Petro.ai Tightens the P10/P90 Spread with Geomechanics and Drainage
Data Science & Analytics

Petro.ai Tightens the P10/P90 Spread with Geomechanics and Drainage

Rosemary Jackson  •  

Q: How does Petro.ai move us past the P10/P90 analysis?

A: This question lies at the heart of what Petro.ai has solved for oil and gas,” Dr. Troy Ruths, CEO of Petro.ai, explained in a recent interview. "The industry has used type curves as a first pass. Why? Because close works pretty well. If you just look at what’s close to you, you can take the data and the design of your neighbor, plug it in and look at what the performance is of that family of wells that you’re targeting in specific intervals with certain completions styles.

“From this close method, you get a distribution of outcome. But that distribution has always been pretty large. We look at this distribution in the oil and gas industry with P10 and P90. If you took your best well and your worst well—the bigger number over the smaller number, that ratio runs about 3. When I started working in this space, this was something that the SPEE monograph was thinking about. How do we take the offset acreage and make a good prediction?

type curves

“Three. That’s a pretty big spread. That’s a ratio of 10 to 30. This means that you need to create development plans that can handle that spread in order to deliver economic returns. It turns out, that spread is too big. You can poke holes in the ground, and you can follow development strategies but what you really need to do is make the size of that spread smaller.

“So, the response to this large spread has been to take the type curve approach and add the multivariate to hone in and make that spread smaller for different types of development styles and different places in the reservoir.

The problem has been that even with that vision of understanding shale at a different level, the industry didn’t focus in on the most important driver of productivity when going through this process of adding a more sophisticated multivariate analysis to the type curve analysis. That most important driver is geomechanics. In study after study, geomechanics has been shown to be the next ingredient we need to fold in.

“Now, the difficulty of folding in geomechanics is that it’s very technical. The industry-blessed tools that incorporate geomechanics take about 8 hours to run one stage. Additionally, tools like simulators can only be history matched to a handful of wells. There are just too many parameters. You’re trying to simulate a very large system.

“Geomechanics needed to be added in so that there’s a tighter, more connected way of updating the system. More of a just-in-time calibration from the field connected tightly with the strategy at work. The tact that Petro.ai has taken is to think of a shale reservoir with a physics informed machine learning program. We’ve incorporated geomechanics and a lot of other processes at work. The way Petro.ai solved the analytical problem is exactly how the industry thought it should be solved.

“There was a belief that this could be done, but there wasn’t an idea of how tight you could make that P10/P90 ratio. Petro.ai can make it much better than 3. Petro.ai is finding we can make that ratio 1.2 even down to 1.0 something. That’s a big deal.

drainage model midland basin

“In the legacy situation where there was a frac and flip; where you come in, frac a well, demonstrate a type curve and not even bother to finish the rest of the acreage—well, those days are over. Now companies need to make money off the reservoir. What do we need to do that in shale? You can’t mathematically make this happen. You can’t optimize a reservoir if your P10/P90 ratio is 3. With that number you’re going to leave so much in the ground.

“Petro.ai believes the best technique is to get the most out of the ground initially by understanding the drainage of your reservoir. Petro.ai talks a lot about drainage and the drainage model.

“Our clients call Petro.ai ‘having a hundred reservoir engineers in a box’ helping you evaluate creative and financially better ways to place wells in your reservoir and improve the model while you develop it.

“And deliver cash flows back to the investors. You can solve for all those pieces with a more accurate engine that doesn’t give you a ratio of 3 on your P10/P90. That ratio becomes tighter, and the drainage is optimized.

PETRO
TEXAS
114 Main Street Ste. 400
Houston, TX 77002
CANADA
4-5609 Avenue du Parc
Montreal QC, H2V 4S8