The Gibbs-NVT ensemble Monte Carlo method is used to simulate the liquid-vapor coexistence diagram and the simulation results of methane agree well with the experimental data in a wide range of temperatures. For systems with two components, the Gibbs-NPT ensemble Monte Carlo method is employed in the simulation while the mole fraction of each component in each phase is modeled as a Leonard-Jones fluid. As the results of Monte Carlo simulations usually contain huge statistical error, the blocking method is used to estimate the variance of the simulation results. Additionally, in order to improve the simulation efficiency, the step sizes of different trial moves is adjusted automatically so that their acceptance probabilities can approach to the preset values.

Keywords:
fluid modeling, reservoir simulation, artificial intelligence, trial move, gibbs-nvt ensemble, mc method, monte carlo molecular simulation, phase equilibria, machine learning, upstream oil & gas
Subjects:
Fluid Characterization, Reservoir Simulation, Fluid modeling, equations of state
Copyright 2011, Society of Petroleum Engineers
You can access this article if you purchase or spend a download.