This paper discusses the integration of microseismic fracture mapping with fracture engineering, well production results and offset well interference data in the Overton Field, East Texas. Target formation is the Taylor Cotton Valley at depths of about 11,500 ft. The fracturing program in this field included different types of waterfrac and linear gel hybrid frac treatments, which were compared to evaluate the optimal type of fracture design in this area. Detailed production analyses of were performed to evaluate well performance in conjunction with fracture geometry measurements provided by microseismic fracture mapping results, calibrated fracture modeling and direct production interference data, which provided interesting insights into effective fracture lengths, reservoir quality, and drainage areas.

A large hydraulic fracture diagnostic project was undertaken in the summer of 2001, which integrated fracture diagnostic technologies including tiltmeter (surface and downhole) and microseismic mapping. The extensive data gathered resulted in a much clearer understanding of the highly complex fracture behavior in the Barnett Shale of North Texas. The detailed fracture mapping results allowed construction of a calibrated 3-D fracture simulator that better reflects the observed mechanics of fracturing in this fractured-shale reservoir. More than just simple calibration was required. Indeed, a whole new understanding of fracture growth was developed.

The Barnett Shale has seen a rebirth of drilling and refracturing activity in recent years due to the success of waterfrac or "light sand" fracturing treatments. This extremely low permeability reservoir benefits from fracture treatments that establish long and wide fracture fairways, which result in connecting very large surface areas of the formation with an extremely complex fracture network. Understanding

Multiple hydraulic fractures have been observed in numerous field settings1-9. Direct evidence for the simultaneous growth of multiple hydraulic fractures has emerged over the last decades by coring through propped hydraulic fracture treatments in reservoirs; mineback experiments; intersection of multiple fracture (planes) through offset wells; multi-planar fracture propagation observed in micro-seismic fracture mapping and tiltmeter fracture mapping; and laboratory experiments showing complex fracture growth. The simultaneous growth of multiple hydraulic fractures beyond the near-wellbore area is therefore probably more the rule than the exception, especially when fracturing from a large perforated interval or in naturally fractured formations.

A comprehensive and consistent approach to approximately quantify and model multiple hydraulic fractures growth using observed fracture pressure feedback has provided significant insight into the mechanism of multiple fracture growth. A simplified concept of "equivalent" (equal sized) multiple fractures "competing" for fracture opening and leakoff has been used to approximate this complex phenomenon. This concept is very useful for quick, real-time fracture treatment analysis, and modeling results have also been calibrated by direct fracture diagnostic technologies to limit potential modeling solutions within reasonable bounds.

This paper presents case histories and field data that details the application of well test data to "calibrate" fracture models, understand fracture performance, and optimize treatment designs. The paper includes field results from gas reservoirs in the Rocky Mountains, West Texas and East Texas. In addition, the case histories illustrate the application of postfracture pressure buildup data to identify reasons for fracture treatment problems and confirm treatment success. The paper also includes a novel data set that illustrates fracture performance from a liquid CO2 stimulation (no sand, no water) in West Texas. The case histories illustrate the importance of integrating well test data with fracture treatment optimization, showing how this process can improve our ability to reliably make multi-million dollar decisions concerning proppant selection and fracture treatment size.

Minifrac tests are performed to estimate the key design parameters prior to propped-fracture treatments: leak-off coefficient, closure stress and net fracture overpressure. It is very important, especially in the case of Tip-Screen-Out or Fracand-Pack designs, to determine these parameters as accurately as possible, since errors will enhance the chances of a premature screen-out of the propped fracture treatment.

The key parameters are determined by analyzing the pressure decline after a pump-in shut-in test. For this analysis, one assumes that the pressure falls smoothly after shut-in. However, in many cases a momentary increase is observed in the wellhead pressure after shutting in the pumps, which could compromise reliable interpretation of the pressure decline. In order to determine the cause of the pressure bump, we examined a group of mini-frac data sets which exhibited this anomalous behaviour and mini-fracs in comparable wells that showed a normal decline.

We found that the cause of the pressure bump was severe under displacement of crosslinked gel at the end of pumping the mini-frac. The crosslinked gel at the bottom of the well- bore may create a plug, which effectively makes the wellbore a closed vessel disconnected from the fracture. Heating of the relatively incompressible fluid in the isolated wellbore section then causes the pressure to increase until the yield strength of the crosslinked gel has been overcome and the plug starts to move downwards. This explains the observed bump in the pressure decline plot.

To avoid this bump it is recommended to displace the crosslinked gel to the perforations or even slightly overdisplace. In our experience, if a pressure bump is still observed, it is best to check the calculated and measured displacement volume and repeat the minifrac. If this is impossible, the simplest correction for the pressure bump is to use the initial pressure decline before the bump, and shift the final decline after the bump downwards by the magnitude of the bump. However, this procedure would be invalid if the point of fracture closure coincides with the bump.

In 1994 and 1995, a series of hydraulic fracture stimulation treatments were performed in the McKittrick field Point of Rocks formation, a thick (>2000') marine sequence of low permeability (0.1 ml)) sandstone intervals separated by shales. The thick interval, 10,000' depth, and a formation temperature of nearly 300 F posed challenges to effectively completing the desired intervals.

In general, the initial series of treatments targeted large gross perforation intervals and incorporated a large pad fraction. The treatments included few procedures to enable engineering evaluation of actual in-situ fracture behavior, which was later found to be very different from theoretical or pre-job estimates. Due to some combination of fracture treatment design and reservoir limitations, the production response results from the 1994-995 treatments were mixed, resulting in marginal overall project economics.

