Unconventional Reservoir Development:  The Need for Speed
Data Science & Analytics

Unconventional Reservoir Development: The Need for Speed

Rosemary Jackson  •  

Petro.ai Scenarios: Options. Lots of options. Users can move the wells, change the inputs, find the best ROI on any pad.

O&G Engineers: You have the variable. You have the reservoir. Now you need an end-to-end tool because one thing you don’t have is time. Tune your completions. Optimize your well spacing. Get your inputs in and the right outputs out, fast.

The need for speed. In unconventional reservoir development you have to crank through your options not in weeks and months but in hours and days. You need rapid calculations on many well configurations to determine right spacing that provides the best return on your investment. The speed of the Petro.ai Platform gives you the ability to generate more scenarios, and with more scenarios you have the potential to make better decisions.

“When I think about improving speed and efficiency in shale workflows,” Kyle LaMotta, VP of Analytics at Petro.ai elaborates, “I’ve always thought that the value you get from speed is not just being able to do things faster or more accurately. Speed also gives you the opportunity to consider more potential outcomes. If it takes me a week to make one type curve, then by the end of that week we’re going to be living with that one type curve. But if I can generate a type curve in 30 seconds then I can change all of the inputs a thousand different times and end up with a thousand type curves in an hour. Then you can evaluate lots of different possibilities.

“The more scenarios you can generate with more variables and more inputs, then the more outputs you have to decide from. You can consider a lot more possibilities. You should really be choosing from a probability of outcomes. There’s never one optimum answer. You have to look at tradeoffs.

“You can say, if we think oil is going to be between 50 and 60 dollars, then here’s the completion design we should consider for this pad. If the price of oil changes, we have another set of scenarios that we can evaluate. The optimum design changes with price. Or if the price of sand goes way up, your wells may not be as profitable as they would otherwise be, but you can still make the best decision on the inputs that you have. If you can’t analyze all the inputs quickly, you can’t determine the effect of say an increased proppant price. You can’t really say that this is the most economical decision at this time. You just stick with the design that you have knowing that you’re drilling and completing suboptimal wells.

sequencing case study proppant and design speed shale

The thing that Petro.ai does to address this need for speed,” Dr. Brendon Hall, VP of Geoscience explains, “is to create a robust cloud-based platform that provides a common data layer. We can rapidly deploy apps, move through dashboards quickly on top of that data layer. Your data stays in the same place and we can do complex computations, deep analytics without the data needing to change hands among different applications and the manual work required for that.

Petro.ai minimizes the tradeoff time between the different legacy tools that are often being used to go from raw data to an answer product.

“Single use tool focused legacy workflows require people to go through a lot of manual steps that are oftentimes very repetitive. These are the kinds of steps that can be replaced by AI and just plain automation to be able to do them repeatedly and very quickly. Petro.ai brings expertise and infrastructure to deploy these automated pipelines and machine learning tools to do the amount of manual work required for shale workflows. Our new design principles are focused on reducing the amount of human input to routine tasks to the largest degree possible so that we can review work and QC it and make sure it’s right while letting the machine do all the tedious manual steps that are required to process the data and get the answer products we require to make optimized decisions.

“An engineer would typically do an analysis on the well data to get the well properties out. And that would be handed off to the engineering team to integrate that in their productivity calculators.They would compute some kind of a type curve or a decline curve in projecting out the amount of oil that’s anticipated and that would be handed to the economists who would run their economic models. The cycle time for that is at least a week, probably more to change hands and have that analysis done. One person can do a cycle of that in Petro.ai in twenty seconds.

sequencing case study proppant and design speed shale

LaMotta adds, “It’s very manual for engineers to find the data, load it up into whatever software they use to do the analysis and then port it back out whether it’s results or a design or whatever it is—all that data becomes very disconnected. There’s a lot of time spent, finding the data and manipulating it before they can get to the point where they’re making a decision.

With Petro.ai, we’re automating those data inputs through data pipelines so that the data is consistently loaded and the user doesn’t have to think about the data. Even a lot of the results that are needed whether a look back analysis on well spacing or updating type curves or planning out new designs for well locations—even those results can be automated so that they can look at a group of scenarios to make a decision.”

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