Geostatistical History Matching with Direct Sequential Transformation of Images

The integration of production data in the stochastic modeling of internal reservoir properties is usually approached by an inverse problem. The inverse methodology consists on an optimization process of perturbation of the fine grid values of permeability and porosity in order to match the synthetic response of the model with the real production data. But as the relationship between internal properties, that one wish to model, and the dynamic data is highly non-linear, the solution passes always for the exercise of running a dynamic simulator, to obtain a simulated response, each time one perturb the values of permeability and porosity. This can lead to prohibitive time consuming in industrial applications. Hence the challenge in this field has been focused in the use of fast simulators or more efficient history matching approaches. <br>In this study, a new approach of geostatistical history matching with a sequential perturbation of original variables (permeability field) is proposed. The field perturbation, is done by Direct Sequential Simulations and Co-simulations. Local models of co-regionalization allow for regional optimization of several producer wells responses.