Producers operating in the unconventional oil and gas basins in North America have found in optimized completion designs the right recipes to make production in a post-downturn industry economic. Over the past half-decade, operators have discovered that longer laterals, higher proppant loads, tighter stage spacing and more fracture stages have led to lower breakeven costs in a $70/bbl price environment.

Now there are emerging signs indicating that the era of optimized completions might be beginning to plateau, that operators might be pumping as much proppant into the well as they can and that laterals have extended so long that it might not be economical to drill out much farther.

“We believe there is a point of diminishing returns [on completion optimizations],” said Jim Miller, senior vice president of operations for Chaparral Energy. “There are several areas in the Stack where operators over the past three years have pushed up the pounds of sand per foot. Probably within the last six months, you’ve seen some of them start to pull back and drop back down. When you compare those wells where they pulled back down or decreased to the EURs of the wells with a higher sand per foot concentration, you can see they reached a point of maximum returns.”

Miller added that in the Meramec Formation in Oklahoma, operators likely have yet to reach maximum levels of sand loading, “but the industry is pretty close to seeing it.”

If, in fact, operators are reaching the limit for optimized completions, where does the industry go from here? Where lies the next opportunity for value creation?

In exclusive interviews with E&P, several producers operating in major basins throughout North America as well as service providers specializing in production technologies said the next era in optimization could well come in enhanced production systems and tools designed to draw out decline curves and improve EURs. Recent events indicate such an evolution.

In late July Equinor announced it will deploy a rod lift technology developed by Ambyint—a company that specializes in artificial lift and production optimization equipment— on its wells in the Bakken Shale, where Equinor will expand the system to full-field development.

During pilot testing, Equinor was able to automate rod pump well optimization through the use of Ambyint’s autonomous set point management functionality, according to a press release. By identifying wells that were overpumping or underpumping, controller set points were adjusted “with minimal human interference,” the release stated.

Equinor said in the release that this type of proactive optimization system resulted in increased production rates and more efficient pumping while reducing the well volatility.

“The Ambyint technology has improved the remote data visibility and has delivered a more accurate diagnostic of downhole conditions to our rod pump wells in the Bakken,” said Jack Freeman, production engineer for Equinor’s Bakken asset, in the release. “The autonomous speed range management tool has leveraged the power of machine learning to optimize our wells by identifying and acting on real opportunities.”

Brian Arnst, director of optimization at Ambyint, said much of the processes and systems used by oil companies even 50 years ago are still in use today and that many of those processes and systems have finally proven to be outdated and inefficient.

“Pump-off controllers were introduced in the 1960s, 1970s and [variable frequency drives] were applied on top of pump-off controllers in the early 2000s,” Arnst said. “But at a very high level, companies are still deploying if-then logic. They are not doing any calculations on site.

“Companies are starting to recognize this, and they are starting to look for what that new solution is. Frankly, there are not a lot of options out there right now. You’re starting to see companies trying to find a new strategy, which focuses on what technologies exist that can allow them to take their production optimization strategies to the next level.”

Arnst said Ambyint identified a trend in the market in which production optimization was being neglected, with companies instead focusing their investments more on completions and advancing reservoir characterization technology.

“Now we’re seeing a lot of companies turn to the production side and realize that is where their cash fl ow in 20 years is going to come from,” he said. “They are realizing they need to figure this out and get their [lease operating expenses] in order. A lot of companies are looking to invest in production technologies that can handle the changes these horizontal wells are bringing.”

Legacy well optimization

One of the principal areas that operators are turning to in their efforts to optimize production are their legacy oil and gas wells, which are older wells whose production has tapered off but also might have ample reserves remaining. Legacy wells run on aging systems and therefore are prime candidates for restoration. Among the challenges of optimizing legacy wells are pressure depletion and higher gas-liquid ratios GLRs), said Jimmy Turnini, production optimization manager at Devon Energy.

“We are pushing the limits of rod pump by setting the pump in the curve toward the lateral with success due to better design and equipment available to handle gas and sand migration,” he said. “Drawing down the reservoir is key and getting the pump down deeper helps.”

Turnini said Devon works to stretch out decline curves by combing lift methods like gas-assisted plunger lift and developing in-house algorithms to provide on-demand supplemental gas to plunger wells when needed.

Amir Gerges, Permian general manager at Shell, said the company focuses on improving the performance of its wells through its Well, Reservoir and Facilities Maintenance Program, which he explained is an integrated dataenhanced performance system designed to maximize value.

