Collision avoidance while planning and drilling complex wells

An aerial view of Wilmington Townlot Unit (WTU) cellars 1, 2 and 3.

Warren E&P, Inc., operates two oil fields in the heart of the Los Angeles basin. Like other operators that have operations within a city, Warren has had to learn to work within a residential community, with homes sometimes only 50 ft away. Warren’s main facility and drilling site is in downtown Wilmington, Calif. The site encompasses 10 acres of land, surrounded by residential and industrial uses. Drilling takes place from two of three cellars. A tailor-made drilling rig is used, and wells are drilled 6 ft apart.

Because drilling started before cellar construction, the wells are not necessarily drilled from the optimum surface location, to prevent crossing and potential collision issues. Adding to the complexity are the 800 original vertical wellbores and 543 sidetrack and redrill wells found within the 1,440-acre unit.

HISTORY

Warren has operations in the Wilmington Townlot Unit (WTU) and North Wilmington Unit (NWU). The company is actively drilling in the Tar, Ranger, Upper Terminal and Ford formations, as part of its redevelopment strategy to recover oil from a field that has been in continuous production since 1933.

On June 10, 2005, Warren commenced drilling of its first well. Since 2005, Warren has drilled 180 wells at WTU, from three different cellars, Fig. 1. Prior to construction of the cellars, wells were drilled, based on the convenience of the rig location and not, necessarily, on the best position to minimize well path conflicts and directional complexity.

Fig. 1. Each cellar contains enough space to drill up to 180 wells.

Warren initially drilled the wells using a periscope tool. While adequate for that time, well complexity, collision avoidance and difficult target requirements have demanded that different tools and techniques be employed.

Warren’s WTU facility contains the entire infrastructure—including separation tanks and injection facilities—for production operations. It is located on the eastern side of the lease, and while the site is of sufficient size to contain the necessary production and drilling equipment, it is not centrally located to easily drill all of the producing horizons.

Drilling takes place in Cellars One and Two. Each cellar can accommodate up to 180 wells in two rows. Piping for gathering, production, testing lines and water injection are attached to the sides of the concrete cellars, Fig. 1.

Warren’s rig is designed to lift—like a hovercraft—on a high-volume, low-pressure air supply, while four 30-hp drive motors move the rig, as needed. While the rig and drill pipe weigh over one million pounds, the Warren rig is quite easy to move from one well to another, in the same cellar, using the Airgo flotation and drive system; short moves typically take 15–30 min. However, the rig is more difficult to move between cellars, with such a move taking up to two days.

While the Airgo system is superior to a normal track-and-piston system, it has limitations. As a result, many wells in the easternmost cellar were drilled westerly under other cellars and wells. Legacy wells on the cellar’s southern end may head northerly. This crisscrossing pattern further complicates collision avoidance issues and may prevent some wells from ever being drilled or even considered.

The resulting combination, of legacy wells, dog legs and collision avoidance issues, makes planning new wells from a centralized drilling location problematic.

PLANNING OVERVIEW

Well planning can now take two or more weeks, with the plans going through numerous iterations to minimize dogleg and angle severity; maximize the distance from existing wellbores; and design a well that can actually be drilled. Torque and drag calculations must also be considered, because few wells are truly vertical or have a target area of hundreds of feet.

Software—such as Petra, Petrel or Geographix—is used to plan wells. Each system has its strengths and weaknesses, depending on the type of formation, 3D visualization or fault depiction. Warren utilizes all three.

FIELD DEVELOPMENT

Land-based operators need to learn from the offshore industry. The planning required to drill from a limited space, such as an offshore rig, can be applied to land-based cellar operations. However, the drive for immediate production and revenue can, sometimes, preclude the rigorous planning phase.

The offshore industry’s drilling and reservoir departments have the benefit of time to evaluate optimal development scenarios, while they await the platform’s fabrication. However, land-based operators usually drill, and then try to figure out how to produce the oil while simultaneously constructing facilities around the rig.

