Parallel Superiorization Method for HistoryMatching Problems Using Seismic Priors
History Matching, Superiorization Method, GPU, ASIC
History Matching is a very important process used in managing oil and gas production, since it aims to adjust a model of a reservoir until it closely reproduces the past behavior of a real reservoir, so it can be used to predict future production. This work proposes using an iterative method for constrained optimization called the superiorization method to solve this problem. The superiorization method is an approach that uses two optimization criteria, where the second criterion seeks to optimize its functional without negatively affecting the optimization of the first. As a comparative approach, a genetic algorithm was chosen since it is widely used in the literature to resolve history matching. Since the addressed problem is an inverse problem that is usually severely underdetermined, several possible solutions may exist for its resolution. Because of this, we also propose the use of seismic data from the reservoirs to reduce the number of possible results by enforcing a meaningful regularization on the second optimization criteria of the superiorized version of the ML-EM algorithm. Another critical factor in these history matching problems is the simulation time, which is usually high. Thus, we also propose to investigate two parallel approaches by using GPUs and ASICs. This proposal contains some preliminary results for a synthetic slab model experiment, adjusted by both the superiorized ML-EM algorithm and the genetic algorithm. Is these experiments, the superiorized ML-EM algorithm was shown to be more efficient relative to execution time while also decreasing the history matching error rate, when compared to the genetic algorithm.