Let’s start with some history. The conventional way of developing a field was to drill a well in the middle, stimulate it, and drain some of the reservoir, then restimulate and drain some more. Why didn’t that work for unconventionals?

Refracturing presents a lot of challenges from both a technical perspective and operational perspective. The initial stimulation of a well is typically performed across smaller, isolated sections of the wellbore at a time. But when you come back in, the whole wellbore is exposed to the reservoir and inadequate fluid and proppant transport restricts the restimulation treatment to the heel section of the well, ultimately resulting in poor production performance. This restriction is one reason why in most cases, it is more economic to drill and complete newer wells for higher recovery.

That seems very logical. What's the challenge with infill drilling?

The challenges span across the entire well life cycle and are mainly driven by presence of an existing producing well. The virgin reservoir pressure around the well reduces with hydrocarbons production. When a new well is placed at a certain spacing from the existing well, this reduced reservoir pressure acts like a sink that impacts the drilling, construction, and stimulation operations, which ultimately leads to poor production performance.

This isn't necessarily unique to the unconventional reservoirs but has been a well-known phenomenon in conventional reservoirs. The difference is the fact that we have to fracture the wells, which fosters the negative interference that occurs during the stimulation at rates and pressures.

It seems like we were developing unconventional wells for a long time before this problem attracted much attention. Why did that change and become such a big concern?

The field development cycle is quite unique in North America. Typically, operators first go through a hold by production (HBP) phase by drilling at least one well in each section to secure the acreage. This process takes some time depending on the operator's acreage position and financial position, during which they delineate the best and worst areas. Once the acreage is secured, they start drilling the infill wells, and then more infill wells—and that's when the problem manifests. In fact, more than 60% of all wells drilled and completed across the US are infill wells.

The industry seems to have two different terms for these wells: infill wells and child wells. Are those terms interchangeable, and is there a definition for those terms?

Generally, yes. Talking about "child" wells denotes the interaction between one well and a specific "parent" well, but that interaction isn't always with the nearest well or the well that's on the same pad. From a technical perspective, "infill" is the better term because it denotes downspacing or drilling wells in proximity of an existing well.

As for a definition, we do have one, and it's based on real field data observation and research. We built an infill data analytics platform covering all of the major basins in North America and looked at the trends. After analyzing the data, we defined an infill well as one that is drilled within 2,000 ft of an existing well that has been on production for at least 6 months. This metric holds true across the unconventional scene.

You seem to have a lot of data already. How long has Schlumberger been working on this problem?

Infill wells are getting a lot of buzzing in the industry right now, but we've actually been developing and deploying a wide range of solutions, with great success, as far back as 7 to 8 years ago. In fact, we sort of stumbled into the problem during a refracturing project in one of the unconventional basins. We realized before anyone else that the ability to stimulate and repressurize an existing producing well had some benefit in mitigating interference from newly completed infill wells in close proximity. We then conducted a chronological audit for these infill wells, i.e. tracing its life cycle and identifying the areas that presented significant challenges, preventing the realization of its true potential. This was a fantastic experience, which blew our eyes wide open, because it clearly showed that the challenges were not just confined to the stimulation operations, but rather holistic in nature. From there, we started to enhance our infill solution offerings beyond the stimulation realm to encompass the entire well development life cycle.

In March 2020, we launched our infill well analytics platform, focused on US Land. This platform was created for internal use only (at this time) and provides sales, marketing, technical, operations, and management functions with insights about the infill well landscape in US Land to drive and sustain win-win engagements with our customers.

What do you mean by "holistic"?

Historically, the discussion of infill wells was centered around the stimulation event. But our audit research showed that the challenges actually begins all the way from the subsurface during the field development planning process up to the production recovery process, with each challenge having a cascading effect on the other.

Think about it: Generally, each engineering team has its own objectives and drivers that potentially creates silos. For example, how do you convince the drilling engineer to focus less on speed, realizing that the hole quality has an impact on well performance? Or the completion engineer who is focused on operational efficiency, with neither of them talking to the production engineer, who has to somehow maximize the production of asset.