Warren drills its wells from cellars that are 12 ft wide, 8 ft deep and 400 ft long. Previously, an offshore rig was used to drill wells in metal caissons located approximately 12 ft on center. After numerous wells had been drilled, the rig was stationed at the north end of the drilling area, while construction was started on the cellars. This entailed digging around the existing caissons to a depth of 12–15 ft, backfilling, laying rebar, and pouring the cellar floor and walls. This resulted in numerous wells drilled from the north end of the facility heading south.

After a 212-ft section was poured at the south end of the site, the rig was skidded over the cellar. During the next 18 months, approximately 45 wells were drilled from this limited area. The rig used was a platform rig, which attached pipe racks that skidded with the rig. The combined unit occupied approximately 182 ft of the available 212 ft. This limited movement resulted in more crisscrossing wells and complicated the drilling during that time period.

APPLICATION

Well planning and drilling should be a methodical process that is reviewed, based on how numerous wells will influence a field’s future development. A perfect plan would have all of the shallow sands penetrated by wells on the extreme ends of the cellar, or on the outer edges of the cellars, if there are multiple cellars.

Deeper horizons should be drilled by wells progressively inside the cellars, and so on. This methodology prevents drilling shallow wells through, or located between, deeper wells. In a perfect world, this might be possible, but surface constraints usually preclude an optimal world.

In the case of Warren, wells were drilled initially in caissons placed in the ground, with the intention that cellars would be excavated around them at a later date. This backwards planning resulted in immediate production, before the 24-to–36-month period required for cellar construction. However, it also caused significant collision avoidance problems later on.

The constraints on drilling rig movement during construction, resulted in wells being drilled in all directions, and to all depths from a very limited area. This resulted in a suboptimal design for well surface placement and created subsequent anti-collision (AC) problems.

A well plan can be thought of as a method of reducing risk, while, at the same time, achieving the goals of reservoir penetration and minimizing dogleg severity. A good plan can reduce the risk of colliding with an offset well, or drilling in the wrong direction. In an environment with a relatively low number of offset wellbores, a planner may focus solely on minimizing the risks inherent with operating the rig. However, in an environment like the WTU, a “well of least resistance,” without considering adjacent wells, would invariably be compromised in order to avoid collision with surrounding wellbores. This adds a second dimension to a least-risky well plan, which can be called the AC dimension, as opposed to rig execution. Hence, a four-phase spectrum of well types arises: 1) complex, high AC risk (worst); 2) simple, high AC risk; 3) complex, low AC risk; and 4) simple, low AC risk (best).

As the types of risk that must be taken into account increase, the planning becomes more difficult. Prioritization of risk becomes a balancing act, because a reduction in rig execution risk may result in an increase in AC risk. As this problem introduces a significant amount of new gray area, certain ground rules must be established, Table 1.

Some of these rules might be established by the operator, others by the contractor. Crucially, they allow the planner to act within a known set of boundaries and enable him to consider the types of wells that might be acceptable in design and worthy of exploration.

In addition to ground rules, there are procedures to be followed. These may be thought of as secondary objectives that, typically, allow room for bending, if not breaking, the rules. For example, procedures used by Warren are tailored to the type of completion that is employed—Warren completes all wells with ESPs.

To minimize rotational stresses on the ESP shafts, a tangent section, for the entire length of the pump, is included. This tangent section is usually less than 1° per 100 ft. The normal tangent length is actually greater than the pump length, to ensure that running the pump equipment does not require a significant number of pups to center it in the straight section. Warren typically drills a straight section of 150–350 ft. While the goal is to install the pump as close to the top of the liner as possible, geological differences from those originally planned may necessitate shortening, or lengthening, the intermediate casing. This dynamic—combined planning and drilling—requires that the tangent section be modified continuously, during drilling, to accommodate well path changes. Warren’s geologists typically plan on less than 3° per 100 ft from the tangent section to the casing point of the intermediate casing. This ensures that in the event the “surface” appears early, there is a sufficient straight section to install a pump successfully.

Between ground rules and procedures, a viable foundation for a well plan can be developed, and the challenge of balancing risks can then be analyzed and mitigated. An individual operator may prefer to run the risk of collision with existing wells—particularly abandoned wells—rather than trying to drill an especially tortuous or high dogleg path. This is especially true, when the directional survey may be suspect or even unknown for older wells. Thus, when trying to balance the factors, the planner needs to understand the priorities of all parties involved.

