Estimation of fracture aperture from petrophysical logs using teaching–learning-based optimization algorithm into a fuzzy inference system

摘要

Aperture, which refers to the opening size of a fracture, is a critical parameter controlling rock mass permeability. Moreover, distribution of permeability within the reservoir is commonly affected by natural fracture occurrences. In a water-based mud environment, borehole-imaging tools are able to identify both location and aperture size of the intersected fractures, whereas in oil-based environment, due to invasion of resistive mud into the fractures, this technique is impractical. Recently, some artificial intelligence techniques facilitated reliable estimations of reservoir parameters. In this paper, a teaching–learning-based optimization algorithm (TLBO) trained an initial fuzzy inference system to estimate hydraulic aperture of detected fractures using well logs responses. Comparing the results with real measurements revealed that the model can provide reliable estimations in both conductive and resistive mud environments, wherever the aperture size is unknown. TLBO, besides of its easier application, outperformed earlier optimization algorithms, which were used to evaluate the method effectiveness.