Increase accuracy by 15%: Data driven AI calculations in a single computational pipeline
Increase confidence by 100%: Surrogate modeling leveraging a robust earth model that includes all your variables, geomechanical to observational
With Petro.ai you’re at the controls, inputting your data and dialing in the drainage, the rock that’s available to be drained by a well or group of wells. Petro.ai builds a surrogate model that can be conditioned based on observational or algorithmic calculations to estimate drainage. Let’s look at those dials one by one: Minimum horizonal stress, reservoir properties like porosity and permeability, the earth model, microseismic, simulator information and actual production numbers.
Minimum Horizontal Stress
Dr. Brendon Hall, VP of Geoscience takes us through the process, “One of the most important things you need to dial in for drainage is geomechanics. This is a word we use a lot here. The state of stress in the earth is really what determines where and how hydraulic fracture stimulation is going to go. It’s this minimum horizontal stress which is the pressure that these fractures have to overcome to propagate, to grow. As soon as they overcome that minimum horizontal stress they’ll move.
“The earth is layered so if the fracture encounters layers with higher stress, then it won’t propagate into those layers. It’ll continue moving in the layer that it’s in. That’s important to have to begin to understand what your drainage is. Stress is one of the major inputs to drainage.”
Hall continues, “Reservoir properties or properties of the rock are important to understand too. When I say reservoir properties, I mean things like porosity and permeability. Two very important things to know about the rock. How easily fluids can move through it. In shale we don’t have very much permeability at all which strongly affects how much you’ll be able to drain.”
“Having an earth model is very important,” Hall explains, “to know where the various layers are. For Petro.ai, the 3D earth model is a digital representation of the properties of the subsurface, the important boundaries and all the wells within the area of interest. This ties into geomechanics. Some layers may have higher stress than others and different rock properties than others. You’ll notice that all these concepts tie together. Knowing where the boundaries are lets you take measurements that you have at one location and make a prediction away from there.
“What’s also important is to be able to say how that rock is going to respond to stimulation.You need to artificially create some permeability by forming connected pathways to all the hydrocarbons that are in that rock in order to extract them. In unconventionals that doesn’t really exist in a natural state. You create it by fracturing the rock.
“Knowing how the rock responds to that stimulation is also important.
“There’s a couple of ways to do that. There’s the observational approach where you can employ various techniques to observe how the rock is stimulated. Common things for that are microseismic. You can use tracers. You can use fiber data. You can observe pressure responses in offset wells.
“What Petro.ai is able to do by building an actual drainage model from those observations like microseismic is to operationalize that data and apply it to a different region. We can then use geomechanics and the geographic model stratigraphy as the link that lets us build a model that is predictive away from the area that you did the observations on.”
“You can also model it. This is where simulators come into play. Hydraulic fracture simulators use physics based models to predict the shape and size of hydraulic fractures that are in the ground. Reservoir simulators let you simulate the fluid flow and how much oil and gas you can extract from the earth once you have it. Those are both computationally intensive approaches and take a long time to run. It’s another way to understand how much rock you’re draining from.
Production Numbers from Existing Wells
“Petro.ai’s approach is a bit of both of those worlds. We build a surrogate model that can be conditioned based on observational data or fracture simulation results to calculate drainage. Other inputs incorporated are actual production numbers that come out of existing wells. This is the Petro.ai data driven approach. We know how a well is completed, where it was completed, what kind of rock it’s sitting in. We’ll look at the production numbers for a whole lot of wells and see what kind of model we can build from that. That lets us know how much oil is coming out of it and that helps us build these models to be able to understand how much we’re actually draining.
“You can take that data from your science pad or observation well and use that across the entire asset.”