Discrete multiscale modelling of fractures in natural fractured reservoir, a case study from a field in the Persian Gulf

The influence of fracture network on quality of reservoir, emphasize the importance of study of fractured reservoirs. The characterization of fractured reservoirs is complex. We are going to study fractures of one of the oil fields of Persian Gulf by discrete fracture network modeling. Fracture modeling is often based on very limited well data and therefore is subject to high uncertainty. Typically, the standard modeling workflow uses interpolation algorithms to predict the fracture spatial distribution. This paper shows an alternative workflow for improving fracture modeling between wells through the use of seismic attributes. The main objective of this paper is to compare fracture intensity models guided by the two approaches: The standard interpolation based approach, and the seismic based approach using attributes sensitive to faults. The difference between these two methods lies in the way fracture intensity is modeled. Fracture intensity is an important fracture attribute because it guides the fracture simulation. It is estimated from the fracture point data derived from the well data and upscaled into the model. Typically, the standard method employs Kriging or Sequential Gaussian Simulation (SGS) to interpolate the fracture intensity of the 3D grid. This method can deliver highly inaccurate results in case of limited well control. The second approach tries to reduce this uncertainty, through controlling the fracture intensity interpolation via seismic attributes. Seismic attributes can be used as secondary input for the interpolation of the fracture. The preferred interpolation algorithm is collocated co-Kriging because one has full control over the radius of influence of the well data. In addition the influence of the secondary input, the seismic attribute, is controlled via its correlation factor with the well data. A comparison of these two methods provides insight into the complexity and uncertainty involved in fracture modeling.