Because hydraulic fracture models include complex physics and uncertainties defined by many variables, the problem of calibrating modeling results with field responses is ill-posed. It is always possible to find a calibrated model that reproduces field data; however, such a model is not unique and multiple matching solutions exist. The objective and scope of the complete paper is to define a work flow for constraining these solutions and obtaining a more-representative model for forecasting and optimization.
Introduction
In this project, a work flow is presented that uses an ultrafast hydraulic fracturing model with well-known physics and high-confidence rock-property inputs to conduct sensitivity analysis for pad-scale field development. Based on a study of uncertainty on the model variables, the two most-uncertain variables (tectonic strain and leakoff multiplier) are selected for model calibrations.