A new approach with multiple realizations for image perturbation using co-simulation and probability perturbation method

History matching is an inverse problem with multiple possible answers. The petrophysical properties of a reservoir are highly uncertain because data points are scarce and widely scattered. Some methods reduce uncertainty in petrophysical characterization; however, they commonly use a single matched model as a reference, which may excessively reduce uncertainty. Choosing a single image may cause the model to converge to a local minimum, yielding less reliable history matching. This work improves on the history matching presented by Oliveira et al. ((2017a) J. Petrol. Sci. Eng. 153, 111–122) using a benchmark model (UNISIM-I-H based on the Namorado field in Brazil). We use a new approach for a Probability Perturbation Method and image perturbation using Co-Simulation. Instead of using a single image as the reference, a set of best images is used to increase variability in the properties of the reservoir model while matching production data with history data. This approach mitigates the risk of the potentially excessive reduction of uncertainties that can happen when using a single model. Our methodology also introduces a new objective function for water breakthrough, improving model quality because of the importance of matching the water breakthrough in the process. Our proposed methodology for image perturbation uses the UNISIM-I-H, which comprises 25 wells and has 11 years of history data. Our methodology made the process of calibration more effective than the history matching by Oliveira et al. ((2017a) J. Petrol. Sci. Eng. 153, 111–122). Cross-influence was minimized, making the history matching more objective and efficient, and consequently, the production forecasts more reliable.

Stochastic inversion of facies from seismic data based on sequential simulations and probability perturbation method

We presented a new methodology for seismic reservoir characterization that combined advanced geostatistical methods with traditional geophysical models to provide fine-scale reservoir models of facies and reservoir properties, such as porosity and net-to-gross. The methodology we proposed was a stochastic inversion where we simultaneously obtained earth models of facies, rock properties, and elastic attributes. It is based on an iterative process where we generated a set of models of reservoir properties by using sequential simulations, calculated the corresponding elastic attributes through rock-physics relations, computed synthetic seismograms and, finally, compared these synthetic results with the real seismic amplitudes. The optimization is a stochastic technique, the probability perturbation method, that perturbs the probability distribution of the initial realization and allows obtaining a facies model consistent with all available data through a relatively small number of iterations. The probability perturbation approach uses the Tau model probabillistic method, which provides an analytical representation to combine single probabilistic information into a joint conditional probability. The advantages of probability perturbation method are that it transforms a 3D multiparameter optimization problem into a set of 1D optimization problems and it allowed us to include several probabilistic information through the Tau model. The method was tested on a synthetic case where we generated a set of pseudologs and the corresponding synthetic seismograms. We then applied the method to a real well profile, and finally extended it to a 2D seismic section. The application to the real reservoir study included data from three wells and partially stacked near and far seismic sections, and provided as a main result the set of optimized models of facies, and of the relevant petrophysical properties, to be the initial static reservoir models for fluid flow reservoir simulations.