|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
Abstract A vast number of the reported cases of increased seismicity of moderate magnitude (Mw > 0) earthquakes seem to be tied to some form of fluid injection activitiy, being it wastewater disposal by injection into deep wells or high pressure fluid injection into oil and gas reservoirs to hydraulically fracture the rock and improve hydrocarbon recovery. Regulations have been proposed to implement traffic light systems to dictate the responses that the industry needs to take based on either the magnitudes or observed particle velocities or accelerations on the surface. In order to relate the seismic hazard potential in seismically active areas, empirical ground motion prediction equations (EGMPE) are used to relate event parameters like magnitude and location to site characteristics such as peak ground acceleration (PGA) or peak ground velocity (PGV) which tend to be how building codes are parametrized. Therefore, local hazard assessment near hydraulic fractures that generate relatively large magnitude events need to be estimated more precisely by developing and using local EGMPEs. Hybrid deployments combining 15Hz downhole and low frequency near-surface geophones can be used to accurately capture both the localized microseismic events and any large magnitude events associated with hydraulic fracture monitoring across North American basins – Horn River, Eagle Ford, Barnett, and Montney for example. In our studies events with M>0 are observed for completions in these formations. While in many cases the magnitude of these events is too small to be felt on the surface, there are reports of higher magnitude events which have been sensed by workers on site and the local population. The exact relationships between magnitudes and shaking are not necessarily one-to-one. Shaking also varies based on the stress release of the events. As summarized recently by Hough (2014) for other fluid-induced seismicity, the lower stress releases typical for these sequences results in on-average less shaking than is observed for equivalent magnitude tectonic events. In order to quantify shaking over a seismogenic volume, we show how to develop EGMPEs based on the North-American examples. The EGMPE methodology developed in this study can be extrapolated for similar earthquakes of larger magnitude and included into future probabilistic hazard and risk analysis for induced seismicity as related to hydraulic fracture stimulations.
Abstract Induced seismicity resulting from fluid injection is a growing concern with a number of operations, including hydraulic fracturing. The vast majority of hydraulic stimulations results in no felt seismicity. However, three examples of larger, anomalous seismicity have been attributed to hydraulic fracturing, which seem to be associated with operations in unique geologic and geomechanical settings. In response, a number of operational protocols have been developed and include specific requirements for seismic monitoring. Seismological aspects are obviously central to these protocols, including characterizing the seismic source strength and associated seismic hazard. The typical microseismicity recorded during hydraulic fracturing represents a small portion of the hydraulic energy associated with the injection. However, the energy balance of the relative amount of seismic energy increases in the cases of anomalous seismicity, which may provide a monitoring tool to potentially help mitigate induced seismicity. Although the number of cases with anomalous seismicity is relatively small, other examples have been observed from geothermal stimulations. In these cases, the ratio of seismic energy is relatively larger but of potentially interest remains significantly less than the hydraulic energy. Furthermore, the ratio of seismic moment to injected volume also increases but typically remains less than a limit suggested by McGarr (1976). Potentially the energy and volume balances could be useful monitoring tools to assist in ongoing operation decision processes.
As regulators introduce operational protocols based on seismic magnitudes occurring during hydraulic fracturing, operation strategies are required to mitigate the seismic hazard. In certain cases, pressure increases associated with the hydraulic fracture injection, can induce slip of tectonically stressed faults leading to triggered seismicity. A coupled hydrogeomechanical model is used here to examine fault activation during multi-stage hydraulic fracturing and to examine operational scenarios. The model shows that the relative stage sequencing relative to the direction of fault slip can significantly impact the fault slip. Changing the viscosity of the fracturing fluid is also explored as an operational mitigation scenario, which is found to have a strong impact on fault slip and associated seismic magnitudes.
