Unconventional oil reservoirs, such as Bakken, have gained considerable interest in recent years because they have become a great resource to produce oil and gas to meet the energy needs of North America. Performance prediction from these tight reservoirs is a challenge because of the complexity of reservoir flow, well completion, and fracture stimulation techniques. Elm Coulee field, in Bakken, is an example of such unconventional reservoirs and is located in Richland County, Montana. The field was drilled using both vertical and horizontal wells, but in recent years the use of horizontal wells has become the standard practice. The objectives of this study were: (1) evaluate the long-term (7 to 10 years) performance of horizontal wells in Montana Elm Coulee, (2) develop a better understanding of how to predict the long-term performance of younger Bakken fields in North Dakota based on the Elm Coulee experience.
Arps hyperbolic decline curve analysis was used as the main forecasting approach. In Arps analysis, q(t) = qi (1 + bDit)−1/b, where q is the flow rate, D is the decline rate, and b is the decline exponent. It will be demonstrated that forecasts using a constant b overpredicts well performance. To match the long-term performance of Elm Coulee wells, the numerical value of b had to be decreased with time.
Analytical approaches (log-log type-curve diagnostic plots and the Fetkovich log-log normalized plot) were also used to decipher the flow regimes, and to determine the varying decline rate from long-term producing wells in Elm Coulee Field. In addition to analytical modeling, numerical modeling was also used because it is more comprehensive in utilizing a larger set of reservoir parameters such as reservoir heterogeneity variations. This is very useful in transferring what we learned from the long-term performance of Elm Coulee Montana wells to the short-term performance of wells in North Dakota by addressing both geology and reservoir property differences between these fields.