Visualization of Hydraulically Driven Fracture Propagation in Poroelastic Media Using a Superworkstation

Abstract

Summary High-performance engineering workstations (HPEW's) provide new opportunities for analyzing complex numerical simulations through the use of innovative visualization techniques. These techniques are applied here to a two-dimensional (2D) fracture propagation in rock, where the rock is assumed to be poroelastic. Understanding of the complex physical interactions between fluid and rock is enhanced when the advanced, three-dimensional (3D) graphics capabilities of such a super-workstation are used. These capabilities are defined and discussed with respect to oilfield-related problems. Introduction HPEW's, also known as superworkstations, have a tremendous potential for exposing complex physical phenomena that otherwise potential for exposing complex physical phenomena that otherwise are difficult to visualize or to comprehend. ["Superworkstation" is defined here as an HPEW with a minimum of 10 million instructions per second (mips) of processing power, 16 or more megabytes of random access memory, several hundred megabytes of disk space, a high-level operating system (such as UNIX), and most importantly, advanced 3D graphics display capabilities supported within the workstation hardware. The workstation should also include a high-definition raster display device or monitor with typically 1,024 × 1,280 pixel resolution. "Superworkstation" has been adopted because it is commonly used by vendors to differentiate these devices from enhanced personal computers and from low-end engineering workstations.] Phenomena associated with oil fields are especially challenging because they often occur at great depths and at scales that cannot be reproduced in the laboratory. Furthermore, these problems are frequently 3D and may involve coupled physical processes-such as mechanical deformation, fracture initiation physical processes-such as mechanical deformation, fracture initiation and propagation, and fluid and heat flow. The need to understand these better has led researchers to rely on numerical simulations. The rapid increase in computer performance and the introduction of supercomputers have made it possible to incorporate extremely complex physical models into numerical simulations. It is now becoming apparent to many researchers that they are limited by their ability to communicate with the computer rather than by the computational power and capabilities of the computer. ("Super-computer" is typically used in reference to computers that can perform on the order of 100 million floating point operations per second. Here we refer to a broader range of computers that have been specially designed for high-performance, scientific computing.) The development of supercomputers has been driven by demands for complex numerical modeling. In turn, the development of superworkstations is driven by a need to comprehend the large amount of scientific data that can be generated by a supercomputer. The superworkstation takes advantage of two important human traits: the ability to recognize patterns in complex arrays of data and the ability to visualize large amounts of data. These concepts have been addressed in a report to the Natl. Science Foundation, which states that "the gigabyte bandwidth of the eye/visual cortex system permits much faster perception of visual relationships than any other mode, making the power of supercomputers more accessible." Although the HPEW initially will be welcomed for its computational speed, we believe that a more significant impact will be the introduction of high-performance, 3D graphics at the engineer's desk. The HPEW allows rapid creation of 3D, solid-colored images and the ability to animate those images. The additional space and time dimensions present tremendous opportunities for grasping time-dependent and coupled processes better. This paper is divided into three principal sections. The next section exemplifies the application of an HPEW to a 2D, coupled simulation of fracture propagation in poroelastic rock. The numerical model accounts for three interrelated physical phenomena: fluid flow in the fracture, deformation of the rock mass, and leakoff in the rock mass. JPT P. 574