Prediction of Vertical Hydraulic Fracture Migration Using Compressional and Shear Wave Slowness

Abstract

SPE Members Abstract Documented field results of vertical hydraulic fracturing suggest that quite often the created fracture migrates vertically away from the formation of interest (the hydrocarbon-bearing zone), thereby producing undesirable results. The single set of information needed to help answer questions concerning fracture migration consists primarily of in-situ stress, tensile strength, and primarily of in-situ stress, tensile strength, and elastic constants of the rock material in and around the formation of interest. This paper describes the use of full waveform data from a sonic wireline tool to determine the relative stress distribution and the resultant induced hydraulic fracture height. Compressional and shear wave slowness, derived from the sonic waveforms, are used to calculate the dynamic elastic rock properties. A transversely isotropic model is used to compute the in-situ stress from the elastic properties. Advantages of the use of wireline measured data are discussed, as are the limitations of the technique. Final evaluation of the technique is shown through the comparison of predicted and poststimulation measured vertical predicted and poststimulation measured vertical fracture height. Two field cases are presented to illustrate the technique. Introduction With the recent advent of uncertain prices for oil and gas, the importance of efficiently developing hydrocarbon resources has increased significantly. Many costs are relatively fixed and cannot be greatly impacted by improved technology. However, new technology can make significant contributions in hydraulic fracturing operations. Often, a hydraulic fracturing treatment not only represents a large fraction of initial well costs, but also determines the economic viability of a particular well or field. Too large a fracture treatment can be an unnecessary waste of completion funds, while too small a treatment may result in such inefficient drainage of the reservoir as to make a well unprofitable. Because of this economic double-edged sword, a hydraulic fracturing treatment must be designed which best exploits the reservoir. Much research, both theoretical and applied, has been conducted in recent years toward greater understanding and control of fracturing treatments. Although several general three-dimensional computer-based models have been developed, their application has often been limited because of poor input data. Consequently, many rule-of-thumb schemes are often employed for local areas. When the proper input data are available, most of the more sophisticated models can predict the size and shape of a created fracture. Therefore, powerful tools exist in the industry for efficient fracture treatment design but have been underutilized for lack of sufficient field data. The consensus among investigators is that the information most needed for making realistic fracture geometry design decisions consists of the elastic parameters of the rock, the in-situ stress conditions, and the created fracture height. In general, this information has been unavailable except on a relatively few research type operations. Therefore, data from a few wells are being extrapolated for use over large regions so that predictions from even the most sophisticated models often have a large degree of uncertainty. Usually, the greatest uncertainty about the created fracture geometry arises from the estimation of the vertical migration or height of the fracture with changes in treatment pressure and pumping conditions. A majority of hydraulic pumping conditions. A majority of hydraulic fracturing models require a realistic fracture height as an input. P. 459