SPE 153042 Identifying Fault Activation in Unconventional Reservoirs in Real Time Using Microseismic Monitoring Michael Kratz, Microseismic, Andrew Hill, Microseismic, and Scott Wessels

Copyright 2012, Society of Petroleum Engineers This paper was prepared for presentation at the SPE/EAGE European Unconventional Resources Conference and Exhibition held in Vienna, Austria, 20-22 March 2012. This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract Identification of fault related microseismicity in hydraulic fracture treatments is crucial to understanding how treatments are stimulating a reservoir. Evaluating b values in combination with event source mechanism provides a reliable and intuitive method for separating fault related microseismic events from standard fracture related events. Typically this analysis is conducted after a treatment is complete and serves as a diagnostic tool to provide possible explanations for reduced production or designing future treatments on nearby wells to avoid an identified fault feature. Being able to identify such features in real time allows the operator to not only identify faults but to stop treatment and avoid these features all together saving time and materials that would otherwise be pumped into an area that doesnt contribute to the overall stimulation of the reservoir, and could reduce production on the well. When evaluating b-values in real time, a technique that can identify faulting early in the initiation of fault stimulation is crucial for preserving the most resources during treatment. Introduction Microseismic monitoring technology has greatly increased in popularity and sophistication as the hydraulic stimulation of unconventional low permeability reservoirs continues to increase. This technology provides crucial information to add to the assessment of reservoir stimulation. Recent advances have added to the post treatment diagnostic information available to the operator, including methods for determining whether event hypocenters are associated with existing tectonic fault motion or with standard fractures along pre-existing planes of weakness. This is of great use for assessing where the stimulation was most effective and where to avoid treating adjacent wells. However, assessing these conditions post treatment does not save on treatment and material costs associated with pumping into a large fault feature. Making this determination in real time, close to initial activation of the feature will allow operators to cease pumping and avoid wasting time and money on ineffective stimulation. Continuing the treatment down wellbore a frac crew would be able to avoid the feature all together. b-values were first defined by Gutenberg and Richter (1954) as the frequency magnitude distribution of a given earthquake catalog. This relationship is represented by log N=A-Bm. Maxwell et. al. (2009) and Downie et. al. (2010) observed that for standard hydraulically driven fractures, b-values will evaluate to ~2, and fault activation will show b~1. When calculating b-values, the Gutenberg Richter equation forms the basis of evaluating frequency magnitude distributions with large natural earthquake catalogs. There have been many improvements to the formula since its discovery in order to make more accurate judgments of b-values and to take into account sources of error such as binning of magnitudes and temporal variability in minimum catalog magnitudes. For the purposes of this paper, we used the Tinti & Mulargia (1987) updated equation.

1.Full events set across the all 4 wells. The events are sized by energy and colored by source mechanism. Red for strike-slip, blue for dip-slip

2 SPE 153042

Where !M is the binning width of the catalog and

Where " is the average magnitude of the catalog and Mc is the magnitude of completeness or the smallest magnitude at which 100% of events are recorded.

Source mechanism inversion of microseismic events can be used to identify failure mechanisms that vary spatially across a treatment. This allows for the separation of microseismic events into two populations, giving a more accurate picture of the forces driving the event. We calculate source mechanisms with a least squares inversion of the observed P-wave amplitudes, assuming a point source, Williams-Stroud et al. (2010).

Case Study This paper focuses on four horizontal wells completed in Montague County, TX. This completion targets the lower Barnett shale which lies unconformably above the Viola/Simpson limestone and the Ellenberger limestone approximately 2-12 miles from the Muenster arch. The treatment was performed in a zippered fashion with the western wells being treated from South to the North and eastern wells in reverse. This data set exhibited to dominant sources mechanisms during the course of the treatment. Maximum horizontal stress in the region is oriented NE-SW, however natural fractures in the area are oriented WSW-ESE (Waters et. al., 2006). While natural fractures would be expected to be the dominant source of microseismicity in the area, source mechanism inversion of the data set shows a predominantly 040/90/-125 (strike,dip,rake) mechanism of fracturing. In addition to this mechanism, in the SE we see a different source mechanism with a failure plane of 240/80/10 and trending along a NE-SW linear trend lining up with maximum horizontal stress.(Fig. 3) The events along this trend are proposed to be a tectonic fault, and have been shown to have b-values that verify this claim (De La Pena et. al. 2010) making this a good data set to analyze from a real time stand point.

