Successful development of the Frontier formation in southwest Wyoming begins with the hydraulic fracture treatment design and implementation. Without a design to effectively stimulate the Frontier formation and a proper implementation of that design in the field, production rates will be limited. This paper will describe the results from an extensive drilling and completion program conducted by Bannon Energy Incorporated in the Frontier formation on the Moxa Arch. Information from over 100 fracture treatments and the production response from over 550 wells will be used to demonstrate the significant improvement in gas production resulting from the fracture treatment designs and implementation methods employed. The effect of fracture conductivity on gas recovery from this formation will be demonstrated along with the effect of multi-phase flow on fracture conductivity. Introduction The Moxa Arch area Frontier formation has been under steady development since the mid 1970's. Over 600 Frontier completions have been made in the study area of this paper. Figure 1 indicates the completion activity by year since 1976. Over half of the total completions have been made since January 1992. Completion activity in the early 1990's was elevated due to incentives for "tight" gas development. The continued activity in 1994 however is due to successful drilling and completion practices which have made the formation more attractive for development. Production from the Frontier is centered along the axis of the Moxa Arch which plunges north to south. Figure 2 is a map indicating the study area along the arch along with the productive sections, Bannon Energy Incorporated (BEI) developed acreage and a boundary between what we will define as Interior wells and field Delineation wells. The geology of the Frontier has been described in great detail in previous publications and studies. While we have not conducted an independent geological study, we have noted from production response and log analysis that north-south offsetting wells have a greater tendency to be similar than east-west offsets. Hydraulic fracturing treatment designs used on the Frontier formation in the late 1970's and early 1980's consisted primarily of cross-linked water based gels with sand as the proppant Production response from these treatments did not exhibit the type of decline associated with a simulated "tight" gas sand. This unexpected behavior was attributed to formation clay swelling and damage caused by the water based fracturing fluids. In the late 1980's, foamed fracturing fluids were used to reduce the amount of fluid in the fracturing treatment. Intermediate strength proppants (ISP) were utilized in these treatments. Observed production response from the foamed fluid treatments was not perceived to be much better than the previous water-based fluid treatments. The cost of the foamed treatments with ISP were significantly higher than the earlier treatments with sand ($300,000 vs. $110,000). Therefore, a return to water-based fluids and the use of sand was determined to be more economic. Production response from these recent treatments has been similar to the response noted in the early 1980's, which were, again, not typical of stimulated "tight" gas sands and thought to be damaged. The goal of the authors efforts during the BEI Frontier development program has been to place a highly conductive propped fracture through the use of intermediate strength proppants and modern fracturing techniques. P. 143

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractFracturing fluid that remains in the fracture and formation after a hydraulic fracture treatment can decrease the productivity of a gas well by reducing the relative permeability to gas in the region invaded by this fluid. This fluid can block the gas flow into the fracture, thus reducing the effective fracture length. Pressure transient tests performed on hydraulically fractured wells often reveal that the effective fracture half-lengths are substantially less than the designed length from fracture stimulation.In this work we used reservoir simulation to determine the relationship between fracture fluid production, effective fracture length, and gas productivity. While the effective fracture length is affected by such factors as non-Darcy flow, it is related directly to fracture cleanup, and increases with time. From this study we found that the rate of fracture fluid production is affected significantly by the conductivity of the fracture. Greater dimensionless fracture conductivity results in more effective well cleanup, longer effective fracture lengths versus time, and greater effective stimulation of the well.The results of this study provide a better understanding of the gas production behavior from wells hydraulically fractured using water-based fracturing fluids.The relationships between fracture conductivity, effective fracture length, and gas productivity presented in this paper can be used in economic calculations to balance the costs of higher fracture conductivity against the additional revenue resulting from longer effective fracture lengths. Results presented will allow operators to better design optimal fracture lengths for typical gas reservoirs.

Low permeability, or "tight", gas reservoirs are being developed at an ever increasing rate in the U.S. The amazing increase in activity in the Rocky Mountain region over the past decade is a testament to this. Currently, there are several "tight" gas plays in the U.S. that involve the commingling of multiple intervals in order to gain economic viability. The Pinedale Anticline of southwestern Wyoming is one of these areas. The Pinedale Anticline completions pose a particularly complex problem when attempting to evaluate the "best" method of stimulation because as many as twenty-two separate stimulation treatments are placed in up to 70 discrete sand intervals over a gross interval up to 6,000 feet thick. Evaluations are further complicated by variation in permeability exceeding two orders of magnitude and pore pressures increasing from 0.44 psi/ft to 0.83 psi/ft. The analysis of "tight" gas reservoirs has been the topic of many SPE papers over the past twenty years. Several have presented data indicating the broadness of the permeability distribution which may be encountered when developing these reservoirs.[1,2,3] The broadness of the permeability distribution, often over two orders of magnitude in breadth, poses a statistical problem when trying to simply compare production response of one set of data to another in a given field. We will quantify the significance of this and present statistical evaluations documenting the probability of obtaining two similar data sets with respect to permeability when broad distributions exist. We then compare the size of the sample set necessary to quantify stimulation effectiveness using production alone with the sample size required when using reservoir simulation. The reservoir simulation analysis presented in the paper demonstrates a process for use in multiple layered reservoirs for evaluating stimulation effectiveness. The process requires significantly fewer field tests than if production rates were used alone. Multiple production logs were utilized over several producing months in selected wells and are crucial to the production history match process. A wide variety of proppant products are investigated and compared to expected performance from published specifications. This paper will aid engineers working in multi-layered reservoirs understand the complexity of the evaluation process and give them a process for evaluating stimulation effectiveness in their reservoirs. Introduction Development of the Pinedale Anticline of southwestern Wyoming has continued at an aggressive pace over the past several years. Massive hydraulic fracturing (MHF) treatments are the only means of stimulating production to economically accecptable levels from the "tight" gas sands present in this area. Each well being completed generally requires between 14 and 22 MHF treatments in order to effectively produce from all potential pay intervals. Over two million pounds of proppant are often used per well, representing a potential high investment cost to the operator. The ability to evaluate the incremental production benefit associated with the use of one proppant versus another can have a significant impact on the profitability of the field development. Several studies have been performed in the past which have utilized a comparison of production values to compare the performance of proppants.[1,2,4,7] Other studies have incorporated the use of reservoir simulation to remove reservoir properties variability from the equation.[3,8] Still others have used a normalization process for the removal of reservoir variability.[5,6] Each of these approaches to quantifying the ability of a proppant to increase stimulation effectiveness has its own pros and cons. In the next few paragraphs we will discuss these advantages and disadvantages to each approach and present the premise that our study is based upon.