In 1996, six additional fracture treatments were performed in a new Point of Rocks well, No. 733-17Z. However, a very different fracture design methodology was utilized, focusing on real-data (net pressure) analysis. Objectives for these treatments included a) to place a more effective stimulation than resulting from past efforts, b) to characterize fracture behavior so that reasons for success or failure were clear, and c) to begin the process of optimizing treatment design. A more surgical fracturing strategy was employed, targeting selected Point of Rocks intervals using small perforation intervals and multiple independent fracture stages. Changes were also made to more effectively place fracture conductivity in the intended intervals, including a dramatic reduction in pad fraction and an increase in maximum proppant loading. These changes were implemented with the support of real-time fracture pressure analysis, enabling on-site diagnosis of proppant placement problems, and final treatment design refinement based on observed fracture behavior.

Compared with past treatment results, the new hydraulic fracture strategy resulted in a dramatically improved production response. With initial commingled production of over 6 mmscfd and 200 bopd (over 1200 boegd), well 733-17Z had a larger IP and has not experienced the rapid production decline of the previous Point of Rocks wells. With this success, the use of real-data fracture engineering will continue; there remains significant potential for treatment optimization in further planned field development.

This paper presents pre- and post-fracture pressure transient analysis, combined with net fracture pressure interpretation, for a well in a naturally fractured geothermal reservoir. Integrated analysis was performed to achieve a consistent interpretation of the created fracture geometry, propagation, conductivity, shrinkage, reservoir flow behavior, and formation permeability characteristics. The interpreted data includes two-rate pre-frac injection tests, step-rate injection tests, a series of pressure falloff tests, and the net fracturing pressure from a massive fracture treatment. Pressure transient analyses were performed utilizing advanced well test interpretation techniques and a thermal reservoir simulator with fracture propagation option. Hydraulic fracture propagation analysis was also performed with a generalized 3-D dynamic fracture growth model simulator.

Three major conclusions resulted from the combined analysis: 1) that an increasing number of hydraulic fractures were being simultaneously propagated during the fracture treatment, 2) that the reservoir behaved as a composite reservoir with the outer region permeability being greater than the permeability of the region immediately surrounding the wellbore and 3) that the created fractures extended into the outer region during the fracture treatment but retreated to the inner region several days after stimulation had ceased. These conclusions were apparent from independent pressure transient analysis and from independent hydraulic fracture propagation analysis. Integrated interpretation, however, increased the confidence in these conclusions and greatly aided the quantification of the created hydraulic fracture geometry and characterization of the reservoir permeability.

Premature screen-outs and/or low proppant concentration are the most likely cause of failure in hydraulic fracturing treatments. Although commonly blamed on a variety of presumed problems-most typically the treating fluid, or large-scale reservoir conditions, such as permeability or stress profile-the true source of most problems has been uncovered only recently by careful analysis of treatment data. The source is referred to as near-wellbore tortuosity, but it can variously arise from deviatoric stress, natural fractures and/or perforation-dominated creation of complex fracture patterns in the wellbore vicinity.

Numerous theories have been formulated to deal with near-wellbore screen- outs and, especially for oriented wellbores from Arctic or offshore platforms, various perforation strategies have been postulated and/or implemented. In contrast to the idealizations and costs associated with those theories and strategies, this paper presents simple cheap solutions that are less sensitive to the wellbore environment. This novel strategy involves injection of proppant slugs into the near-wellbore region and, when necessary, immediate shut-ins upon small slugs, with three important results: the response of the near well-bore region can be measured and characterized; a large part of the near-wellbore tortuosity can be removed, by simplifying the near-wellbore fracture pattern; and the true nature of the large-scale reservoir response can be determined, e.g. from the greatly modified pressure fall-off obtained after placing slugs near the wellbore.

The paper reports the concept and implementation, in a number of commercial fracturing environments, in both gas and oil reservoirs, with both foam and liquid-gel jobs. These show the effective removal of tortuosity varying from 20 to 200 bars and associated elevation of allowable proppant concentrations.

In order to increase gas production in an under-producing formation, the German utility/energy company RWE-DEA stimulated the Wardboehmen Z1 well with a propped fracture treatment on December 6, 1991. Two days prior to the main fracture treatment, a step-rate test and mini-fracture were pumped for the purpose of evaluating more accurately the characteristics of the reservoir, including closure stress, leak-off rates, and permeability (and stress) profile. This information was then used to substantially improve the originally proposed design, provided by the service company, in order to create an effective/optimal propped fracture in this reservoir, incidentally also producing major savings in job cost. For the first time in Germany, electronic bottomhole pressure-measuring equipment with surface readout was used, during the minifrac. Availability of the bottomhole pressure data in the minifrac and repeated abrupt flow-rate changes, including shut-ins of about one minute, during both the minifrac and main fracture treatment allowed realistic simulation of fracture development with an on-site real-data 3D simulator. It was possible to determine the continually changing friction losses in the near-wellbore vicinity of the perforations: this tortuosity was very high initially but reached acceptable values after pumping the re-designed pad volume. This finding was important because it minimized the risk of a premature near-wellbore screen-out resulting from the planned proppant concentrations, up to 1,250 g/l (10 ppg). The fracture treatment led to a consistent four-fold increase in the gas production rate. This success was due, at least partly, to careful planning and use of novel technology in the fracture treatment, which allowed the on-site determination of actual reservoir conditions, with growth influenced by extremely variable permeability as against idealized models used for initial design.