2. Event that have occurred at moment of positive identification of the fault

3. Equal area stereographic beachball plots of the two source mechanisms present here.

4. Frequency magnitude distribution curves used to determin Mc by maximum curvature.

SPE 153042 3

Statement of Theory and Definitions or Methods Monitoring hydro-frac treatments in real time poses a cadre of challenges. The overall goal is turn around, or the ability to return meaningful results faster and faster. In order to analyze how b-values for this data set could be of use in real time, they were sorted by time and then separated by source mechanism, as they would be in a real time operation. Identifying separate

source mechanisms within the same treatment tells us that there is a fundamental difference in the character of the events being generated during the treatment and is the first clue that a fault may be present. Separating them allows us to dig deeper into the driving forces behind the events being generated. Once sorted by source mechanism, what we have is two independent and growing catalogs. As soon as these catalogs reach a minimum of 50 events in length, they are complete

Standard Fracture Events Fault Events

Standard Fracture Events Fault Events

6. scatterplot of energy released by each event through time. The shaded area represents the period of activation of the fault feature, from onset to full activation. The red line reperesents where the positive identification of a fault signature occurs.

5. b value time series separated by mechanism. Red dotted line indicates point in time series in which a positive fault signature is confirmed.

4 SPE 153042

enough to perform a b-value calculation. In this data set, this occurs at 30 minutes into the treatment for both source mechanisms. Once both catalogs reach 50 events, the b-value for the catalog is recalculated using an expanding window, such that each new event is added to the appropriate catalog and the b-value is recalculated for the whole data set. The minimum magnitude of completeness is recalculated at each step by maximum curvature. However, each mechanism showed a unique magnitude of completeness over the entirety of each data set with the 040/90/-125 mechanim showing a magnitude of completeness of -1.7 and the 240/80/10 mechanism showing a magnitude of completeness of -1.8. (Fig. 4) In order to identify faults in real time we would need the b-values for one of the two source mechanism to reach a stable trend of ~2. This means that as we watch the curve in real time a b-value signature stabilizing around a value of 1 is a signal that this mechanism is associated with faulting. (Fig. 5) Once we have separated the catalogs by source mechanism and identified a positive fault signature in the b-value plots we look at energy release rates for a threefold confirmation of fault related activity during the treatment. We calculate the energy released with each microseismic event as log10J=1.5Mw+4.8. Plotting the energy released event by event for each source mechanism through time should show that events with a b-value signature of 1 should show larger energy values during the time of the first activation of the fault through to the fault achieving its full length. What was noticed in this data set is that there seems to be a zone of reduced microseismic density where treatment energy jumps across from the treating stage into the fault feature. This damage zone surrounding the fault seems to represent the maximum length at which the feature can capture fracture energy and may serve as a way to provide some guidance for how far to move down the wellbore to completely avoid the fault when it is expressed in the data. What causes this and if it is common enough to be a rule is unclear, but this zone forms a uniform spacing on either side of the feature within this data set.

Conclusions On this treatment the proposed fault hits full activation at ~6:00 am on the 19th , while positive identification of a fault signature occurs 24 hours earlier . In this case we define full activation as the feature realize its full length. We see a positive identification in all three indicators well in advance, before it is able to reach full activation and waste time and resources. By skipping past the fault feature before continuing the treatment, its potential effect on the treatment and production of the well can be minimized. Here this would mean saving three stages worth of fracing materials, not to mention 24 hours of treatment that will not be contributing to the effective stimulation of the well. References De La Pena, A., Wessels, S.A., Gunnell et. al., 2001. Fault or Frac? Source mechanism and b value detection of fault fracturing-A Barnett case study, EAGE C001 Downie, R.C., Kronenberger, E., Maxwell, S.C. 2010. Using Microseismic Source Parameters to Evaluate the Influence of Faults on Fracture Treatments - A Geophysical Approach to Interpretation, SPE 134772 presented at the Society of Petroleum Engineers, Florence, Italy, September 19-22. Gutenberg, B., and C. F. Richter, 1956 Magnitude and energy of earthquakes, Ann. Geofis., 9,1-15 Maxwell, S.C., Jones, M., Parker, R. et. al., 2009. Fault Activation During Hydraulic Fracturing, SEG Expanded Abstracts 28, 1552. S. Williams-Stroud, L. Eisner, A. Hill et. al., 2010. Beyond the Dots in the Box Microseismicity-constrained Fracture Models for Reservoir Simulation presented at EAGE Barcelona, Spain 2010. Tinti, S. and F. Mulargia 1987. Confidence intervals of b-values for grouped magnitudes, Bull. Seismol. Soc. Am., 77, 2125-2134. Waters, G., J. Heinze, R. Jackson et. al., 2006, Use of Horizontal Well Image Tools to Optimize Barnett Shale Reservoir Exploitation: SPE 103202, presented at the SPE Annual Technical Conference. DOI: 10.2118/103202-MS.

MAIN MENUTable of ContentsSearchPrint