Insights of new-generation reservoir property modeling (MPS methods) in assessing the reservoir performance for different recovery methods

Usage of an advanced geostatistical method for reservoir characterization results in more reliable insights of reservoir performance. Multiple-point statistic (MPS) methods which borrow ideas of texture synthesis could mimic the variation of complex reservoir permeability. This study implements this complex variation of permeability for enhanced oil recovery (EOR) purposes. Thermal EOR which reduces the viscosity of reservoir could be nominated as one of possible EOR method for any reservoir. The effect of high-order statistics of permeability is studied against the variation of oil viscosity. To this end, a modified version of SPE 10 model-1 is used to apply a thermal EOR scenario. Constant rate water injection creates a front of water which moves through the reservoir based on synthesized MPS-derived permeabilities as well as the reference case. In this study, multi-scale cross-correlation-based simulation (MSCCSIM) and image quilting (IQ) are selected for generating reservoir permeability. The results show that synthesized permeability map is able to create a correct trend of oil production rate for whole reservoir life compared with the reference case. However, similar behavior at one oil viscosity value does not guarantee similar behavior of reservoir for another set of oil viscosity values. For highly viscose oil reservoir, reference permeability shows later water breakthrough than oil with normal viscosity while in synthesized permeability, water breakthrough happens earlier for oil with normal viscosity. This indicates the uncertainty of EOR scenario prediction due to imperfectness of MPS methods for characterizing the studied reservoir. Incidentally, as an important advantage, the reference general shape of water front and saturation spectrum is satisfactorily regenerated by one of MPS methods. Moreover, flanging and fingering of water front are mimicked to some acceptable extent.