When a new well is being fractured, anyone with a producing well nearby needs to look out for hits.
The hydraulic pressure used to fracture a new well is likely to be felt in older wells nearby because fractures tend to grow toward zones where the pressure has been depleted by production.
“It is unrealistic to expect total avoidance in formations such as the Eagle Ford, Woodford, and Bone Springs,” said Mike Rainbolt, completions engineer, senior advisor for Apache Corp., adding “let me rephrase that—it is impossible.”
The risk of a hit is high in those plays in Texas and Oklahoma because hydraulic fracturing there is likely to create long planar fractures, whose growth can be magnified by aggressive fracturing designs used to maximize liquids production.
Damage caused by “frac hits” is the high-profile problem that has fanned interest in the larger, but less obvious issue of how fracturing affects the reservoir between tightly spaced wells. Rainbolt presented a 43-page paper by Apache (SPE 187192) at the 2017 SPE Annual Technical Conference and Exhibition in October, which offered a broad look at the problem with examples and ideas on how to manage it.
The impact of a hit can be big and immediate. The Apache paper included a look at a well whose production dropped 65% after a hit and remained down until it was treated.
But these fracturing-driven interactions may only be apparent with production analysis. Apache’s study found one pair of wells where the area fractured by the second well overlapped with the original well. As a result, it said “the existing well could have produced the reserves without the infill well.”
Tightly spaced wells fractured using more water and sand are producing more oil, “but if you look at the incremental production per frac, you see it is going down,” said George King, the lead author on the paper and the global technology distinguished engineering advisor for Apache.
King is leading an effort with Ali Daneshy, president of Daneshy Consultants International, to produce an SPE technical paper commissioned by the SPE Board of Directors on fracturing-driven interactions due in the spring.
The growing range of fracture impacts, bad and sometimes good, has spawned a number of names to try to describe this phenomenon. The label “frac hits” was used in the Apache paper because the phrase is the commonly used term for “well-to-well communications, usually within the pay zones, by the intersection of created hydraulic facture energy with fractures or perforations in nearby wells.”
The definition covers a range of possibilities. It can be high-pressure surges of fluids flowing well-to-well through connected fracture systems, damaging the older well’s fractures and reducing production. Or it can be small pressure increases observed during fracturing that may signal reduced production ahead for reasons that are difficult to diagnose. There are even wells in which production improved after a hit, but the formations at the top of that list are not known as big oil producers.
Whatever the name, the issue is likely to become increasingly important. It is the product of more aggressive fracturing designs, with more water and proppant, which are pumped in tighter-spaced areas. Major liquids-producing plays will require more tightly spaced development.
The Apache paper said that fracturing methods need to be adapted to work within those lines. “Let’s not try to change the scope of the work to increase recovery from these reservoirs,” King said. While the tools used for fracturing likely remain the same, the designs need to be tailored to fit within the space available. During pumping adjustments need to be made based on pressure signals in nearby wells to ensure that it does not “screw up the well next to it.”
Why Call Them Frac Hits?
The word “frac hits” is the most widely used description of the phenomenon also known as “well-to-well fracture-driven incidents.”
This follows the lead of the Apache technical paper that opted for using the term frac hits, which it admits is less descriptive than “fracture-driven interaction” or “well interference” but more commonly used.
The reality is more complex than the words frac hit imply. It is related to well-bashing, an extreme hit that can knock out production but not a good description for interwell interference. In those cases the indirect pressure of new fractures growing near old ones can cause small pressure changes that could affect a well’s output over time.
Unpalatable Options
Breaking the ingrained habits that lead to frac hits will be hard.
For example, it is clear that more aggressive fracturing designs can lead to hits in places with tightly spaced wells. But Rainbolt said that concern is up against the argument that completing wells using more water, more sand, and more fractures will result in more production from new wells.
Wider well spacing reduces the number and severity of the hits, but that limits the number of higher-producing new wells. “We can control it with well spacing. We do not want to, but we can,” he said.
Drilling wellbores that do not undulate up and down, known as porpoising, would eliminate low spots that can fill up and plug a well. Straighter wells are the goal, but there is a limit to what drillers can do while rapidly drilling longer laterals in formations where the best path is often not a straight line. “The reality is it will porpoise up and down,” Rainbolt said.
Perforating wells so the fractures created by a new well do not line up with the ones in an older well will reduce the risk of fractures growing into each other, creating a flow path from well to well. But varying the plan runs counter to the need to mass produce wells with many evenly spaced fractures. “With so many [perf clusters 17 to 25 ft apart], good luck keeping them from being across from each other,” Rainbolt said.
Completing all the wells on a pad at one time can reduce the risk, but the effort required for simultaneous fracturing exceeds the patience of many operators. It is also not an option for later infill development.
Adding wells close to older ones is not likely to stop because operators need to maintain production in plays where the decline rates are steep. While fracture interaction may mean new wells will not ultimately produce as the ones drilled 3 years ago, they will initially be producing several times more barrels of oil per day.
