Efficient Semi-Analytical Engine for Performance Prediction of Unconventional Reservoirs

Abstract

Abstract Reliable and fast prediction of fractured well performance is vital for asset evaluation and prudent field development decision making in unconventional, low permeability oil and gas reservoirs. Simple low cost techniques such as decline curve analysis are ill suited of properly accounting for extended transient flow regimes. Numerical simulation techniques on the other hand are unrivaled in capturing geological details and high accuracy in fluid flow description. However, they suffer from high computational expense and often too little relevant data is available. Our objective is to develop an efficient semi-analytical solution that combines the advantages of speed and accuracy from both techniques for forecasts on single well, pad and even field level. Contrary to conventional oil and gas reservoirs, the tight reservoir characteristics allow for the application of analytical solutions to predict the behavior of fractured wells. The depth of investigation is small with spatial property variations having limited influence considering the large local pressure gradients. Fluid flow is mainly of transient nature and less dominated by multi-phase flow effects within the reservoir. Moreover, the typical configuration of an individual well with multiple fracture stage, entire pads with multiple wells or even entire assets allow the application of the superposition in space, providing a cheap means for accurate solution for multi-well problems. This paper presents a transient semi-analytical model based on a tri-linear formulation. It accurately captures fluid flow to a single or multiple fractured wells over the entire life cycle of production, from early linear flow to pseudo-steady boundary driven flow. It is applicable to oil or gas reservoirs with permeabilities ranging from nano-Darcies to approximately one milli-Darcy. It is capable of handling both the vertical and horizontal well type, with either single or multiple fracture stages respectively. Well-bore, stage and pad configurations can be freely parametrized. A tubing head pressure boundary condition is incorporated to model production from those wells more realistically. For shale type reservoirs, a desorption process is incorporated. The model is particularly suitable for quantitative evaluation of field development alternatives involving a large number of fractured wells. The accuracy compared to a high resolution numerical simulation model generally exceeds 90% with a computation speedup factor in the order of one hundred or more. Besides conventional benchmarking with a numerical reservoir simulator, actual field production data from the Marcellus and Eagleford shale wells shows the utility of the new solution. Being both fast and accurate, the technique presented in this paper is ideal for supporting high quality, cost efficient decision making which is crucial in the current low oil price environment.