Reservoir Rock Permeability Assessment Using Images Analyses of Thin Sections and Intelligent Systems

Permeability is the ability of porous rock to transmit fluids. Permeability is one of the most important properties of reservoir rocks because oil production from a reservoir depends on the permeability of the reservoir. Permeability is determined using a variety of methods, which are usually expensive and time consuming. Reservoir rock properties can be determined using image analysis and intelligent systems to reduce time and money. This study presents an improved model based on the integration of petrographic data and intelligent systems to predict reservoir rock permeability. Petrographic image analysis was employed to measure the inter-granular porosity, inter-crystalline porosity, moldic porosity, micro porosity and optical porosity, amount of cement, limestone, dolomite and anhydrite, type of texture and mean geometrical shape coefficient. The permeability was predicted using individual intelligent systems including a neural network, a fuzzy logic, and a neuro-fuzzy model. The mean squares error of the neural network, fuzzy logic and neuro-fuzzy methods were, respectively, obtained as 0.0168, 0.0107 and 0.0095. Afterwards, committee machine based on simple averaging was used to combine the permeability values calculated from the individual intelligent systems. As a result, the mean squares error of the committee machine was obtained as 0.0072. These results showed that the committee machine performed better than neural net, fuzzy logic, and neuro-fuzzy models acting alone.

Keywords:

Permeability , Thin section image analysis , Intelligent systems