Stochastic History Matching of core flooding experiments on carbonate samples

History Matching and uncertainty quantification are two important aspects of modern reservoir engineering studies. Finding multiple History Matching models for uncertainty quantification with minimum/lowest misfit is one of the focus of this research in Assisted History Matching methods (AHM). In this paper, the laboratory coreflooding analysis on representative reservoir rock samples (particularly carbonate rocks); and translating them into a real reservoir scenario through numerical modelling of reservoir coupling stochastic History Matching using the Laboratory results to reduce the uncertainty in porosity and permeability of the carbonate rocks was the key focus of this thesis. The study goals and objectives of this research is to perform a laboratory experiment on carbonate rocks and find their fluid saturation and the residual oil saturation, permeability and porosity values, the pressure values which can be used in the numerical simulations. The potential of Improved Oil Recovery (IOR) on carbonate rock samples by low salinity (brine) coreflooding has been investigated through both the laboratory measurement and the fluid flow simulation. Results show that both investigations indicated that the potential is high as it can be seen from the Oil Recovery which was between 38% - 42% percentage and this is reasonable for a secondary recovery coreflooding with brine. Ageing done on sample K4 shows that ageing increases the oil recovery during water coreflooding and aged samples provided the best History match result with the best misfit. The Assisted History Matching (AHM) method used raven epistemy software which enabled a more fast and reliable process. The History Match parameters for the production history from numerical modelling and from the laboratory experiments of hours of pressure build up test produces nearly identical effective reservoir permeability of 0.01 - 0.02mD and 0.146mD, therefore, we can conclude that there is a presence of interconnected micro-fractures as the main contributing factors. Using the Particle Swarm Optimization Algorithm (PSO), a lower misfit provides a better match of the production history and also, the sampling behavior of the optimization algorithm has a direct impact on the prediction/forecast. For a heterogeneous Model, using higher particle number and a higher iteration provides more realistic results and gives a better understanding of the result because the convergence can occur at a higher iteration. Uncertainty Quantification using stochastic - Particle Swarm Optimization Algorithm seems to be a good approach for predicting and forecasting models in the Reservoir modelling field. Finally, the uncertainty quantification in porosity and permeability obtained from laboratory experiments where these uncertainties were optimized through a Stochastic History Matching using Particle Swarm Optimization Algorithm (PSO), also shows that given a heterogeneous system, the best misfit provided results of the spatial distribution of permeability and porosity of close range to the values from the laboratory results.