An ensemble-based technique has been developed and successfully applied to simultaneously estimate the relative permeability and capillary pressure by history matching the observed production profile. Relative permeability and capillary pressure curves are represented using a power-law model. Then, forward simulation is performed with the initial coefficients of the power-law model, all of which are to be tuned automatically and finally determined once the observed data is assimilated completely and history matched. The newly developed technique has been validated by a synthetic coreflooding experiment with two scenarios. The endpoints are fixed for the first scenario, whereas they are completely free in the second scenario. Simultaneous estimation of relative permeability and capillary pressure has been found to improve gradually as more observation data is assimilated. There exists an excellent agreement between both the updated relative permeability and the capillary pressure and their corresponding reference values, once the discrepancy between the simulated and the observed production history has been minimized. Compared with coefficients of capillary pressure curve, coefficients of relative permeability curves, irreducible water saturation, and residual oil saturation are found to be more sensitive to the observed data. In addition, water relative permeability is more sensitive to the observation data than either oil relative permeability or capillary pressure. It is shown from its application to a laboratory coreflooding experiment that relative permeability and capillary pressure curves can be simultaneously evaluated once all the experimental measurements are assimilated and history matched.
