Cockrell School of Engineering
The University of Texas at Austin

UT PGE assistant professor John T. Foster and a team of researchers have developed a more sophisticated hydraulic fracturing model to simulate complex fracture networks.

Paraphrasing Einstein, “Everything should be made as simple as possible, but not simpler.” According to Foster, who is working with UT PGE professor Mukul Sharma on the project, today’s hydraulic fracturing models have too many basic assumptions.

“They are simpler than they should be,” says Foster. “We know the existing models are incorrect as they are easily falsifiable during field operations, specifically through higher injection pressures to initiate fractures and more fluid loss.”

The objective of the model is to increase oil and gas production, while minimizing the cost and environmental impact of hydraulic fracturing in the field. “What we want from the well is the fewest amount of stages, the least amount of water and sand and the maximum production – that is the Holy Grail,” says Foster.

foster web story

Dr. John Foster

The model, recently published in a series of journal articles, simulates hydraulic fracturing propagation and trajectory. It is based on the theory of peridynamics, which was born at Sandia National Laboratories as a method to predict pervasive failure of materials.

“The peridynamics theory was only applied in the past to mechanical loading of materials, but we developed a fluid flow coupling so we could apply it to hydraulic fracturing,” says Foster. “Recently, through our computational model, we have validated it against a set of experiments where hydraulic fractures are interacting with natural fractures.”

In a paper published in the Journal of Petroleum Science and Engineering, Foster confirms that if a hydraulic fracture interacts with a natural fracture it will do one of three things: 1) Cross over the natural fracture 2) Turn along and extend the natural fracture 3) Stop at the natural fracture. “In the suite of numerical experiments in our paper, many of these scenarios were tested and our model predicts the behavior,” says Foster.

Foster and his team are currently running the models on the Texas Advanced Computing Center’s’s (TACC) supercomputers as they take a large amount of computing power. Their goal is to make it more computationally efficient so smaller companies who do not have the computing resources can utilize it. They are currently coupling the peridynamics fracture part of the code with more traditional numerical methods away from the fractures, which can run faster.

“Typical simulations might run overnight on a supercomputer, but we want it to run for a few hours on a laptop,” says Foster.

As a next step, Foster is building a robust interface for the research software so industry has the opportunity to potentially enhance its operations.