Reducing Uncertainty in Reservoir Modelling Through Efficient History Matching

摘要

Inverse methods play an important role when the characterisation of spatial distribution of petrophysical properties is necessary. Using the inverse procedure it is possible to reduce the uncertainty in reservoir characterisation by integrating different types of data combining probabilistic (geosta-tistical simulation) and physically-based (reservoir fluid simulation) modelling. One characterises the distribution of field permeability in oil reservoirs through the geostatistical history matching procedure. First of all, a set of equiprobable realizations of the reservoir properties is generated using direct sequential simulation (DSS). The field permeability is obtained and updated iteratively with successive co-simulations, preserving geological continuity, until a realization that matches both the stochastic model (variogram and histogram) and production data (i. e. fluid rate) is found. The algorithm proposed shows the ability to produce a good match with a small number of simulations.