Apache’s goal is to find ways to adapt the tools now used to monitor and fracture wells to minimize the waste caused by fracture interference and maximize the production from each well.
Best Defense?
A common industry response when fracturing is happening nearby is to shut in wells and sometimes pressure them up by pumping in water. It seems like a logical solution since fractures tend to grow toward lower-pressure areas.
The Apache paper said that just adding pressure will not block an incoming hit. “Shutting wells in to limit hits, or pressuring wells up, is not very effective” at averting production losses, Rainbolt said.
On the positive side, it may limit wellbore damage and speed well recovery if there is a hit. Most important, it allows pressure monitoring, critical for managing fracturing and using it to learn about how fractures develop in a location.
A pressure increase of as little as 15 psi may be a warning sign of fracture interference. King said reacting to that signal by reducing the pumping rate by 10 bbl/min may reduce the risk of fracturing-induced damage.
“We are getting closer and closer to understanding at least what are the causes and what are the monitoring points that we can get from simple gauges,” King said, adding, “The next step is how do you control it? It is up to us to take action rather than put it on autopilot.”
Instead of pushing the injection rate up to the maximum in as little as a minute, the paper offered an example where the pressure rose for 1 hour.
That would give those at the controls time to observe and react based on the pressure responses in the older well.
Making this change presents an organizational challenge. It requires setting up monitoring systems, training decision makers on when to make adjustments, and changing computer-controlled systems to allow quick changes of the programmed routines.
The process could be aided by fracturing slower. Rather than pushing the pumping rate up to the maximum in as little as a minute, a slower ramp-up would provide those at the controls with time to observe and react to gradual pressure changes, King said. Slower changes would also affect the fractures developed.
“A steady rise in surface injection pressure at a constant injection rate may be indicative of fractures staying in zone,” the Apache paper said, advising that “changes in the surface injection rate may be a first action in preventing or minimizing frac hits.”
Monitoring using chemical tracers is also recommended, which can show if there are direct well-to-well connections because of connecting fractures.
“If you are not monitoring the offset well and casing pressure, why not? If you do not, you do not know how bad the hit is,” Rainbolt said.
Is a Frac Hit Model Possible?
With growing concerns about frac hits reducing production, there is a demand for ways to model it. But the specifications for such a model set out in the technical paper by Apache could be daunting for developers.
Models to simulate how hydraulic fracturing in one well affects others nearby are needed because the paper (SPE 187192) stated that “many prevention and remediation failures are the result of misinterpreting the cause of production loss during the frac hit.”
Getting that right will require realistic accounting for a list of geologic variables, complicated by the fact that the rock makeup varies over short distances and the properties may be interacting. “Models that predict frac hits might be possible if the rock and potential flow paths are understood, but formation heterogeneity within the shale, including the interaction of structure, rock fabric, mineralogy, in-situ stresses, and induced stresses form a very complex array of influences, with data availability and quality concerns challenging model use at present,” according to the paper.
As for using the available fracturing models, the paper noted that they “tend to yield somewhat different results.” Deciding which model, or models, to use would be time-consuming and not currently worth the trouble because while they were “developed by brilliant scientists, computer programmers, and engineers” the results are questionable because “most of the time we do not have as much information as is needed to input in the models and [must] make numerous assumptions and ‘best guesses.’ ”
Trial and Error
Many technical papers outlining the ideal spacing for unconventional wells have been written over the years, but there is no simple rule of thumb for finding the safe interval.
Fracturing sends out a high-pressure stream of water that can end up traveling to unexpected places for reasons that are hard to know. The Apache paper said the fractures created are based on the interaction of controllable factors—it lists 11 of them—and the uncontrollable ones range from the natural fractures present to natural barriers that block fracture growth.
As a result, the results in the ground achieved by identical completion designs will vary. “You are not manufacturing something—you are altering Mother Nature a bit,” King said.
In some plays, a frac hit event has a good chance of adding production. Those include the Marcellus, Haynesville, and the Barnett, King said. But in the Woodford, there is a 94% chance that a hit will hurt.
If the reason for damage cannot be determined in a hit well, it is often hard to know what should be done to restore production. The producing fractures, whose shape engineers can only imagine, are the starting point of any such investigation. What has been observed from the few reports that have directly observed fracturing by collected cores, as discussed in a recent paper by ConocoPhillips (URTEC 2670034), shows how much we do not know.
“Whenever we directly measure created hydraulic fractures or the SRV [stimulated rock volume] we are always surprised and our mental images are altered and begin to change,” said Kent Perry, executive director of the Gas Technology Institute, which has organized test sites by exploration and production companies in the Permian Basin.
Trial and error is often the best available tool to figure out which spacing allows maximum development without excessive interference. A single well test may show that 300-ft spacing is not a problem on one well pad, but generally tighter is riskier.
Which tools are available are based on what can be observed in the well. “Diagnostics to determine where production losses are occurring and what caused them are still being formulated, but several approaches described include rate and pressure transient approaches and chemical analysis of fluids produced after a fracture treatment or a frac hit,” the Apache paper said.