“A multidisciplinary team performs surveillance and optimization activities across the production system,” Gerges said. “This process consists of gathering, managing and interpreting data that are then used to identify and execute production-enhancing opportunities throughout the field. In our Permian asset, we conduct daily, weekly, monthly and annual reviews.”

According to Rocky Seale, national product line sales leader, completions and well intervention at Baker Hughes, a GE company, a common method the company implements to drive cost savings and improve production are wellbore cleanouts. Seale explained that debris accumulates in the wellbore and impacts the functionality of artificial lift systems such as electric submersible pumps (ESPs) and can cause them to fail prematurely.

“That’s a big value add for the operator, where they can go clean out the well and then, instead of having to change out a pump every nine months or so, it extends the life [of the ESP] to 12, 15 months,” he said.

Robert Turnham, president of Goodrich Petroleum, stressed the importance of maintaining a conservative approach to choke management during production operations, particularly in overpressured wells.

“What we and others are doing is limiting our pressure drawdown on a daily basis to 30 to 50 pounds per day,” Turnham said. “You restrict the flow in an effort to not pull the reservoir too hard, because if you pull it too hard, or open the choke too much, then you’ll see premature depletion. So choke it back to a rate and pressure where you are very gentle on the reservoir and you flatten your decline curves.”

Turnham said such a strategy would ultimately generate better recovery of the gas that is in place versus seeing a higher production rate early on, but have a much steeper decline curve, which Turnham explained as being damaging to the reservoir.

The key component to any production optimization strategy is cost, he said. Turnham identified the best correlation to determining cost efficiency in production operations is the amount of proppant per foot compared to EUR.

“So the higher proppant concentration equals better EUR up to a certain point,” he said. “Then you have diminishing returns.”

Turnham explained how Goodrich pumped 1,000 lb/ ft to 1,100 lb/ft of proppant in the Haynesville, which led to about 141 MMcm (5 Bcf) of return on a 1,402-m (4,600-ft) lateral, or 31 MMcm (1 Bcf) per 304 m (1,000 ft) of lateral length.

“We’re now pumping 3,000 to 5,000 pounds per foot and getting as much as 3 billion cubic feet [85 MMcm] per 1,000 feet [of lateral],” he said, “… so a lot more sand but a much higher EUR. The real question is, by going from 4,000 pounds per foot to 5,000 pounds per foot, is your IRR [internal rate of return] the same, better or worse? Because we know the EUR is going to be higher based on the proppant concentration being higher at 5,000 pounds per foot.”

Turnham said Goodrich is finding that the optimum return might be to “dial back” proppant concentrations from 5,000 lb/ft to about 4,000 lb/ft. “So it’s not the highest EUR, but it’s the highest IRR,” he said.

Well refracturing

Although not all operators are implementing refracturing operations, companies often refracture some of their older wells to squeeze more hydrocarbons from their respective reservoirs, particularly if the well was initially poorly completed or understimulated.

“The biggest thing that has changed is refracking of existing wells,” Turnham said. “In the Haynesville back in 2008 to 2014, we were predominantly drilling 4,600- foot laterals and pumping about 1,000 pounds of proppant per foot. The frack interval was 250 to 400 feet [76 m to 122 m] wide. Basically, we would [fracture] 10 to 12 stages with very low stimulation per foot compared to what we’re doing now. So clearly the vintage wells were understimulated, which is why they are good candidates for refracks.”

Turnham said a few years ago “Hail Mary fracks” were common—pumping without much direction and hoping the well stimulation fluids would get into the formation where operators needed it to get more gas back. Now, Turnham said Goodrich is revisiting its wells by introducing smaller pipe, cementing in smaller casing of the existing well’s casing and running plug-and-perf fractures.

“You can’t pump at similar rates as a new well because the interior casing is smaller, but you’re going into a wellbore that was clearly understimulated,” he said.

David Elkin, senior vice president of asset optimization at EQT, said for refracturing options to be viable in the Marcellus Basin, the initial completion must have been very poorly designed. Although EQT has tested methodologies of the past, the company has yet to observe a need since its early well designs featured what Elkin called “reduced cluster spacing.”

“EQT has very few understimulated wells and therefore very few refracturing candidates,” he said.

Turnini said Devon is engaging in refracturing in the Barnett, focusing on opening new perforations or adding additional intervals that were not originally completed.

“The general approach is to re-perf tighter clusters between existing clusters and try to pump new completions with diversion,” he said.

Oasis Petroleum has been experimenting with refracturing possibilities for about the past three years, said Taylor Reid, president and COO.