RISK DISTRIBUTION

When planning a well, operators choose the well plan elements that are the least expensive, in terms of their countervailing risk. For example, if a build rate is increased from 4.5° dog leg severity (DLS) to 5° DLS, and an offset well is moved from 7 ft at 1,000 ft, MD, to 15 ft at 1,000 ft, MD, then AC risk has been lowered substantially, even though a very manageable amount of additional DLS has been added to the plan.

As this appears to be a good deal, it should be considered and implemented. If on the other hand, build rate is increased from 4.5° to 8° DLS, and an offset well moves from 7 ft to 8 ft, the proposed changes should probably not be implemented, because there’s little AC risk improvement and increased difficulty for the rig.

When in a complex environment, planning can become very judgment-driven and intuitive; it may require the planner to weigh the relative value of the options to be exchanged. Although these values may differ, especially when the potential variety of operator preferences is taken into account, the planner and geologist should feel as though they gained more than they lost.

In a sparsely populated field, the planner has the luxury of emphasizing and minimizing rig risk, because there is no cost for doing so, and the chance of collision is low. However, in a densely drilled field, emphasizing rig risk may come at the cost of AC risk. A simple way to envision risk distribution, in a densely populated environment, is to split the risk into two categories: AC risk coming from the field, and rig risk coming from the
actual drilling.

It is important to understand the difference between positional certainty and positional control. For example, a shallow wellbore up to 1,500 ft typically carries with it a low degree of positional uncertainty because, with modern survey instruments, there can be confidence that the well might only be a few feet off from the designed plan. The same may hold true for all surrounding wells. Nevertheless, a plan should not be so aggressive, that it passes wells by the thinnest of margins, because fine directional control can be, surprisingly, difficult to maintain. Drilling operations, and positional control, can easily drift from the planned well path in excess of 10 ft, especially in tilted, unconsolidated sand shale sequences.

Since steering problems can be difficult to predict or control, a plan should be established, based on positional uncertainty, but it should also be buffered against the realities of directional drilling or positional control.

Warren, typically, drills three types of wells—horizontal, “S”-shaped and sinusoidal, Fig. 2. Each well requires significant planning and execution; each requires different concessions to maintain positional certainty and positional control.

Fig. 2. Warren typically drills three types of wells: horizontal, “S” shaped and sinusoidal.

DRILLING EXECUTION

Planning is not limited to collision avoidance or placing a well in a particular spot in the reservoir. It also includes all of the other aspects necessary to make a successful and economic well, including the actual drilling, running casing, completing the well and determining the type of equipment that will be used to produce the well. Tangent sections must also be included to ensure long-lived ESP runs. Low dog leg turns must also always be considered to ensure minimizing stresses on casing, improving gravel packing success and maximizing ESP run life.

The drilling process, in and of itself, is a mechanical process and should be viewed as such. Too much reliance on computers and graphics, which are not related to the actual drilling process, can lead operators astray. The wells that Warren drills are only 6 ft apart at the surface, and inaccuracies in tools lead to uncertainties in the actual bit location—this is known as the “cone of uncertainty.”

Gyro tools work well, but they are not always close enough to the bit. A simple technique that Warren’s drilling manager uses is to place his hand on the drillstring and feel his way past the adjacent wellbores (Warren’s rig is a top drive unit, so safety is not compromised). Contact can be recognized by tapping transmitted up the drillstring, if contact with an adjacent wellbore is made and the bit reoriented. Reliance on the tools—whether it be computers at surface or gyros downhole—does not, necessarily, alert you to a problem.

One of the most important things that a company can do is to drill the least difficult wells first and progress to more complex wells as the crew gains experience. Warren has seen that a well-oiled drilling team will easily save 25% to 35% in drilling time, even though the wells will be longer and more difficult to drill after a six-to-eight-month campaign.

Visual aids cannot be stressed enough—placing a survey that includes graphics (both plan and cross-section views) on the rig floor and explaining it to the rig crew is invaluable. Areas of potential collision, high pressures due to past injection, and under-pressured reservoirs should be noted verbally and on the aid.

