There is a pressing need to compare and evaluate hydraulic fracture models which are now being used by industry to predict variable fracture height. The fractures of concern here are vertical fractures which have a pronounced elongation in the direction of the payzone, i.e., there is a dominant one-dimensional fluid flow along the payzone direction. A summary is given of the modeling entailed in the basic ORU fracture model, which calculates fracture height as a function of distance from the wellbore in the case of a continuous sand bounded by zones of higher (but equal) minimum in-situ stress. The elastic parameters are assumed the same in each layer, and injected flow rates and fluid parameters are taken to be constant. Leak-off is included with spurt loss, as well as non-Newtonian flow. An advantage of the model is its small computer run time. Predictions for wellbore height and pressure from the ORU model are compared separately with the AMOCO and MIT pseudo-3D models. In one instance of high stress contrast the ORU wellbore pressure agrees fairly well with the AMOCO model, but the AMOCO wellbore height is greater by 32 percent. Comparison between the ORU and MIT models in two cases (also high stress contrast) indicates height disagreement at the wellbore by factors of 1.5–2.5 with the MIT model giving a lower height. Thus it appears there can be substantial discrepancies between all three models. Next we compare the ORU model results with six cases of elongated fractures from the TERRA-TEK fully-3D model. Although two of these cases are precluded due to anomolous discrepancies, the other four cases show reasonable agreement. We make a critical examination of assumptions that differ in all the models (e.g., the effective modulus-stiffness multiplier approximation in the AMOCO model, the effect of finite fluid flow in the vertical direction in the MIT model, and the effect of 2D flow and limited perforated height in the TERRA-TEK model). Suggestions are made for reconciling some of the discrepancies between the various models. For example, the ORU/AMOCO height discrepancy appears to be resolved; for other discrepancies we have no explanation. Our main conclusion is that the AMOCO, TERRA-TEK and ORU models for fracture height and bottomhole pressure are in reasonable agreement for highly elongated fractures. Despite the difficulties in understanding the different models, the comparisons herein are an encouraging first step towards normalizing these hydraulic fracture models.
Fully Coupled Multi-Scale Model for Gas Extraction from Coal Seam Stimulated by Directional Hydraulic Fracturing