“We now have an active program where we refrack parent wells as we are fully developing each drilling spacing unit [DSU],” Reid said. “Generally, we frack the parent well first using a combination of gelled fluid, slickwater and heavy doses of diverters.”

Reid said in Oasis’ early wells they had challenges getting adequate distribution of the fracture across the lateral, which resulted in the toe of the well preferentially taking most of the fracture. The company looked to solve the problem by leveraging microseismic evaluation and fiber coil to confirm the fracture distribution across the lateral.

“In later jobs we adjusted by using combined liquid and particle diverters at higher loadings to much greater effect,” Reid said. “In fact, the microseismic and fiber coil confirmed distribution across the lateral. Once the parent wells are fracked, we then fracked the new offset well. The result has been that wells, which were on rod pump before being fracked, generally flow and then produce at higher rates post-frack, adding unique reserves that more than justify the capital expenditure.”

Reid also disclosed that Oasis is working on EOR designs in the Williston Basin. He said the evaluation is in the early days and the hopes of Oasis and other operators testing EOR in the Williston are to increase the amount of oil recovery from what is typically about 10% to 20% of oil in a reservoir.

“There are a number of options, but the most popular under evaluation currently involves the injection of natural gas and/or natural gas liquids from the reservoir to liberate incremental oil,” Reid said. “As I said, it is early in the evolution of these techniques for the Bakken, but it could have a substantial impact to the production life of a well and to the overall reserves for North Dakota and Montana.”

Seale said refracture design trends in the Eagle Ford and Haynesville include operators initially running a 3½-in. to 4-in. pipe inside a 5½-in. sealing and reperforating sections between the initial wellbore perforations. In the Eagle Ford, some operators are implementing larger-sized casing to plan for refractures at a later date, he said.

“For example, in the Eagle Ford you are seeing operators that instead of running 5½-in. casing, they are running 6-in. casing to give them that little bit of extra room,” Seale said. “That reduces the friction pressure on the initial frack, and then they can run 5-in. flush joints through there and through their refrack if they choose to do it at some later time. So they are planning the wellbore for refrack and re-entry down the road ahead of time.”

Artificial lift trends

Graham Makin, vice president of sales, marketing and investor relations at Silverwell, a service provider that specializes in digital artificial lift systems, said there is ongoing discussion among operators in unconventional plays about the cycle of artificial lift techniques through the life of the well.

Makin said current discussions are centering on how long a well should flow under its own pressure, when to implement ESPs, and when to turn to gas-lift systems and eventually to rod and plunger lift systems.

“People optimizing that process are trying to decide which lift technique to use when, and that process is being driven in different ways depending on if they want to maximize production in the early phases of a well, from an economic point of view, or if they want to go for a longer life cycle and take a reservoir management approach to the well,” Makin said.

Devon has moved away from ESPs and toward gas-lift systems as a result of well flow characteristics and economic benefits because an ESP system is often costly to maintain, according to Turnini.

“IP with steep declines, increasing GLRs and sand migration played havoc on ESP reliability so we’ve moved to gas lift as we’ve developed better designs and, through the use of modeling, determined that we can draw the reservoir down farther than once thought,” Turnini said. “The failure rates are much lower, making the economic benefits for gas lift superior to ESP in many of our cases.”

Oasis Petroleum implements a variety of artificial lift methods in its production operations, including gas lift, ESPs, jet pumps, rod pumps, plunger lift and traditional beam pumps. Reid said selecting the right artificial lift method is dependent on the well and its time in the life cycle.

“The first phase of high-capacity lift is critical to realize the benefits of the high-intensity fracks we are employing,” Reid said. “Rod lift systems simply do not have the capacity to move the volumes of fluid at this stage in the well’s life. As production drops off over time, we will eventually place the wells on rod pump.”

Chaparral’s Miller said they primarily apply ESPs on their Stack wells, which he said have proven more productive than gas-lift systems.

“We primarily determine whether to use an ESP or gas lift by a GLR of greater than 1,000,” Miller said. “If it’s above 1,000, we’ll typically go with gas lift. If the GLR is lower than 1,000 or closer to 500, then we’ll usually move toward the ESP. But when we look at the production of wells on ESP versus gas lift, we see peak IP rates of less than 60 days on an ESP versus peak IPs on a gas lift of sometimes up to 120 days.”

Digital technologies

While operators turn to refracturing and determine the optimal artificial lift strategy to enhance their production returns, Devon’s Turnini said the most substantial impact on production optimization comes in the form of analytics.

“In my opinion, the biggest opportunity to enhance production operations is to leverage real-time data,” he said. “This will enable us to provide the field with information to act on quicker and move us from being reactive to more proactive.”

