Reconciliation of Fracture Performance and Fracture Geometry with Design

Abstract

For more than 50 years a vision of a hydraulic fracture that is vertical with two symmetrical wings has been accepted by two highly diverse sciences in petroleum production: fracture mechanics and production/reservoir engineering models. The fracture in both sciences has a height, a length tip-to-tip, and an average width. A fracture propagation model is usually employed to determine fracture dimensions. Linear elastic fracture mechanics points towards a relationship between fracture length (and, implicitly, height) and fracturing pressure. Therefore, pressure analysis during execution is supposed to 1) determine the generalized fracture geometry and 2) to quantify fracture dimensions such as length and width. Once the well is put on production, the fracture geometry can be determined either through a well test or through long-term production data analysis or both. The models employed for this exercise have been in wide use and have been credited with considerable success in hydraulic fracture treatment evaluation. Pressure patterns would lead to the determination of the apparent fracture half-length and fracture permeability-width product. Unfortunately, often there is a discrepancy to a severe discrepancy in the fracture dimensions obtained from the two methodologies. We present here several theories describing the discrepancy and we quantify the impact of reservoir and fracture parameters. These include reservoir areal permeability anisotropy, damage to the proppant pack, discontinuity in fracture conductivity and, of course, turbulence effects. We are applying this approach to 23 hydraulically fractured wells in a less-than 1 md oil and gas reservoirs in Western Siberia and achieve reasonable reconciliation between the results of the diverse methods of fracture geometry determination and the impact of reservoir and fracture variables. Fracture treatments, designed for other wells in the field take these conclusions into account.