3D Earth Modeling: A geomechanical model of the subsurface that represents the state of stress in the earth. Bringing your simulator input into this AI built surrogate model creates high levels of dependable accuracy in determining fracture propagation.
FracFingerprint: Petro.ai is tying the simulator information into the geomechanical information and creating a surrogate model. The output is in the Pad Designer App as the FracFingerprint where drainage is accurately estimated. Petro.ai simplifies the output of the frac simulators, processes the results to train the model and determines the FracFingerprint.
Petro.ai is a web-based platform that quickly evaluates scenarios and connects geomechanics to economics. Those scenarios include the simulations that you’ve run. The Petro.ai 3D earth model takes the results of those simulations and builds a surrogate model that uses AI to learn by observing what these frac simulators produce.
“We know the equations of fluid motions and solid mechanics, so we couple those together in a numerical way and create the idealized physics models that goes into hydraulic fracturing,” Dr. Brendon Hall, VP of Geoscience explains. "Petro.ai takes that as an input to the state of stress in the earth which we get from the 3D earth model.
“This complete system, 3D earth model and simulator input calculates how fractures propagate by simulating the physics with AI to provide the full surrogate model of how the rock has broken, where the fracture is propagating and how much proppant and fluid are involved.
“What Petro.ai does is reproduce a lower aura model of the simulators which lets you run it faster. You need those simulators to originate the work. Once the simulator has done its work and it puts out those results, you can train the Petro.ai machine learning to reproduce that work very fast and efficiently.
“This simulator information is combined with the 3D Earth modeling in Petro.ai, a geomechanical model of the subsurface that represents the state of stress in the earth. That state of stress is the primary component you need to be able to predict how fractures might propagate. If you know there’s a very high stress layer that’s sitting above where you’re trying to frac, then the fracs that you’re making might be blocked by that stress barrier or not propagate upwards anymore. If that changes as you move around, you need to know about that.
“Or suddenly you might be at a part of the barrier where that stress layer doesn’t exist. And those fracs will propagate upwards and that would be good to know so you don’t keep doing a pad design that requires those fracks to be blocked at that location. The 3D earth model sets up the inputs that you need to make further predictions like fracture modeling.
“Fracturing is a complex physical phenomenon. It involves solid mechanics or how solid materials change, are broken, bent, and deformed. It also involves fluid mechanics because we’re pumping a lot of high-pressure fluid into the subsurface to create a high pressure that cracks the rock, creates void space so fluid will propagate into that space. Then, as in limited entry, increase the pressure again to the point where it continues to fracture the rock. In those limited entry designs, Petro.ai can determine with a certain high pressure, what the response is going to be.
“It’s hard to test theoretically, to recreate the earth dynamically changing—the pressures and the temperatures and the sizes involved are very hard to reproduce in the laboratory. So, numerical simulation is used to study and quantify the effect. Bringing together all of a company’s resources and inputs is a hallmark of Petro.ai’s success in advancing completions to the next level of predictive accuracy.”