Prior to drilling, injectors located in proximity to the well path should be throttled back to help lessen the reservoir pressure. Taking a kick usually ends up contaminating the mud system. As a result, hole cleaning suffers, and shale sloughing increases. Both will cost time and money, as the mud system is built back up. Shutting in nearby producers, and having check valves in common headers to prevent contaminants from flowing down the gas lines into wells, is another simple solution to prevent costly mistakes.

The choice of directional tools is important. While a “push-the-bit” system is cheap, complex well paths and collision avoidance may require a “point-the-bit” system.

When drilling only 6 ft away from two wells during the surface and intermediate casing phase down to the kick-off point, it is critical to be on the plan. Even small errors add up and, when coupled with the errors of incorrect placement of previously drilled wells, collisions can, and do, become issues. Trial and error resulted in the creation of Table 2 for notifying the geologist in charge of drilling the well.

It is amazing how easily a bit can split open 75/8-in. J55 casing and damage five or six wells, due to mud infiltrating the production system. Planning, communication, shutting in adjacent wells and common sense can save $3 million, or more, very quickly, Fig. 3.

Fig. 3. Complex well paths and potential collision issues.

Drilling complex wells requires learning new things and trying things that are not in the rule books. Weight placement is, probably, one of the most critical elements in ensuring sufficient weight-on-bit (WOB) to drill efficiently, especially in highly deviated wells with significant offset.

The ability to transmit weight to the bit is a function of where the heavyweight drill pipe is placed in the near vertical section and the length of the near horizontal section. Time and distance in underpressured reservoirs can make or break the operation, even before getting to the zone of interest. Differential sticking is the kiss of death and quite often the loss of $3.5 million in tools. Managing mud weights, when passing through low-pressure reservoirs, is critical. In this situation, shutting in injectors in other nearby zones that the bit will pass through becomes imperative.

A simple trick to getting weight to the bit involves pumping frac beads. Warren has used 20/40 beads at the rate of 2 gal/100 ft drilled in long, near-surface horizontal wells. The complex solution is to use a plastic tube that will fit inside the drill pipe and is sealed at one end with duct tape. The tube is filled while waiting for “kelly-down,” inserted before the next stand is connected and dumped.

Complex wellbore geometries can be drilled easily, but the degree of bend in the pipe must take into consideration the stresses on the casing boxes and pins. Regular, 8-round LT&C threads are commonplace on casing, but have limited ability to handle doglegs in excess of 15°. Buttress thread will handle the bends, but it is more expensive. Doglegs should, theoretically, not be over 6°. While casing can be run through 10°-to-12° doglegs, this tends to place undue stress on the string. Warren has run 75/8- in. casing in 14°-to-18°/100 ft wells successfully, but it has also experienced pin breakage when running poorly cut threads. Doglegs in the well path will quickly prevent the transfer of weight to the bit and the running of casing or placement of stainless steel wire-wrapped screen in extended length laterals. Planned tangent sections are quite often forgotten when attempting to achieve a target entry angle into the reservoir. This is detrimental to the rotating shafts in the ESP that will be installed later.

Bones or calcified shales can cause significant problems during drilling. The angle of penetration is important. Angles of less than 65° are, usually, required to ensure staying on and not being moved away from plan. If the bit angle is less than 10° between the drill string and the bone surface, it is nearly impossible to drill through. The bit will continue to slide on the surface, resulting in significant deviation.

Hole cleaning becomes critical to drilling a successful well. Pump rates of 350 to 360 gallons per minute (gpm) for a 6¾-in. hole and 450 to 500 gpm for a 95/8-in. hole are, typically, preferred. Attention to doglegs, and the eventual reaming, is imperative to a successful completion.

The authors have experienced a number of failures during gravel packing that could be attributed directly to doglegs in excess of 10°. The greater the number of doglegs, the greater the chance of sticking the drill pipe and failing to run the liner to the bottom. Three or more doglegs in excess of 12° almost always reduce the chance of successfully gravel packing the well, as the alpha wave fails to continue to propagate.

