Investigation of Matrix-Fracture Mass Transfer by History Matching

Two common approaches to model naturally fractured reservoirs are Discrete Fracture Network and the Dual Continuum Models. The main Problem of DFN method is it cannot be applied for real fieldexamples because of computational costs and the main challenge of dual continuum method is thedifficulty to define appropriately the interaction between matrix and fracture, i.e. transfer function. The recovery curve technique decouples the local matrix depletion calculation from the regional fracture flow and can be considered as a compromise between the two approaches. This method is believed to describemore credibly the reality of the matrix-fracture interaction. In this method the recovery curves aredetermined for water imbibition and gas gravity drainage processes and then are used to describe matrixfracture communication. This paper illustrates an approach to determine the recovery curves through history matching. A column model from a real NFR is extracted and by matching the movements of the phase contacts in the model, recovery curves are determined. The determined recovery curves can be employed to forecast the production. Hybrid Genetic Algorithm is used as the optimization tool. The results show very good agreement with the history of the model and confirm that the proposed method has the capability to determine the heart of naturally fractured reservoir simulation, i.e. matrix fracture masstransfer

Keywords:

Naturally Fractured Reservoirs , Dual Porosity , History Matching , Matrix-fracture Mass Transfer , Genetic Algorithm