Tight reservoir oiliness numerical simulation based on a Markov chain Monte Carlo (MCMC) method: A case study of the upper Triassic Yanchang-6 formation (T3ch6 Fm.) outcrop of Ordos Basin

Abstract

Numerical simulation of the tight reservoir oiliness has always been an important issue in tight oil resources exploration. It is hard to take some special situations into simulations such as the micro-nano thermodynamic system, the multi-fluid phase behavior, and the superimposed tectonic stress field of tight oil reservoirs. In our research, we use MCMC methods to simulate the tight reservoir oiliness constrained by some visualized, spatial-related and interacted geological parameters based on the statistic Bayesian theory, instead of taking the complicated accumulation mechanism into consideration. In this research, we built a posterior probability density distribution (Post-PDD) matrix using the spatial relation between oiliness, fracture, and lithofacies parameters of a typical Yanchang tight oils outcrop. In the simulation, one Monte Carlo Step (MCS) was settled N = 80 × 36 × 5000, where 80 and 36 represent the horizontal and longitude grid number respectively, and 5000 represents the sampling times for each grid. In total, simulations of 50000 MCSs were made for each realization. For the simulation results, the variation in the fluctuating trends of the fitted and actual values matched well with the analytic variogram for the MCMC realizations. Additionally, we found that the simulation uncertainty was consistent with the reservoir heterogeneity. Even though the lamina, intercalation or the barrier was not defined in the grid model, the oiliness realization of the different types was remarkably consistent with the reservoir heterogeneity, which may be caused by the sedimentary rhythm similitude between different scale sedimentary cycles, particularly in the lacustrine tight reservoir.