A new objective function designed for the calibration of stratigraphic forward models

Abstract

This paper presents a novel and quantitative method of comparing facies successions of distinct sizes and resolutions, adapted for the automated calibration of stratigraphic forward models (SFM). The method, which corresponds to the calculation of an objective function, consists of correlating time-sampled SFM results with space-sampled well logs using a strategy inspired by methods of automated well log correlation. The method was designed to be generic and to accept multi-variate definition of facies, so that it can be applied in a variety of geological contexts (e.g., mixed carbonate-clastic environments). The method demonstrated properties suitable for automated calibration, such as a smooth and almost convex relationship with individual simulation parameters. The method was applied to a case-study using gradient-based, meta-heuristic, Bayesian, and random optimization strategies in an 8-dimensional uncertain space. The results show that calibrated models can be found, but there are still non-uniqueness issues due to compensation effects between uncertain parameters. Therefore, a calibration strategy capable of identifying these multiple solutions is necessary to propagate geological uncertainties in petroleum system assessments and field production predictions.