We realized that we had to address this problem a level higher than our individual product lines, to develop a true holistic solution that spans the entire end-to-end infill well value stream.

The industry has talked about breaking down silos for decades, but they never seem to go away. How is Schlumberger trying to change that?

When we reviewed everything we have done, it became extremely clear that the only way to achieve consistent, reliable outcomes—meaning the operator getting maximum return on investment—three conditions have to be met. You need integrated reservoir-centric workflows, proprietary technologies that are pervasive across the value chain, and communication and information sharing among all stakeholders.

OK, let's say everyone is willing to talk. One of your discussion points was how hole quality affects well performance. How much does it really matter?

On the drilling side, the KPI has typically been speed, regardless of whether the well is properly placed. In some cases, this results in some weird-looking laterals with high tortuosity. The burden then falls on the stimulation engineer to figure out how to effectively stimulate that lateral. And of course, this also affects the production engineer because it might change the artificial lift system that was originally scoped out.

But if you go back to the stimulation problem, when you don't have a smooth lateral, even if you use best cementing practices, you will still have mud channels behind the casing in the horizontal part of the wellbore. Even with a perfectly straight lateral, you still have a high potential for mud channels on the bottom of the pipe. When you stimulate the well, these channels act as a conduit and promote communication between stages and nearby wells. So instead of that perfect frac drawn on a piece of paper—completely avoiding that existing well—you get fluid migrating to previous stages and propagating in the proximity of that pressure sink around the exiting well.

So a drilling problem that doesn't seem like a very big deal cascades exponentially in the completion phase. But it's an easy problem to solve at the construction phase with our proprietary Fulcrum cement-conveyed frac performance technology. Basically, it reacts with the residual drilling fluid, reducing the mud mobility and stops frac fluid flowing through the channels. So this technology that used during the well construction phase helps the stimulation engineer to achieve better stimulation effectiveness resulting in better production performance.

Is cementing integrity the only thing that's affected by lateral tortuosity?

If the wellbore isn't in the right place, that's going to affect perforating as well because you're not going to have an effective connection to the reservoir as desired.

Then, the number of perforation clusters accepting fluid from the stimulation treatment is like a black box in the industry. Basically, you have good traceability of what's going on from rigging up on location to knowing where your wellbore is—even if it's wonky, to the number of clusters shot. And during the treatment, we can trace the fluid/prop at a particular rate down the pipe, but once it gets to the perforation, it goes dark, i.e. we have no idea which clusters are accepting the fluid. Of course, we have technologies like fiber optics that provides some insights to this in real time, but it isn't pervasive across the industry.

Is there a way to force fluid into the other perforations?

So imagine your stimulation design had 10 clusters and all frac models predicted perfect 300 ft half length across each cluster. But in reality, only four clusters are taking fluid. The entire operation, rig up, horsepower, materials, etc. was planned for 10 clusters, but now those four clusters are basically taking double. So instead of the planned 300 ft, your frac might grow much longer—right into that depleted zone of the nearby existing well.

That's one area that we've really worked on with significant success. We have technologies that increases that perforation efficiency from the typical 50% average to as high as 85%. One of such technologies is BroadBand Sequence fracturing service that enables the engineered, sequential stimulation of perforations with increasing initiation pressures for better well performance. This is achieved by using a proprietary blend of diversion material to create a temporary seal off clusters taking fluid, then forcing fluid into difficult to break down clusters. It wasn't originally designed for infill wells but has been modified to effectively deal with the more acute perforation efficiency situation for infill wells.

Before you can use a technology like that, don't you need to first have a better picture of where the fluid is going downhole?

Absolutely! That's where our WellWatcher Stim stimulation monitoring service helps to optimize the infill well completions. It's a high-frequency measurements that sort of give us insights on both the infill well and also nearby existing. That means we know where the fluid is going—how many clusters and how efficiently—as well as the communication or interaction between wellbores.

