Child Vs. Parent Wells

Dear Subscriber,

In this week's newsletter, we share with you one of our URTeC papers from this year that shows how ML models can create precise PDP forecasts with very little data, leaving traditional methods behind.

Building on our previous newsletter's theme, we've got two must-read posts by our VP of Product Management, Ted Cross: "Understanding Parent-Child Well Dynamics in US Shale Plays" and "Where is the Sweet Spot?"

Additionally, we want to remind you of the various events we have in November. Don't miss out! If you're thinking of attending any of the listed events, don't hesitate to connect with Ted on LinkedIn for a chat.

Understanding Parent-Child Well Dynamics in US Shale Plays

10+ years into unconventional oil and gas developments, you'd think almost every well coming online would be a child. But "child fraction" is only 70% in the Permian and 80% in the Eagle Ford.

The Bakken, by contrast, has >90% of wells coming online being child wells.

Where are these "parent" wells coming from? Of course many of them are co-developed parents, so not standalone wells. In the Eagle Ford and Permian, many units are held by production of vertical wells, and operators are also exploring new parts of the field to a greater degree (e.g., Karnes County is 99% children, but Webb County is only 60%).

[URTeC Paper] How Much Data is Needed to Create Accurate PDP Forecasts?

Young wells can be difficult to forecast with decline curves. But Machine Learning can be very accurate!

In this URTeC paper, we highlight how ML models with minimal data, outshine traditional methods.

Where is the highest quality rock? Deriving Rock Quality Index from Machine Learning Models and Principal Components

Where is the sweet spot? Geologists can argue about this until they are blue in the face, but data-driven methods offer a better way.

In our work, we will often find teams fixated on building the best oil in place map, iterating on it over and over again. Of course, hydrocarbon presence is essential to any commercial accumulation. But we have found that the sweet spot within a basin is often determined by other factors: pressure gradient, fluid composition, clay volume, or other geomechanical properties.

Every time someone builds a Novi model, we generate a data-driven rock quality index that we call geoSHAP. This metric takes its name from Shapley values, which are derived from game theory mathematics and applied to machine learning models.

GeoSHAP has several advantages: it's data-driven, so free of the bias of interpreters. You can generate it for any set of subsurface data, whether that is sophisticated interpreted maps, basic log measurements like gamma ray and resistivity, or even just depths and lat-longs.

You can use geoSHAP to quickly analyze a basin, benchmark operator performance, and more.

SPE Austin: November Luncheon

Just 4 days until the SPE Austin November Luncheon!

Join us for an in-depth exploration of Midland Basin well performance trends with our VP of Product Management, Ted Cross .