Quality criteria for automated history matching of hydrocarbon reservoir models

Eremyan G.A. eremyanga@hw.tpu.ru Petroleum Learning Centre of Tomsk Polytechnic University Tomsk
Rukavishnikov V.S. Petroleum Learning Centre of Tomsk Polytechnic University Tomsk
DOI: 10.24411/2076-6785-2020-10117

Abstract
The paper considers identifying of high-quality models obtained after automated history matching of hydrocarbon reservoir models. Numerical reservoir modeling consists of three main stages: creation of a static geologic model,
construction of a dynamic filtration model on its basis and an iterative process of history matching of the created reservoir model. Based on the selected history matched models, predictive simulations are made to estimate the
forecast uncertainties and associated risks. The purpose of this work is to study the quality criteria for the history matched models and the validity of their use on the example of a synthetic reservoir model. The results of this study
show disadvantages of using the value of the objective function as the only quality criteria. Also, ways are indicated for a more reasonable selection of desired quality models.

Materials and methods
The main methods in this study are numerical reservoir simulation, comparative analysis and computational experiment using a synthetic reservoir model.

Keywords
history matching quality, quality criteria, objective function, optimization algorithms, automated history matching, reservoir simulation, oil field, numerical model
Download article