BP has started production from a prolific new natural gas well in the Mancos Shale of New Mexico, a discovery that points to the area's potential as a large new gas supply source for the United States. Early production rates at the NEBU 602 Com 1H well in San Juan County are the highest achieved in the past 14 years within the San Juan Basin, a large oil- and gas-producing area spanning southwest Colorado and northwest New Mexico that includes the Mancos Shale. The well achieved an average 30-day initial production rate of 12.9 MMcf/D. The successful well test took place on assets that BP acquired from Devon Energy in late 2015, which expanded the company's position in the basin and provided improved access to the Mancos. The NEBU 602 Com 1H well was drilled with a 10,000-ft lateral in an area known as the Northeast Blanco Unit (NEBU), a section of federal lands in New Mexico's San Juan and Rio Arriba counties and an area where BP has been present since the 1920s.
This work is a novel attempt to analyze rock/brine and oil/brine interactions in heterogeneous unconventional liquid reservoir rocks and the effect of these interactions on oil recovery. There is very limited literature on the impact of brine salinity on shale wettability alteration and spontaneous imbibition experiments. This study includes a comprehensive approach that analyzes rock-fluid interactions through contact angle and zeta potential measurements followed by imbibition studies monitored by CT scan technology to understand the effect of salinity on CT penetration profiles and the resultant oil recovery.
The rock samples were obtained from sidewall coring of an ultra-tight liquid rich shale in South Texas composed predominantly of carbonate while sodium chloride and calcium chloride brines of varying concentrations were used as testing fluids. Contact angle and IFT experiments were conducted on unaltered samples aged in oil obtained from the same well the sidewall samples were retrieved. The captive bubble method was used to measure the contact angles of oil droplets on the rock while IFT was measured using the pendant drop technique. Zeta potentials were measured to assess the stability of thin brine films on the powdered rock samples and oil droplets. Finally, spontaneous imbibition was carried out at optimum salt concentration that resulted in maximum water wetness to measure oil recovery while Computed Tomography (CT) based imaging was used to analyze imbibition profiles and quantify penetration of fluids into the rock.
Experimental results suggest that both sodium and calcium chloride were able to alter wettability of samples from an initial intermediate-wet state to water-wet with the maximum water wetness observed for sodium chloride. Increasing salt concentration also lowered the crude oil/brine IFT marginally. Zeta potential measurements suggested sodium brine films were more stable compared to calcium brines while double layer expansion does not completely explain the wetting changes observed. Sodium and calcium brines at optimum salinities recovered more oil compared to water alone while maximum recovery was obtained with sodium chloride which also penetrated the most into the rock matrix as found by analyzing the CT scan images. This study therefore validates the potential of affordable low salinity injection brines which have the capability to alter shale rock wettability and improve oil recovery by penetrating deeper into the reservoir rock.
Water produced with coalbed methane in the San Juan Basin presents a costly disposal problem. Production increased from 15,000 bpd in 1988 to 115,000 bpd in 1992, and is projected at 180,000 bpd in 1995. Over the next 20 years, 644 million to 1.04 billion barrels of water production is anticipated. Currently, the main disposal method is underground injection. The rate capacity of existing disposal wells in part of the Colorado portion of the basin will not be sufficient to handle all the water to be disposed of. The volume of produced water during the next ten years will exceed the volumetric capacity of the deep disposal zones there, and other disposal zones or methods will be needed. Alternative treatment technologies are available to treat produced water and make much of it suitable for surface discharge. A smaller volume of concentrated brine remaining after water treatment would still require underground injection or other acceptable disposal methods.
The San Juan Basin of southwest Colorado and northwest New Mexico is the heart of the coalbed methane play in the U.S. and the world. More coalbed methane is produced from the San Juan Basin than the rest of the world combined. In addition to natural gas rates of 1.5 Bcfd, brackish water is produced along with the gas, posing a costly disposal issue and a possible constraint to future development.
