An Analytical Model for History Matching Naturally Fractured Reservoir Production Data
Conference Paper
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
AbstractA method to estimate reservoir properties and predict future performance using production data for naturally fractured reservoirs is presented. The method uses an analytical solution describing dual-porosity reservoirs with constant bottomhole pressure production in both infinite-acting and finite reservoirs. The solution is rigorous for slightly compressible liquid flow, and it is modified using a normalized time to model gas flow in finite reservoirs.A method is presented for analyzing production data. The method includes procedures for validating the dual-porosity model with production data, determining the suitability of simpler reservoir models, and determining the "best" parameter estimates for the selected model. Example applications using field data from naturally fractured reservoirs are presented.IntroductionNaturally fractured reservoirs are typically characterized as having two distinct porosity systems a primary system associated with the reservoir matrix and a secondary system associated with the fractures. An idealize 1 reservoir model developed by Warren and Root, in which the reservoir is represented by a regular fracture system that is superimposed upon the primary porosity system, has formed the basis for many investigations into the performance of naturally fractured reservoirs. This dual porosity model is also adopted in our study.A key exercise in reservoir engineering is forecasting, or predicting, the response of a producing well. Production data can be an important source of information for determining reservoir properties for use in simulations of reservoir behavior and well response. We address here the estimation of reservoir properties from production data for naturally fractured reservoirs, and the subsequent forecasts of well response.Finite-difference reservoir simulators configured as dual-porosity reservoir models, can be used to analyze production data. However, analytical solutions may be obtained for certain idealized situations. In such cases, analytical solutions provide a number of advantages compared to finite-difference simulators. Typically, they can provide solutions comparable to finite-difference simulations in a fraction of the computing time. Grid breakup, which normally impacts the accuracy and stability of finite-difference simulators, is not required for the analytical solutions. Consequently, data input for a reservoir simulator using analytical solutions can be greatly simplified compared to finite-difference simulators.
