Development of a new workflow for pseudo-component generation of reservoir fluid detailed analysis: a gas condensate case study

Abstract

In this work, a new automatic workflow for accurate optimal pseudo-component generation from gas condensate mixtures with a large number of components is presented. This workflow has a good insight into thermo-physical and critical properties and introduces only a small amount of loss of information and EOS flexibility. In this regard, the fuzzy clustering is used to classify the components in the mixture based on the similarities in the critical properties. The mixing rules are then applied to find group properties. Two different approaches for components association in clustering process are investigated with several numbers of groups. The mathematical validity of the groups is controlled with a proper validity index. The fluid phase behaviour is analysed to investigate the proposed workflow under physical feedback for different numbers of groups. The comparison of equilibrium calculations, for the extended and grouped mixtures shows a close agreement. The average absolute deviation percent (AAD%) from the extended analysis for the liquid dropout percent in constant volume depletion reaches to 0.32 for 20 groups and 0.93 for 14 groups. The AAD% for the gas compressibility factor over the pressure steps is 0.00 for 20 groups and 0.04 for 14 groups.