Numerical Simulation And Artificial Intelligence Techniques For Modeling Low Salinity Water Flooding (LSWF) In Sandstone Oil Reservoirs

Recently,low salinity waterflooding enhanced oil recovery(LSWF-EOR)has drawn the attention of many researchers due its additional oil recovery than high salinity waterflooding(HSWF).Previously,practitioners main focus was the quantity of water injected into petroleum reservoirs.However,the quality of water injected is as important as the quantity injected which has led to the emergence of LSWF-EOR technique.Furthermore,several studies over the past decades have focused mainly on the theoretical and experimental aspect of LSWF EOR method with the exact mechanism still poorly understood and debatable in the literature.Additionally,the ability of numerical reservoir simulators to accurately model and predict the underlying mechanism of this novel EOR process is still a challenging task.Therefore,this research work seeks to establish the prediction methods for this LSWF-EOR technique by applying computer programming,fluid flow mechanics and mathematical modeling approaches.The main contents of this research are:(1)To develop a LSWF-EOR numerical simulator to model the exact underlying mechanism in sandstone oil reservoirs by validating it with LSWF-EOR experimental data.(2)To develop a chemical tracer numerical simulator to evaluate the impact of reservoir heterogeneity on fluid flow performance through a newly developed geostatistical method of generating porous media(3)To develop a numerical reservoir simulator to model low salinity polymer(LSP)flooding impacts on ultimate oil recovery in porous media(4)To develop history matching models by hybridizing backpropagation artificial neural network(BPANN)and particle swarm optimization(PSO)to history match LSWF-EOR experimental data.The obtained results from this research work are as follow:(1)The exact EOR mechanism of LSWF-EOR mechanism establishes wettability alteration as the main primary mechanism.(2)The mathematical models for LSWF,LSP flooding and chemical tracer numerical simulation were established.Finite volume method and Newton Raphson iterative method were used to solve the numerical models developed.Also,numerical oscillations in the numerical simulators developed were eliminated by using total variation diminishing(TVD)schemes.Then,analytical methods were used to validate the numerical methods developed in this research work.(3)The three numerical reservoir simulators and the history match models(BPANN and PSO-BPANN algorithms)developed were implemented in Matlab programming environment.(4)The chemical tracer numerical reservoir simulator was used to simulate viscous fingering and preferential flow channeling subsurface instabilities commonly associated with water flooding schemes.Thereafter,LSP flooding was used to minimized the effects of these subsurface phenomena which improves oil recovery efficiency.The ion exchange model coupled to the LSWF numerical simulator was validated with PHREEQC geochemical software with excellent results.Additionally,from the obtained numerical results,p H was found to increase for LSWF while it decreases for HSWF.The physical and chemical mechanism of LSWF-EOR displacing oil in porous media was fully investigated with LSWF-EOR numerical simulator developed.The LSWF-EOR simulator and the history match models developed were used to simulate and interpret reliable coreflooding oil displacement data which gave excellent results.Finally,this research work has added a new novel insight into the understanding of LSWF-EOR mechanism in sandstone oil reservoirs for future full-scale implementation.