Joint History Matching of Production, Tracer, and 4D Seismic Data in a 3D Field-Scale Case Study

Abstract

To improve the reliability of reservoir models, it is essential to condition reservoir models on available field data sets and reduce uncertainties through a history matching process. There are different types of field data that one can use to estimate uncertain reservoir model parameters. Among them, production data are the most used in history matching, but others also provide valuable complementary information. In this work, we take inter-well tracer and 4D seismic data as the extra sources of information for their high potentials for improving the understanding of reservoir heterogeneity, identifying drainage patterns, improving sweep efficiencies, and so on. However, in practice, it remains challenging to simultaneously history-match multiple field data sets in a proper and consistent manner. This study can be considered as a step towards addressing this problem. To this end, we propose an integrated ensemble-based history matching workflow, in which reservoir models are conditioned simultaneously on production, tracer and 4D seismic data with the help of three advanced techniques: adaptive localization (for better uncertainty quantification), weight adjustment (for balancing the influence of different types of field data), and sparse data representation (for handling big data sets). The history matching workflow is implemented and tested in a 3D benchmark case, and its performance is investigated through a set of comparison studies. Our studies indicate that jointly history matching production, tracer and 4D seismic data results in better estimated reservoir models, in terms of both improved data match, and reduced model discrepancies. Furthermore, we show that with the help of the correlation-based adaptive localization scheme, we are able to maintain substantial ensemble variability even in the presence of multiple types of field data, which appears beneficial to achieve a better performance during the forecast period. Overall, utilizing more types of field data can lead to extra performance improvements, which, however, is achieved at the cost of increased complexity of the history-matching workflow.