller in the progressive treatment with frequent phases of depressurization.

Description: 〈span〉〈div〉Summary〈/div〉Since 2013 to date more than 1000 seismic events have been recorded by the Colombian Geological Survey (GSC) in the municipality of Puerto Gaitán (Colombia). A total of fourteen earthquakes are moment magnitude M〈sub〉W〈/sub〉 greater than 4.0. The largest event ever recorded in the area occurred in November 2015 with M〈sub〉W〈/sub〉 4.8. It seems like the case of Puerto Gaitán is associated with the deep injection of co-produced wastewater from oil and gas extraction. The data presented in this work suggests a close relationship in space and time between injection operations and seismicity. An analysis of temporality between both datasets resulted in a time lag equivalent to about 218 days. For this paper, we computed the input and output energy during injection operations from 2013 to 2015 in order to estimate the fraction of total input energy that is radiated as seismic waves. Our results suggest that the seismic energy is only a small fraction of the total energy into the system. Although Puerto Gaitan is one of the places with the most significant volume of wastewater injected among the ones reported in the literature, the energy efficiency of the system is the lowest reported to date in comparison with other applied technologies. The low efficiency seems to be associated to the aseismic deformation of the reservoir rocks. The observed clustering of earthquakes is delimited by the basement crystalline depth. From an operational point of view, we determine that, like most cases associated with fluid injection, volume of fluid is the variable that determines change in the seismic Moment released. Furthermore, the sequence of events in Puerto Gaitán may not fit into a well-known correlation between the volume of fluid injected and the maximum expected magnitude. The observed magnitudes in Puerto Gaitan are well bellow compared to those reported in the literature for similar volumes of injected fluid.〈/span〉

Description: Earthquake focal mechanism solutions (FMS) form the basic data input for many applications, e.g. stress tensor inversion or ground-motion prediction equation estimation. In these applications the FMS data is usually binned spatially or in predetermined ranges of rake and dip based on expert elicitation. However, due to the significant increase of FMS data in the past decade an objective data-driven cluster analysis is now possible. Here we present the method ACE (Angular Classification with Expectation-Maximization) that identities clusters of FMS without a priori information. The identified clusters can be used for the classification of the Style-of- Faulting and as weights for FMS data binning in the aforementioned applications. As an application example we use ACE to identify FMS clusters according to their Style-of- Faulting that are related to certain earthquake types (e.g. subduction interface) in northern Chile, the Nazca Plate and in Kyushu (Japan). We use the resulting clusters and weights as a priori information for a stress tensor inversion for these regions and show that uncertainties of the stress tensor estimates are reduced significantly.