Q: How do ISIPs and DFITs work together in the Petro.ai models?
A: We frequently encounter questions from our clients about the interaction in the models of DFITs and ISIPs. (Reposted blog from March 2021)
Leverage the data type you already have. And don’t work off an arbitrary average, quality control your ISIPs, Instantaneous Shut In Pressure, and extract an accurately calculated SHmin. Using the Pump Curve Pressure Model App in Petro.ai, the machine selects your ISIP fits across all your data. Then you clean and extract the ones that create an accurate Minimum Stress model.
You can use DFIT, diagnostic fracture-injection/fall-off test, to estimate SHmin. But, DFIT is costly and time intensive. And according to JPT has inherent issues in data collection which can result in “poor or incorrect results.” Incorrect stress profiles based on incorrect methods keep the Reservoir Engineer from becoming the expert on a well or set of wells. Analyzed and curated ISIPs are the answer to developing a full and confident stress profile.
“When we work with operators, ISIPs are still a throw away data type that have enormous value,” Dr. Troy Ruths, CEO of Petro.ai stresses his concern about this underused resource.
“The industry is getting tons of diagnostic information. Our approach is to do a lot of sampling. In a sixty stage well, you have sixty potential samples. You don’t need to use all of them. Just use the ones you know point to an accurate model.”
Stress is hard to predict and easy to measure. “You can predict stress,” Ruths continues, “but most of the ways we predict stress from petrophysical logs are actually quite poor, not because it was bad science but because it wasn’t developed for shale. They were developed for sandstone which is where conventional reservoirs live and that’s what we’ve been producing from for generations.”
Unconventionals means fracking. The frac moves in the direction of SHmax. The frac grows in the direction of low SHmin. Fracs don’t have to grow up or down or even stay in zone, they go where SHmin determines they should go. Ruths explains, “With shale, early on, it was just a thing you had to go through to get to sandstone. Sandstone has a much lower SHmin than shale. So whenever we fracked in conventionals, if you had shale you didn’t even need to measure it. You just knew it was higher than the sandstone. Now, we’re fracking the frac barriers. Problem number one.”
Shale is a laminar rock. The biggest shale deposits have stacked layering where the vertical interval is very tall. The heart of the issue is understanding the vertical stress profile, namely SHmin. Shale is a highly brittle rock and measuring that is challenging. And that brings us back to DFIT. It’s common at the toe of the well to run a clean fluid DFIT. Expensive and time consuming.
Enter Petro.ai and ISIPs for SHmin determination. There are several complexities with ISIPs. First you have the job itself—poor cementing, bad perfs, the manifold at the pump that might have distorted your ISIP measurement. Secondly, the tortuosity of the well, the pipe bending that can also throw off the ISIP calibration.
In the Pump Curve Pressure Model App, Petro.ai allows you to quality control your ISIPs. Cleaning for logistics is the first step, by looking for a good decaying signal at the end of the pump. In the platform QCing can be done quickly and easily. Let’s take a look below.
All these points in the first screen capture are ISIP’s. They weren’t manually picked. The machine picked them. Grab some points and then view the ISIPs individually through the Actions dialog.
What you’re seeing on each of these graphs is the process of treating the well, shutting it off and the pressure dropping down. You’re watching the pressure decay as you take the rate off the pump. What you’re looking for is a curvature change.
The noise is the physics of fluid moving around in this very long well bore. Pick the ISIP at the juncture of the two lines, as close to the pump shutdown as possible. That is the Initial Shut In Pressure.
“Another difficulty is the shut in time is really fast so there’s no follow up data. What we’re looking for is at least a few seconds where you see a clean fall off signature,” Dr. Nitin Chaudhary, Senior Data Scientist, clarifies. “So, I’m going to go through the selected items and take out the ones that have issues. I’m going to remove those ISIP picks. That’s fine. We made them automatically. The system can redo if necessary. Now we’re left with the ones that we believe will predict an accurate SHmin.
“In the case where the ISIPs seem high or if you don’t have lots of ISIP’s and you don’t want to throw certain ones away, the engineer can come in and help the machine out. We’re using the blue curve which is the treating pressure. I’m looking for that joint. When I come in I’m picking four points, low, medium, and high because there is an uncertainty range when you see the water hammer effect. This gives me the ability to put in a range bar on those ISIP measurements. The last dot is the max treating pressure. We can use those four dots to help triangulate what that decay signature looked like at the end of the well. And that’s how we characterize the SHmin which is the black line.”