The Marlim Field: Incorporating Time-Lapse Seismic in the Assisted History Matching

Abstract History matching is one of the most important and time consuming tasks in reservoir simulation. In traditional history matching approaches engineers attempt to honor production data by manually modifying their simulation models. Even though such approaches are being gradually replaced by assisted history matching methods the engineer is still responsible for choosing which simulation model parameters must be modified. On the other hand optimization algorithms available in assisted history matching tools allow changing parameter values automatically and also to consider 4D survey data in the objective function. Transmissibility multipliers associated to regions located between wells are among the most traditional simulation model parameters used for history matching purposes. Such procedure is very efficient in order to match water production data, but usually leads to transmissibility maps which have no geological meaning. This paper describes the results of the assisted history matching study of Marlim Field, the biggest offshore field in Brazil. The assisted history matching presented here included both production and 4D seismic. The seismic information was used qualitatively and quantitatively, i.e., to delimitate regions associated to transmissibility multipliers and also in the objective function itself. The assisted 4D history matching was carried out with a commercial history matching tool, which allows the synthetic seismic calculation. The production history data of most part of the wells was matched with similar accuracy one would expect to obtain through a careful and time consuming manual history matching approach. However the total elapsed time achieved here was orders of magnitude smaller. In addition the shape of the regions associated with model parameter values was totally supported by 4D seismic information. It turns out even with the aid of assisted history matching tools, the quality of the results strongly depends on adequately choosing which simulation model parameters must be modified in order to match the production data. Once 4D seismic data can give a realistic picture concerning reservoir fluid replacement it can also highlight preferential permeability pathways. The results of the present study have shown these indications can be very helpful to delimitate regions with history matching purposes that have geological significance.