A hybrid GBPSO algorithm for permeability estimation using particle size distribution and porosity

Abstract

Particle size distribution measurements can be used for permeability estimation, and it is widely accepted that there exhibits a certain degree of correlation between permeability and porosity. In this paper, an efficient, low-cost, and reliable approach is used to develop an empirical correlation for estimating permeability based on particle size distribution characteristics and porosity in two modes: mode #1 includes 5% (D5), 10% (D10) and 60% (D60) of the cumulative passing particle size distribution curve and porosity for situations where porosity is known, and mode #2 where porosity is unknown. To optimize the coefficient of the proposed relationships, genetic-binary particle swarm optimization algorithm is used. A database consisting of 50 samples collected from four wells drilled in two neighboring pads in Western Canada were used, and their permeability values were predicted successfully. A validation based on a reference study and an application of sand completion design based on the finding of this study are also discussed. The novelties of the proposed approach are examining the effect of fines content, investigating the full range of particle size distribution curve, and using a hybrid intelligent method to optimize the coefficients of the correlations. In addition to sand completion deigns purposes, the proposed method can be used in enhanced oil recovery studies, reservoir management, and reservoir simulation applications.