He Figured Out Shale and It’s All Tactical
Data Science & Analytics

He Figured Out Shale and It’s All Tactical

Rosemary Jackson  •  

Knowing the rock: Means understanding the tactical logistics of the subsurface, of the fracture network, of the wells nearby, of your drainage. What’s being produced and where’s it going. Variables that go into the AI.

Knowing AI: Means knowing how to tactically pull together all the variables in complex calculations that algorithmically assemble a complete understanding of the subsurface and generate the decision matrix of possibilities.

Knowing the decision matrix of possibilities: Means having all the tactical options weighted by inputs, weighted by economics. Moving quickly. Changing fast. Deliberating towards a decision with confidence in the parameters and the outputs.

You have a fortress of shale to siege. Smashing walls, launching small projectiles to hold open the progress you’ve made and scouts that return with limited intel. Shale staunchly shields its access points and variables leaving your strategy confidence low. Then, your first lieutenant, Petro.ai CEO Dr. Troy Ruths, lays a battle plan across your worn, field desk and you see with astonishment…He figured out shale.

Not by himself. He pulled together a Petro.ai team of excellence. He resourced experts like Dr. Mark Zoback. And then. He figured out shale.

There are three parts to figuring out shale: Knowing the rock. Knowing AI. Knowing the decision matrix of possibilities.

Knowing the Rock

To know the rock means you understand that network of natural and created fractures in your shale reservoir. It means knowing the every which way oil and gas will go after you’ve pounded across a distance of perforations along a pipe. It means understanding the set of variables and equations to calculate that volume of oil or gas that you’re draining. To know your rock means you know your drainage.

Shale shields that drainage information inside its fortress of access points and variables.

To know your rock, Petro.ai builds a subsurface geo-mechanical model with inputs that include geo-spatially correct well locations, the area dependent vertical stress profile and frac geometry all algorithmically combined to create the predictive outputs of the Pad Design Scenario App, your rock information siege engine.

Niobrara with staggered pattern

Petro.ai is interested in tactical decision making,” Dr. Troy Ruths, CEO of Petro.ai shared recently. "Let’s put that in the perspective of the operator who’s creating an opportunistic offensive approach to drilling and completing a well. You need to know your rock. You need to know the wells in place. You need to know the network of fractures that are draining your rock.

Using a type curve is not tactical. It’s information about one well and then what’re you going to do? Get a general sense of your production by copying and pasting that type curve for another well? To be tactical you need to look at the full scope of your deployment in your rock.”

What are engineers asking when they’re trying to make these tactical decisions?” Kyle LaMotta, VP of Petro.ai continues, “The specifics are the number of stages to complete, the spacings of the stages like how far apart they should be, the design of the stages which includes the number of the perforations, the diameter of the perforations, the interspacing which includes how far apart your different clusters of perforations are going to be. And then at the stage level how much fluid is going to be pumped, how much proppant is going to be pumped. And then what is the response of pressure rates based on those designs?

“It’s going from a really large, 1 mile by 1 mile area, down to hundred-foot stage level. The model can take those things into consideration. We can model at the well level, or we can model at the stage level.”

And that’s a lot of variables to be understood in relationship at one time.

Knowing AI

We know the rock. Now we have to strategize all those variables—all those entry points in the fortress that is shale—taking them in as so many dependent variables across the disciplines that are analyzing the best tactical move. And we’re going to do that with the most beyond complex multi-calculator, beyond simulator, beyond business intelligence (BI) tool. We’re going to do that with AI (artificial intelligence).

Ruths explains, “We’ve come to a point of diminishing returns with simulators and BI and it’s because of the granularity of the information we need to provide to the machine learning model and the noise in the results, there’s just no way you can write a mathematical function that would capture the changes and the results without getting more technical or granular information in there. We can’t build the fortress model from an intelligence report on one tower.

“We’re not providing enough differentiated information to the model for it to actually pull it out. It’s very easy for a company to think that multi-variate regressions (MVR) aren’t working anymore, but an MVR is just a math fitting function that takes into account all these variables.

Bringing together all the rock variables is actually all about advanced tactical decision making. What we need to get is the higher level of granularity information that we have when we’re going through the well development life cycle and the planning and the execution of that into MVR so we can get more tailored results for that particular well.

Right now a lot of companies do batch updates to their models. They collect all the new information and batch update all the models. Even the way they think about it they’re doing it holistically. That’s not tactical.

Drainage model: Midland basin

“Your shale fortress is a Harry Potter-like world of shifting doors and windows. You need to be able to differentiate variables, move your attack, analyze your next options piece by piece as the system changes not in one batch movement. You need to more tactically align your variables.

“If you drill a well then the first thing you might ask is wells per section. You drill a pilot hole. Then you may want to fill in additional wells. You’re trying to think, should I change that variable. You may not want to change that variable right now and that’s okay. Right there someone might say, well then I don’t need to include that into my model.

“But the fact of the matter is, it’s not that you haven’t changed it but you need to present the realization of that shift because you’re going to drill these wells. And it might be different, there might be a parent well in there. It’s still important for it to be a variable in the model.

“Once you drill those wells, you’re going to have DUCs (drilled but uncompleted wells). In general, you may have an idea for what your stage design should be but now you want to tailor it more. It may be different based on where those wells ended up landing. Maybe you had to drop a well.Those kinds of things happen, so then do you change the design. These are all the tactical decisions, because we take a more detailed model into account, that we can answer.

“Because our model is 3D and because our model ties to geoscience and it has assumptions about drainage and depletion, it can be used to give you a solid tactical advantage.”

And there isn’t going to be one answer, but a series of tradeoff decisions that you need to make.

Knowing the Decision Matrix of Possibilities

This is the real differentiator for Petro.ai. All of this intel about your fortress, all of this tactical determination, all of everything you do is tied to economics. How much money can you throw at this siege before you have to go home? If you know your rock, if you have the complex calculator, then you can compile all your strategies and decide which one is economically viable.

A lot of people out there are saying they see microseismic or they use type curves or they run simulators. Those are all indirect measurements. Production, drainage volume is the only thing that is real. It’s the only thing that can be driven to economics,” Ruths continues.

Your tactical decisions need to be within the context of production and its impact on production. Petro.ai takes the rock intel from the well bore model, far field model, interface model, and then regional models too that are folded into it. It takes all of that and creates an interactive model that you move and change. Then you watch as your variables impact the model, and you create a decision matrix of options for your best economic scenario.

Sieges are expensive and you don’t have time or money to throw away. Determine your best strategy on your field table then head your team out with confidence. Petro.ai has figured out shale.

PETRO
114 Main Street Ste. 200, Houston, TX 77002
4-5609 Avenue du Parc, Montreal QC, H2V 4S8

© 2020 Petro.ai All Rights Reserved