Presentation Date: Tuesday, September 26, 2017
Start Time: 11:25 AM
Presentation Type: ORAL
Summary Induced seismicity associated with hydraulic fracturing has become a significant regulatory issue in Western Canada, after the occurrence of felt events in three specific reservoirs. Seismicity results from elevated pressure of the stimulated fracture network reducing the effective clamping force and triggering slip of tectonically stressed faults. Several published examples indicate activation over progressively larger fault regions, as multiple stages interact with critically stressed faults in various relative orientations to the treatment well. Geomechanical modeling is used to examine progressive fault slip from multi-stage fracturing and the associated seismicity. The modeling explores different operational scenarios to examine progressive fault activation as hydraulic fracture stages sequentially pressurizes more of the fault. Testing these mitigation strategies on faults in different orientations provides potential operational guidelines to support seismic traffic light systems, typically used to mitigate injection induced seismicity. Simulations show that the amount of injected fluid interacting with the fault plane controls the intensity of observed seismicity for a specific fault. Stages farther away from the fault can have an impact on fault slippage but with a delayed effect. Sequence of propagation of the hydraulic fracture stages compared to fault orientation is important. If the first stage is closest to the fault, more of the injected fluid will interact with the fault, triggering a large slipping patch on the fault plane. Successive stages will have a lesser effect due to stress shadowing. However if the first stage is the most distant from the fault, slippage on the fault plane will be gradual, thus reducing the amount of seismic moment release. The sequence in which wells on a multi-well pad are stimulated could also impact the associated seismicity. Introduction There has been increasing occurrences of recent seismicity associated with fault activation during hydraulic fracturing, resulting from elevating pore pressure on optimally-oriented, pre-existing faults leading to triggered release of stored tectonic energy (Maxwell, 2013). Anomalous seismicity is similar to microseismicity, although larger magnitude fault seismicity corresponds to inelastic slip over a larger area. However, triggered fault slip has in certain conditions lead to felt ground shaking. For example, three Western Canadian reservoirs located close to the tectonically-active trust belt have experienced anomalous activity: specifically at localized regions of the Horn River Basin, Montney and Duvernay Shales (Atkinson et al., 2016). Operators and regulators in Canada have proactively engaged the issue, establishing operational practices including seismic monitoring for a traffic light system to guide seismic hazard mitigations strategies. Clearly the topic continues to be of concern and establishing mitigation best practices is increasingly important.
ABSTRACT: Seismic events near the stope face of six longwalls at Western Deep Levels Limited are discussed in terms of their relation to blasting time. First, the time period is established during which seismic events can be regarded as directly blasting-induced. Later, the different seismicity levels of the six longwalls are discussed. Finally, the ratio of the number of seismic events, which occurred during the blast and after the blast is evaluated for several magnitude categories. Longwall geometry, geological discontinuities of large and small scales did however result in anomalies such as abnormal concentrations of rockbursts during and outside blasting time. It is also shown that seismicity levels of longwalls, which advanced at an oblique angle to geological features, were strongly reduced when compared with other longwalls. The potential of production blasts to trigger impending seismic events under certain conditions becomes apparent.
The gold mine Western Deep Levels Limited (WDL) is situated in the Witwatersrand Basin approximately 75 km south-west of Johannesburg. The lease area totals nearly 45 km2, extending 11km on strike and about 4km on dip. Mining operations began in 1957 with shaft sinking operations. The first gold pour took place in 1962. Two economic gold-bearing reefs, the Ventersdorp Contact Reef (VCR) and the Carbon Leader Reef (CLR), are extracted (Fig. 1). The VCR is worked between 1500m and 2300m below surface, dipping on average 21 degrees South-east and sub-outcrops in the north-west. About 900 m below the VCR lies the CLR horizon which is continuous over the whole lease area. The VCR consists of a conglomerate with a great variation in pebble sizes and channel width Which exceeds 2,5 m in places ("reef roll"). The CLR is formed by a narrow, carbon-rich conglomerate with a channel width of a few centimetres.
For the exploitation of the VCR both, No.2 and No.3, shafts were sunk in the northern part of the lease area to 1930 m below surface. To gain access to the CLR sub-vertical shafts were
(Figure in full paper)
sunk down to 2975 m below surface. Further development in the form of tertiary vertical shafts was necessary to enable mining in the lower sections of the CLR-horizon.
The mine adopted a longwall mining system in which approximately 75 % of alliongwalls are protected by systematic stabilizing pillars. Mini-longwalls consisting of six panels, with a total length of about 200m on dip, are separated by 40 m wide strike stabilizing pillars. Stabilizing pillars were introduced in 1980 to address the eminent rockburst problem of the late 70's. Since 1987 backfill in the form of classified tailings is added in some areas to improve the regional support.
The reef is extracted conventionally by drilling and blasting. The broken rock is scraped to boxholes which lead to haulages in the footwall. These haulages ("follow-behinds") are developed some distance behind the actual face-position to avoid high field stresses.
Both horizons are intersected by a number of dykes which are mainly north east - south west orientated.