Bayesian probabilistic dual-flow-regime decline curve analysis for complex production profile evaluation

Abstract

Decline curve analysis (DCA) has been broadly applied in the oil and gas industry to predict future production and estimate reserves. The Arps decline model is one of the most popular DCA models used by the oil and gas industry. This model was designed for boundary dominated flow in conventional reservoirs. In practice, it has also been widely applied to both transient flow and boundary dominated flow in conventional reservoirs. Applied to a complex production profile involving dual flow regime from artificially enhanced tight/shale reservoir, each production segment needs to be analyzed separately. Although several analytical decline curve models have been developed in recent years to address the problem, none of them is flexible enough to model both production profiles from a single flow regime and a dual flow regime.In this study, we introduce a new production decline analysis model that can fit both production profiles from single flow regime and dual flow regime without the need to separate the production segments. We also introduce a probabilistic measure using Bayesian Markov Chain Monte Carlo simulation to determine the existence of dual flow regimes, the location of the regime switch, and uncertainties in parameter estimations with the new model. The proposed Probabilistic Dual Regime Decline Curve Analysis (PDR-DCA) method is tested on 344 horizontal Northern Montney gas wells.This proposed Bayesian approach provides a workflow to generate probabilistic decline curve for both single and dual segment production profile, probabilistic type well curve for a well group, and production prediction for well with very short production history.