Coupled with that, our integrated fracturing stimulation software provides a visual representation of what is happening in the reservoir: We inject this fluid and here's how the fracture is growing toward that depleted region. Then we can use those pressure measurements to redesign the fluids or pumping parameters to constrain that fracture growth as needed.

Does flowback after the frac have any effect on production performance?

Yes! It’s been well known that flowback after the stimulation treatment has a direct impact on the well performance. As a result, there have been some rules of thumb, but there was not much research specific to unconventional wells and especially infill wells with those pressure-depleted regions. Think about it though, we've done all this fantastic work to drill the best well, isolate it perfectly, created a box for that frac to prevent it from communicating. Everything went fantastic, everyone is excited, and then we switched this great well onto production—and killed it.

That's one of those things we learned from those audits and how everything cascades from small problems into big ones. The most critical part of that fracture system that we placed is right there at the pipe. If you flow back all that near-wellbore proppant, you pinch shut the connection between the reservoir and the wellbore. That well is not going to reach its potential.

We can avoid that now with our AvantGuard advanced flowback services, which use our Vx Spectra multiphase flowmeter to give us deep insights into preserving the integrity of everything we've done to make this a great well.

And the data from production brings us full circle as we start the next rounds of development, right?

Data-informed planning is definitely the way to move forward, but I think that's one aspect that the industry hasn't really caught onto yet. We see most operators taking the quick and easy solution to minimizing infill well interference problems, which is to stick with larger well spacings. Of course, that has a lot of implications on their reserves, on number of drill locations, and on how long they can stay in business.

There's another option, which is to develop a full field development plan, run the models and all the sensitivities for this static perspective. And that gives us an idea of what the optimal well spacing should be today. That assumes all the wells will be drilled and completed at once. But the reality is that a section development typically occurs over wider time frame, with wells drilled and completed at varying time periods. For example, maybe when the first sets of infill wells are drilled the existing first wells have been producing for six months, and maybe that changes what the optimal spacing should be.

Then as you keep adding infill wells, maybe one has been producing for two years and another one for six months. Now the pressure sink is deeper, so the concept of spacing is actually a 4D problem, extremely dynamic and complex. And that's one of the reasons it hasn't really caught on yet: the tools that are required to solve these problems require deep, cross-functional experience. Schlumberger has the tools and the cross-domain expertise and knowledge to wield them.

All right, so we now have this perfect well and it's producing for some time but starts to decline. What happens next—is there an option for workovers, or how do we know when another infill well is a better idea?

This question goes right back to the data we're constantly gathering for all of these wells. When you look at the industry, the bulk of the data we create during the postfrac operations is individually structured unstructured. What I mean is each operator or service company has a structure but it's not uniform. And it's mostly PDFs—hundreds of pages of notes, images, scans, and pictures. That means it's not time-series data like SCADA systems. It turns out that as an industry, less than 1% of all that unstructured data is actually used to optimize or work over the well performance.

If we keep saying that data and the Internet of Things (IoT) is the new jewel of our industry, we need to be more efficient with how we use our data. And that brings us right back to the DELFI environment, which has the ability to rapidly ingest all of the unstructured data and organize it so you can use it. So you can get a very clear understanding of how an event in a neighboring well may have created a drop or a spike in this well.

The next step in enrichment is to basically be able to quantify the appropriate workover profiles for wells of a particular type within a field. For example, a model might be able to identify characteristics of the best candidates for a perf squeeze or an acid wash and the best timing for it. That's critical to manage economics in unconventional fields.

Do you have any closing thoughts to share?

I would like to bring it back to the idea of parent and child wells. I have two kids, and if you ask any parent about their desire or ambition for their children, we all want our kids to outperform us. For example, if we only achieved 70% of our potential, we start them off with the aspiration that they can achieve 100%. No sane person will start off telling their kids, “Statistically you can only achieve so much; hence our expectations shouldn’t be very high.”

At Schlumberger, we believe that with solid planning and design using proven technologies, reservoir-centric workflows, and digital solutions, every child well has the ability to outperform its parent.