EQT is among the growing number of companies adopting data analytics and predictive technologies into its operations. Elkin said such an approach aids in extending producing well decline curves.

“We do find ourselves with a larger and larger base production made up of these later-life wells, and we have to think about how to extend and smooth out that production as best we can, even with new volumes constantly being onboarded,” Elkin said. “The answer to that is, of course, getting to lower field pressures but also leveraging our current plunger systems with sensors, transducers and SCADA, and optimizing in real time through production profile algorithms, potentially getting into pattern recognition and AI [artificial intelligence] so we are predicting performance ahead of time and adjusting to it.”

Predictive maintenance algorithms have proven particularly useful for companies like Devon and Chaparral, both of which have leveraged their capabilities to predict ESP failures, which are a common and costly problem in production operations.

“A key for us has been lowering the number of failures we’ve had on ESPs,” Chaparral’s Miller said. “We spent a lot of time evaluating historical data and determined how many failures have to occur before we realize better returns running a gas lift versus an ESP. It’s around one-and-a-half. Typically we would want to shift an ESP to a gas lift or a rod pump sometime between six months and a year, so our ultimate goal was to determine how to keep failures below that oneand- a-half mark threshold during that time frame so that we’re a lot better off economically. Thanks to the enhancements we have made with our automation systems, we’ve been able to do that and seen a significant reduction in ESP failures.”

Josh Walker, vice president of completions and operations at Chaparral Energy, said the company widely implements data analytics, but the analytics systems only work well if the data are managed the right way and “kept clean.” Walker said Chaparral can leverage the company’s historic production data to determine a specific reservoir landing location from a specific well pad and what the first method of artificial lift should be for that well.

“Beyond that, our goal is to get to real-time optimization,” he said. “That’s where the automation comes in so that we can control things from anywhere. We can get to the point, for example, historically with a plunger lift well we might have a lease operator tweak something one day and we’ll watch it for a week, and [then] we’ll make another tweak and we’ll watch it for a week. Where with the real-time optimization, we can see everything in real time and make immediate changes. You can have a machine make 30 changes within a day and get a result that took weeks in the past ironed out in a matter of hours or days.”

Parent/child well challenges

One of the most significant challenges facing the oil and gas industry in regard to production is the relationship between producing “parent” and “child” wells, or well bashing. Well bashing is a result of drilling infill, or child, wells that interfere with existing parent wells as fractures from the new wells connect with the old ones, which can lead to a loss of fracturing fluids and underperformance in both wells.

Well bashing can significantly reduce the amount of recoverable reserves from a reservoir. Operators in every basin acknowledge the problem, but solutions are so far proving elusive.

Elkin said that although EQT does face issues with parent/ child well interactions, the Marcellus Basin is relatively more “forgiving” than other plays. He said EQT manages the effect of offset depletion by managing pressure on the parent well prior to completing the child well.

“Unfortunately, there is no one-size-fits-all solution,” he said. “Both produced volume and the completion design of the parent well can affect the outcome, and alterations to the spacing and completion design of the child well may also be necessary to mitigate this interaction.”

Turnini said Devon is still “exploring options” but has seen some success in controlling the effects of well bashing by recharging the parent well with water ranging from small to larger jobs equivalent to the original fracturing job.

In its Permian Basin operations, Shell conducts reservoir development for the entire reservoir area, taking into consideration the optimal well spacing and completion design for each horizon, Gerges said.

“Dependencies between horizons are managed as co-development, with a focus on managing interference and minimizing parent/child depletion to maximize recovery and value,” he said.

Oasis’ Reid said that in North Dakota the company is applying full-field development strategies, so Oasis does not have much activity cycling in the same area, which helps allay the effects of well bashing.

“We generally have one parent well in each of our DSUs, and when we come in for development drilling we are drilling out the full DSU at all depths in the Bakken and Three Forks to minimize future interference effects,” he said. “In addition, we are often refracking the parent well to improve its performance and to minimize the disruption to the new wells from a depleted parent well.”

As operators turn their attention to production optimization, they will face a multitude of challenges, from implementing artificial lift techniques to addressing the interactions of parent and child wells. Both service providers and operators alike agree that as production optimization plays an increasingly important role in the industry a major step change moving forward will be the application of data analytics, predictive and optimization programs.

“The industry as a whole has always optimized production; it’s just been an extremely laborious and slow process,” Silverwell’s Makin said. “What we are now on the cusp of being able to do is make that process immensely efficient with a much faster cycle time.”