Improved Production Profiling Using Thermal Balance and Statistical Modeling in the Pinedale Anticline of the US Rocky Mountains

Abstract

The Pinedale anticline is located in the Green River Basin of Southwestern Wyoming, USA. The field is the largest tight gas discovery for the onshore region of the United States in the last twenty years (Robinson and Shanley 2004). Gas production is from very tight, stacked clastic reservoirs that are Upper Cretaceous in age, with productive intervals in excess of 6000 feet. The large productive intervals require multiple hydraulic fracture stages to complete. Time-lapsed production analyses are performed to optimize well spacing and to characterize the gas bearing reservoirs. Production logs are also run to determine the effectiveness of the hydraulic fracturing and to identify water entry points that may lead to premature completion failures. Typical wells produce relatively small amounts of water, usually less than 5 percent by volume. It is nearly impossible to detect such small watercuts with conventional methods of production analysis. However, a probabilistic production analysis method simultaneously modeling flowmeter and temperature can take advantage of the high contrast between the heat capacity of gas and water and therefore provide good estimates of the water and gas production profiles, even in small watercut wells. This paper describes the technique used to improve the production flow profiling, supporting its assertions with case study results. Introduction Productive intervals in the field are stacked-lenticular tight sands with porosity ranging from 6 to 12% and permeability in the submicrodarcy to 20 microdarcy range, with an average value of 4 microdarcies. Water saturations vary from 30 to 60%, with comparably low water production. Condensate ratio with an API gravity of 52 is 8 to 10 bbl/MMscf (Eberhard and Mullen 2003). Numerous faults exist in the region, adding to the complexity of the reservoirs and creating over pressured gas zones in wells whose nearby offsets encounter normally pressure zones at similar depths. Pore pressure variability with depth does not follow a linear pattern, with intervals in the normal range bounded by layers in the geopressured zones. For this reason, individual production intervals must be hydraulically fractured in isolation with other intervals to assure effective treatment. Such complexities in the fracturing methodologies require an effective method of assessing flow contribution, both by phase and flow rate, of each productive layer. Additionally, the method used to estimate layer properties in multi-layer low permeability gas reservoirs (Spivey 2006) requires that accurate flow contribution be measured for each productive layer. Since the method requires multiple measurements over time as an input to the history match, consistent measurements and production log analyses are an absolute necessity. The combination of accurate production logging analyses and layer property determination provides the reservoir and production engineering teams with much more precise data than surface production data alone, maximizing the effective management of resources. With the cost of proppant representing as much as 30% of total completion cost (Huckabee et al 2005), accurate production analysis of the fracturing effectiveness can enable significant cost savings. Production Logging Methodology A typical completion (figure 1) consists of up to 24 frac stages each of which may have six perforated intervals ranging from 2 to 6 feet in length. Perforating shot density is 3 shots per foot. A 7 inch protection casing is set 600 feet from TD, with 4 ½' casing set from TD to surface. Flow-through composite fracture plugs (Eberhard, et al 2003) are used to isolate hydraulic fracturing treatments, then drilled out prior to production.

Enhanced Production Profiling of Gas Wells Using a New Method to Model the Geothermal Temperature Gradient

“…These theoretical responses are then compared to the actual measured data and the well model is modified until the best possible match is achieved. The model is optimized through an iteration process to minimize the error between the theoretical and measured data for all depths of the wellbore to produce the most accurate production profile (Donovan et al 2008).…”