Using FLOTHRU in the open hole section at 350 to 360 gpm with 38 to 40 sec viscosity, and turning the drillstring at 90 to 120 rpm, results in suitable hole cleaning and the lifting of solids. Actual reaming is accomplished by going from the bottom of the hole back to the casing shoe three separate times. While this may be perceived as excessive, collision avoidance, and the complexity of trying to maintain specific distances from shale boundaries, typically, results in holes with doglegs sometimes approaching 18°. These doglegs should be eliminated to lower the chance of casing failure, improve the chances of running the casing to bottom, and limiting damage to long ESPs that will be installed.

When drilling a well, everything has a consequence, especially in complex wells. Minimizing time in an underpressured zone can result in tight turns, into and out of that zone. Too many sharp turns or doglegs contribute to drilling difficulties, including WOB to torque and drag, and running the actual logs. Sharp turns can result in key-seating of the drill pipe or logging tools. Running the casing can be almost impossible, and the chance of successful gravel packs drops exponentially with doglegs.

The only time that Warren has come close to contacting other wellbores, other than directly under the cellar, was when those wellbores were omitted from the AC database, or the surface coordinates were surveyed or converted incorrectly. If a breach of casing occurs, or the bit is near a producing well, copious amounts of drilling fluid may be lost, with the normal reaction of the drilling crew being to pump more fluid down the hole. In the event a well has a rod pump, the rods will eventually part when there is enough mud in the flowlines, solving the rig’s fluid loss problem. In the case of ESPs, it can be a very different story. Warren over-designs its equipment to obtain long lifecycles. A surface SCADA system ensures that the wells maintain a certain amount of fluid over the pump. In the event of a breach, or additional fluid influx into the wellbore, the variable frequency drive (VFD) will automatically increase the frequency, and the drilling fluids can flood the production systems.

It is from experience that Warren employs general rules of thumb. A cone of uncertainty can be generated for the well path that is being drilled and the wellbores that already exist. Depending on the age of the survey, or lack thereof, and the depth from the surface, this cone varies in size. The AC program calculates the distance to the other wellbores and the likelihood of collision. Warren then manages that risk by either shutting in producers and injectors, to reduce the near neighborhood pressure and problems to the drilling rig, or by reducing the rate of penetration when passing another apparent wellbore.

In the event that the drill bit comes close to a wellbore, it can be detected, albeit late, by looking at the logs, Fig. 4. In this example, the drill bit was between 10 in. and 20 in. from the adjacent wellbore—a wellbore that was omitted from the data base at the time.

Fig. 4. In the event that a drill bit comes close to an existing wellbore, it can be detected by looking at the logs. This figure shows a sample of a near collision.

Drilling from cellars can be a very complex, challenging process. The benefits of drilling from a single site can far exceed cheaper vertical drilling. Surface use can be significantly reduced, and fluid spills can be contained more easily. Physical monitoring and access to the site can also be easier.

Drilling in urban environments requires techniques that are normally reserved for offshore drilling and will become more commonplace as cities and towns grow and encroach on operational sites. The downside to drilling from a single site can be non-standard well paths and the fact that they require significant planning.

CONCLUSIONS

Complex wells are, largely, limited only by the imagination of those planning the wells. While it is inevitable that old wellbores will be intersected while drilling new wells, planning beforehand can mitigate risks and damage to expensive downhole equipment. Where live wells are present, it should be considered necessary to either reduce injection rates or shut in producers.

Complex wellbore geometries can be drilled easily, but the degree of bend must take into consideration the type of casing and the thread connections to be used.

Planning should not be considered just for the current well, but for how all of the wells in the project will come together, and how all reservoirs will be developed. If at all possible, planning should be undertaken, or at least the ground rules defined, prior to commencement of drilling. Complex field development should be thought out thoroughly and sufficient time allocated to all applicable resources—including meetings between drilling, reservoir and completions personnel, as well as vendors—to plan for exceptions and for the future.

Well planning is ultimately about simplicity and assessing risk. A thorough discussion between the operator and the directional service company about the level of risk, and what can be done, must be conducted. In the event that issues arise, a post-mortem analysis, conducted in a professional manner, will help resolve future problems and save money. 

ACKNOWLEDGEMENTS

The authors thank Ellis G. Vickers, Esq. of Warren Resources, Inc., for reviewing, editing and providing invaluable input in this article.