Use of Deconvolution for Pressure Transient Analysis in Layered Reservoirs – A New Method

Abstract

Many reservoirs exhibit multilayer behavior. Multilayer oil/gas reservoirs are usually classified as (1) systems with formation crossflow and (2) systems without formation crossflow (commingled systems). The focus in this work is the analysis of pressure transient data of commingled layers. Pressure transient analysis for multilayer reservoirs to estimate individual layer properties is usually difficult and suffers from many limiting assumptions. The available interpretation methods are usually based on a subjectively presumed model (usually homogenous and isotropic layers, and radial infinite reservoirs) that has many unknown parameters (e.g., permeabilities and skin factors) to be estimated through a history matching process. The individual layer properties obtained from such analysis methods may not be reliable. The reliable way to evaluate individual layers’ characteristics is to isolate and test each layer separately, which is challenging due to high costs and occasional operational constraints. In this work, we suggest a testing procedure and an analysis approach to analyze well test data of commingled reservoirs that allows reliable characterization of individual layers. The approach benefits from modern deconvolution techniques to eliminate the rate variation (rate transients) effects from the bottom-hole pressure signal acquired during the test. The methodology presented does not make assumptions about the individual layer reservoir models and recovers distinct pressure signals of the individual layers. The recovered pressure signals of each layer are also free of wellbore storage (WBS) effects. In addition, the individual layers’ pressure signals are stretched over the whole test duration (both drawdown and buildup periods). Successful application of deconvolution requires complete sandface rate data of the individual layers in the commingled system. However, the continuous measurement of individual layers sandface rates is usually not available. A simple model is introduced in this work to make good estimations of the layer rate profile using few measurements of the production logging tool (PLT). The developed rate profiles are then used in deconvolution and individual layer's pressure signal can be recovered for further analysis. The approach was verified against simulated cases with variety of layer models. The results obtained from the developed approach are in good agreement with the true solution. The findings of this study can be used to characterize commingled reservoir systems and determine the individual layers properties. It has applications in optimizing injection/production well performance of commingled systems.