Facies Simulation in Practice: Lithotype proportion mapping and Plurigaussian Simulation, a powerful combination

Abstract

4 Abstract Geostatistical simulation of facies has become part of the mainstream workflow when building stochastic geocellular models. However, modelers are presented with a plethora of challenges when attempting to produce models based on real data, including honoring depositional facies boundary conditions and their proportions, honoring data in the presence of numerous or closely spaced wells, capturing post depositional overprinting, and accounting for non- stationarity. These challenges often require unavailable tools and even the lack of modeling skills. Although the technology has evolved during the past 60 years and many sophisticated techniques exist, only of a few of the methods are available in commercial software. The most commonly used facies modeling algorithms tend to satisfy some but not all of these issues, but fall short on many more. Sequential Indicator simulation, the most popular method, lacks the ability to honor facies boundary conditions, Truncated Gaussian Simulation handles simple facies transitional boundaries, and while Object Simulation manages most non-overprinted complex facies sets, but it is unstable in the presence of numerous or closely spaced wells. Non-stationarity compounds the problem and a simple detrending of the data often results in an unsatisfactory solution. One powerful combination of methodologies is the use of a Lithotype Proportion Matrix with pixel-based facies simulation algorithms. There are a numerous advantages to this