A Practical Hydraulic Fracturing Model Simulating Necessary Fracture Geometry, Fluid Flow and Leakoff, and Proppant Transport

Abstract

SPE-AIME member now with Schlumberger Well Services Abstract Hydraulic fracturing model using various sets of fracture flow/geometry equations are available in the industry. The majority of these models assume a constant fracture height selected at the start of the design, and simulate two-dimensional fracture geometry (width and length) and one dimensional fluid flow in both the fracture and the formation. The two-dimensional fracture geometry simulation can lead to optimistic estimates of fracture lengths and the one-dimensional flow may not allow adequate representation of proppant transport and fluid loss. Highly sophisticated hydraulic fracturing models are available that simulate three-dimensional fracture height and two-dimensional fluid flow throughout the entire fracture process. These models are versatile and are recommended for highly complex, layered, reservoirs where rock material properties, in-situ stress distribution, and flow properties are variable at the wellbore and also throughout the reservoir. For everyday use by the completion/production engineer, hydraulic fracturing models need to be generated that are more advanced than the "conventional" two-dimensional models but simpler approach and less costlier than the fully dimensional models. The hydraulic fracturing model present this paper can be classified as a pseudo three-dimensional fracture geometry accounting for simple two-dimensional fluid flow. Fracture height at the wellbore, width, and length are computed simultaneously. These calculated parameters are then compensated for two-dimensional fluid flow which accounts for friction pressure drop and gravity. An iteration process is set-up until a satisfactory convergence is attained. Knowledge of appropriate fracture geometry and two-dimensional fluid flow enhances the accuracy in fluid loss calculations and proppant transport. Special data required as input proppant transport. Special data required as input to this model include the in-situ stress and mechanical properties distribution in and around the pay zone. The paper presents hydraulic fracturing theory and the basis of the model under discussion. A detailed explanation of the model is also presented where field examples are used to illustrate its use and importance. INTRODUCTION AND BACKGROUND There are currently two basic types of hydraulic fracturing models available in the industry. One that simulates two-dimensional hydraulic fracture geometry and one-dimensional fluid flow (TWO- D). The other type simulates fully three- dimensional hydraulic fracture geometry and rigorous two-dimensional fluid flow (THREE-D). These models have their advantages and disadvantages and limitations in their application. The models are analogous to reservoir models. Some complex reservoirs may require equally complex hydraulic fracturing models, others do not. Equations used in TWO-D models, are based on descriptions by Geertsma and deKlerk, Khristianovitch and Zheltov, Barenlatt, Perkins and Kern Sneddon, Howard and Fast, Daneshy, and Nordgren. In a recent publication by Veatch' a detailed explanation of these models are presented. Basically these models can be grouped in two divisions based on their approach to calculating fracture width. The Geertsma and deklerk approach is based on width calculation in relations to the fracture length. The other approach is the Perkins and Kern, where the model begins with fracture width calculations in terms of fracture height. Both the approach and all of these models assume a constant height selected at the start of the design. The two-dimensional fracture geometry simulation can lead to optimistic predictions of fracture lengths because of assumed small fracture height. The one-dimensional flow may not allow adequate representation of proppant transport and fluid loss Reservoirs that are bounded by very strong barriers to fracture migration are only adequately represented by these models. Where strong barriers do not exist, as is the case with many reservoirs, the constant height assumption will provide overly optimistic fracture lengths and inappropriate fluid loss estimates and proppant transport phenomenon. phenomenon. Among the currently available THREE-D models are those presented by Clifton and Abon-Sayed Cleary, and Palmer. P. 463