چكيده به لاتين
The primary production of oilfield are reducing in the world. These resources are mainly in carbonate fractured reservoir and mostly have oil-wet characteristics with high degree of fracturing. So the flooding processes do not work properly, therefore enhanced oil recovery methods must be used for oil production from these reservoirs. One of theses methods are surfactant processes that are recently paid attenetion for high oil prices. One of the surfactant processes is surfactant-aided gravity drainage that is applied for these reservoirs under water drive. The most important controlling forces in this process are gravity and capillary forces. Actually, the interaction between these two forces handle the reduction or addition in oil recovery. Gravity force plays a positive role, however, capillarity prohibits the flow in this process.
There are some unknown problems in surfactant-aided gravity drainage in the literature. For example, comprehensive influence of surfactant in flow performance, influence of different parameters on surfactant performance, and no proper physical simulation of static imbibition test that is done for evaluation of surfactant on the flow. In this work it is tried to overcome these limitations with numerical simulation models and surrogate model.
Numerical simulator (ECLIPSE) is used in analyzing the surfactant-aided gravity drainage with two different points of view. One is single matrix block analysis to investigate gravity drainage mechanism and the other one is multi matrix block analysis to investigate re-infiltration mechanism.
In both analayzes, the simulation model is imbibition test of surfactant that is done on carbonate oil-wet rock. Another model that is constructed in this work, is surrogate model based on multi-layered, non-linear, training neural network to produce recovery curves. This model is developed to replace numerical simulator for water-oil displacement.
The simulation models (SMB and MMB model) with the presented form are new, with three reasons. The first reason is that nearly for the first time, ECLIPSE is validated with experimental data to model surfactant-aided gravity drainage. The second is comprehensive analayzes consist of all impressed parameters on surfactant-aided gravity drainage. The third is proper physical simulation of imbibition test according to the experiment condition. The surrogate model laeds to evolution in multi-phase flow displacement, because without any installed simulator on computers, the recovery curves are produced with proper accuracy.
Keywords: gravity drainage, imbibition, surrogate model, re-infiltration, surfcatant