Water production in 1988 totaled 15,000 bpd from 235 wells, and increased to 1 1 5,000 bpd from 1800 wells by late 1992. 1,200 additional wells had been drilled by year-end 1992, and 460 new well locations had been approved. As these new wells come on stream, further increases in water production are anticipated. Total dissolved solids (mainly sodium bicarbonate) in Fruitland Formation water range from 1,000 ppm in the northwestern portion of the basin where fresh water recharge has occurred, to more than 20,000 ppm toward the center of the basin. These levels are too high for surface discharge of the water without treatment.
Underground injection is the most widely used method to dispose of produced water, accounting for over 95% of the water produced, with the remainder being evaporated. Disposal wells are a known technology that is acceptable to regulatory agencies. The main problem with disposal wells is their cost, ranging from $400,000 to $1,200,000, depending on depth and stimulation type. This leads to a total disposal cost for underground injection cost averaging approximately $1.00 per barrel, with a wide range from about $0.50 to $1.50 per barrel. The high capital cost is a deterrent for small operators, and for operators with only a few wells in a particular area. In addition, there is a risk that disposal wells will not have the capacity to accept all the water an operator wishes to dispose of. Treatment to appropriate permitting and environmental standards may allow surface use or discharge.
Water Production Analysis and Forecast
Figure 1 shows historical gas and water production from the basin, along with well count and water injection. 2,041 Fruitland coalbed methane wells had produced through mid-1992. The majority of the gas (84%) was produced from the New Mexico portion of the basin, although 67% of the water came from Colorado. Figure 2 shows that much of the gas has been produced from two townships in New Mexico (T30N, R6-7W). Water production (Figure 3) has been high over most of Colorado, and much less in New Mexico except for the two high gas producing townships. The cumulative water-gas ratio (Figure 4) shows the two high gas rate townships in New Mexico have low water-gas ratios, similar to the townships around them.
The water production forecast was based on extrapolation of historical production trends. The average rate per producing well was computed for each township, and extrapolated using an exponential decline. Many townships in New Mexico are showing rapid declines in water production (up to 70% per year decline), indicating the wells may be interfering with each other. Other townships are showing little or no decline; in those locales, a minimum 10% decline was assumed for the base case water production forecast.
Scale formation is a major problem with oil and gas production, particularly in the Continental United States. As reserves are particularly in the Continental United States. As reserves are depleted, the ratio of brine per barrel of oil, or gas equivalent, increases. Accurate prediction of operating conditions, under which scale formation will occur, is still needed. Most scale prediction algorithms are based upon pH and brine chemistry measurements, which are either difficult or impossible to obtain at high temperatures and pressures of gas and oil wells. In the present paper, a saturation pressures of gas and oil wells. In the present paper, a saturation index for calcite formation is developed and tested with field data. In the development of the index three issues were addressed which have not been previously included in an easily useable scale prediction algorithm: 1. the variation of the mole percent carbon dioxide with pressure and gas:brine:oil volumetric ratios; 2. the fugacity pressure and gas:brine:oil volumetric ratios; 2. the fugacity coefficient of carbon dioxide in a mixed gas: and 3. for most wells, the change in saturation index relative to shut-in downhole conditions is shown to be independent of all chemical parameters of the system, i.e. only a function of normally measured physical parameters. In addition, equations are developed to calculate the brine pH.
Production decline curves of three representative low permeability gas wells in the Piceance Basin are analyzed. These wells produce from the Mancos "B", Mesaverde and Wasatch formations, respectively.
It was found that long term production in these wells could be approximated using linear flow equations. This observation leads to the development of a decline curve method for predicting rate-time behavior based only one or two years of production data. The method is easy to apply and requires only data which is routinely reported to state oil and gas regulatory agencies. This type of data is public information and is readily available in most states.