Proppant conductivity is an important design criterion in hydraulic fracturing treatments. Knowing how different proppants behave under changing stress conditions is important to fracture stimulation success. Conductivity and non-Darcy flow effects has been laboratory measured for all ceramic proppants. Unfortunately, almost all laboratory measurements are performed with an increasing stress and cyclic stress behavior is not observed. Often, in the production of oil and gas wells, shut-in periods occur and pressure in the wellbore and proppant pack increases causing stress to be relieved on the pack. When production begins again, stress is increased on the pack. This is stress cycling and past publications1 have noted that stress cycling can cause a reduction in proppant pack conductivity. The work presented in this paper chronicles a vast series of crush tests performed on several proppant types and sizes. The crush tests were run using exactly the same starting sieve distribution for each particular proppant tested. After a sample was subjected to stress or multiple stress cycles, it was sieved for a detailed analysis. The change in the sieve distribution was noted and a median particle diameter was calculated along with the standard deviation of the distribution. A relationship was observed between these values and those of conductivity and Beta Factor. From these relationships it can then be possible to estimate the effect that stresses cycling has on conductivity and non-Darcy flow effects. The papers results can help the fracture design engineer to better understand how conductivity, of the designed proppant pack, will change with time as stress cycling occurs. This understanding can lead to the recovery of more oil and gas through improved stimulation results. Introduction Conductivity is a very important consideration in the design of hydraulic fracture treatments. Many publications have documented this to one degree or another2,3,4. The conductivity that is achieved from a placed hydraulic fracture treatment is a result of many factors. Fracture closure stress, proppant pack concentration, proppant type, proppant sieve distribution, proppant embedment, gel residue damage, stress cycling and others all affect the proppant pack conductivity that can be achieved from a treatment. Baseline conductivity properties, evaluated using API standards5, are available for most proppants available for use in hydraulic fracture treatments. These conductivity tests are performed under controlled laboratory conditions and usually represent the maximum attainable conductivity for any particular proppant for the conditions tested. Proppants come in many different sizes and materials. Some are naturally occurring, while others are man made. Some have coatings to increase cohesiveness and strength. Most proppant sizes are referred to by an API sieve size designation. For instance, a 20/40 mesh proppant would have a sieve distribution in which 90% would fall through the 20 mesh sieve and stay on the 40 mesh sieve. Other typical API proppant designations are 30/50 mesh, 16/30 mesh, 12/18 mesh, 16/20 mesh, and on and on. Unfortunately, all 20/40 mesh proppants do not have the same sieve distribution and can vary greatly in Median Particle Diameter (MPD) depending on whether the sieve distribution is weighted more on the coarse or fine end of the 20 to 40 mesh scale. The importance of MPD becomes apparent in Fig. 1 which shows a plot of MPD versus baseline conductivity at 2,000 psi closure for a wide variety of commercially available ceramic proppants. From this graph we see that conductivity is very strongly related to MPD. As stress increases on the proppant pack, individual proppant balls begin to crush and become smaller pieces. This is why the baseline conductivity at 2,000 psi was chosen, very little proppant crushing, if any, should be occurring at this stress with ceramic proppants. The graph also seems to indicate that the type ceramic material is not as important to conductivity as MPD, since all proppant materials plot on the same trend.

The Mesaverde formation in the Piceance Basin of western Colorado is a series of tight gas sands contained in a complex geological environment. Pay selection and completion techniques have varied greatly during the development of this formation. Production response varies greatly from well to well, and decisions on where and how to optimize completion economics are difficult. Consistent treatment sizes and designs are usually used even though some treatments may be over- designed while others are under-designed. Knowing where to spend Completion dollars and where to save them can substantially impact the economics of a development project in this Basin. This paper will present a reservoir description technique that uses standard openhole well logs and helps operators to predict reservoir quality for individual sands. This prediction allows the operator to optimize the completion design and therefore maximize his economics. This technique has been applied to many wells, and the accuracy is documented through production analysis. Both total well production and individual sand production, determined with production logs, have been predicted through the use of this technique. This paper also examines the economic benefit of using this analysis technique will be presented. The authors believe that the methodology used can be applied to similar reservoirs to achieve similar results. P. 129