The observed long-term linear flow behavior indicates that fracture lengths are much longer than would be expected from hydraulic fracturing treatments. Possible explanations for this behavior are discussed. The possibility of using short-term test data to define the long-term possibility of using short-term test data to define the long-term production decline curve is also discussed. production decline curve is also discussed
Production histories of fractured low permeability gas wells in the Piceance Basin are characterized by a sharp initial decline followed by a Piceance Basin are characterized by a sharp initial decline followed by a long transition into exponential decline. These two decline periods correspond to linear and pseudo steady-state flow, respectively. Predicting rates and reserves based on test data or short production Predicting rates and reserves based on test data or short production histories is difficult using conventional decline curve analysis.
The usual approach to predicting reserves by decline curve analysis, in this type of well, is to arbitrarily assign a high exponential decline rate for the first two or three years, followed by a lower decline. Another approach is to find a hyperbolic decline curve to fit the early tine data and extrapolate to estimate future rates. Both of these approaches can result in large errors in calculated reserves.
Early production data of low permeability gas wells should exhibit some linear flow characteristics. The constant pressure solutions for fractured wells presented in the literature predict that linear flow (the half unit slope) ends at tDxf=0.015. For reasonable values of fracture length and reservoir rock properties, linear flow should end in a few days or months, yet actual field data shows apparent linear flow lasting four years or more in some wells. Therefore, if fracture type curves are used to project well rates, based on a match that does not properly describe the linear flow system the predicted rates and reserves may be high.
The basic problem, therefore, appears to be one of reservoir description. Long natural fractures, long narrow reservoirs or thin high permeability beds can all result in long term linear flow behavior. permeability beds can all result in long term linear flow behavior. Hydraulic fractures are superimposed on the natural system. Short-term testing may characterize the hydraulic fracture, but long-term production may be controlled by the naturally occurring reservoir properties. This paper proposes a decline curve analysis technique for wells exhibiting paper proposes a decline curve analysis technique for wells exhibiting long-term linear flow characteristics and explores some possible explanations for this phenomenon.
Three wells which product from typical low permeability gas reservoirs in the Piceance Basin were selected for analysis. Case I is a Mancos "B" well in the Douglas Arch area, Case II is a Mesaverde well near the DOE MWX site, and Case III is a Wasatch well near Rifle, Colorado. The general well locations are shown on Figure 1.
Only readily available information was considered in this analysis.
The Wilson Creek field is situated in mountainous terrain of Rio Blanco County, Colo. It was discovered in early 1938 and produces from both the Morrison and Sundance sands of Jurassic age. The underlying Sundance reservoir is fully developed with 18 wells; no dry holes have been drilled. The overlying Morrison reservoir of greater areal extent is almost completely developed with 21 producing wells and three dry holes around the periphery of the field; in addition, all or most of the 18 inside Sundance wells will eventually serve to exploit the central portion of the Morrison reservoir.
The Morrison reservoir was moderately undersaturated with an initial average pressure of 2,180 psig and a bubble point pressure of about 1,100 psig; a partial water drive is in effect. A top structure gas injection program was started early in the life of the pool, resulting in a dampened pressure decline rate and increased oil recovery.
The Sundance reservoir is greatly undersaturated and exhibits a prolific water drive. Individual well fluid capacities are generally high and edgewater-type encroachment is evident throughout all producing levels. A down-flank gas injection project has recently been instigated.
The field is completely electrified and power is generated in a field plant. The operators own and maintain an 18-mile pipeline through which over 7,000 BOPD may be shipped by gravity flow. Major operational problems are related to extended winter periods of extreme cold and deep snow, rough terrain, paraffin accumulations, water disposal, and high water:oil and gas:oil ratio control.
The Wilson Creek field is located on the northern edge of Rio Blanco County, Colo., about 25 miles southwest of Craig, the only railhead in the area, and 10 miles north of Meeker. For geographical orientation see Fig. 1. The local terrain is mountainous, being in the more rugged portion of the regional Danforth Hills anticlinical trend. It is readily accessible year-around via well-maintained county roads connecting the local federal and state highway system.