At a remote shale gas project, a robust manifold system to allow simultaneouspreparation of multiple wells significantly cut downtime and associated costs.
Karl DeMong, Apache Canada Ltd.; and Kendall Keene, T3 Energy Services
In most of the North American shale plays, gas has been developed in pads of one or a few wells. However, for a remote development like the Horn River Basin of northeastern British Columbia, Canada, the economics of this small-scale type of development are daunting, even though individual well performance has been encouraging. The requirement to develop pipelines, roads, compressors, camps and supply infrastructure drives up costs. In addition, costs are rising for completions—the single largest ongoing area of spending in a shale gas development budget.
The planning of Apache’s Horn River development emphasized improvement of both time and cost efficiency in the completions portion of the project. Apache wanted to avoid expensive nonproductive time (NPT) by drilling multiple wells on the same pad and by preparing most or all of the wells simultaneously so operations could always be continued regardless of the situation in a particular well. The key to having all the wells ready to frac at all times would be a robust manifold system designed to handle the severity of fracture treating conditions.
The decision was made to space the wellheads as closely as possible (Fig. 1), which would reduce the size and cost of the manifold and other wellsite equipment, make it easier to heat the manifold, frac trees and plumbing in the subzero environment, and reduce the time needed to move and rig up intervention equipment from well to well. In addition, the close spacing promoted environmental stewardship by reducing the surface area disturbed per well.
 |
|
Fig. 1. Aerial view of an Apache pad site in the Horn River Basin in midwinter. The manifold is under the hoarding to the right of the yellow frac tree shelters.
|
|
Apache approached T3 Energy Services to partner in the development of the manifold design and build the components. A team of operations and engineering personnel from both companies evaluated various design concepts for the manifold as well as other elements of the completions concept for Apache’s Horn River Basin development. Through the use of the resulting systems, NPT on the critical path was greatly reduced, and the project’s completion rate increased from less than one frac per day to 2.5 fracs per day, beating the best performance previously achieved in the play.
DESIGN REQUIREMENTS
Among the key design parameters for the manifold, the layout had to allow easy rig-up for pumping and intervention equipment, Fig. 2. All wells would have to be tied in prior to commencement of the frac program. The manifold would need to be able to withstand a working temperature range of −75°F to 275°F and slickwater slurry rates of over 100 bbl/min. for multiple pads, with metallurgy supportive of freshwater, treated produced water and frac flowback water use. For safety, the manifold main valves would have to be operated remotely and their positions indicated remotely. Other design parameters were backup manual valves to be supplied for every actuated valve; design of manifold and frac tree systems for linear close interwell spacing; and use of iron erosion management best practices.
 |
|
Fig. 2. The manifold during rig-up.
|
|
The manifold was conceived as a semi-permanent system, in the sense that it would be stationary for the duration of stimulation on a single pad and then moved to the next pad. The general concept was to create a single station from which to frac all the wells on a pad, with the frac pumps, blenders, chemical mixing equipment, bulk chemical storage, sand blenders and sand storage all centrally located on the pad and plumbed to the manifold to distribute the sand, water and chemicals to each well in turn. To avoid interruptions caused by logistical problems, such as weather delays on the roads, a minimum one-week supply of all stimulation materials would be stored in bulk onsite and tied in at all times.
The manifold designers faced several considerable hurdles to success.
Equipment limitations. The design requirements subject the manifold and associated equipment to more severe conditions than those usually experienced by frac lines and equipment over several years. Even compared to the shale gas frac benchmark, the Barnett, three to four months’ service in the manifold is roughly equivalent to one year in the Barnett.
Continuous duty, though a long-term goal for frac equipment, had not yet been achieved to the team’s knowledge. Hard duty is usually defined as 40% pump time, and some companies have claimed equipment improvements allowing 60% duty or better. Planning for appropriate backup equipment and replacement parts and scheduling maintenance, service personnel and contingency plans are all of critical importance to pumping uptime.
Manifold-to-frac tree interface. At the location of this interface, the close interwell spacing on the pad precludes the use of a multi-line frac head (commonly called a goat head) and dictates that the lines come in from one side. Well placement is not precise enough to hard-plumb the piping with prefabricated parts, so pipe jumpers were fitted onsite after the frac trees were installed and the manifold was laid, Fig. 3. This process meets the reliability and safety requirements but is time consuming and on critical path, and therefore not ideal.
 |
|
Fig. 3. Jumper lines coming off the manifold inside heated hoarding.
|
|
It is preferable for the manifold-to-frac tree interface to have some flexibility to allow for easy rig-up and compensation for variations in frac tree location and dimensions. There are two methods of providing this flexibility: hoses and swivel joints. Neither provides the desired reliability or safety at this time, but work is continuing. The hard lines provided additional bracing for the frac trees and manifold that helped to prevent pulsation vibration leaks. Other methods have been identified and will be employed to handle vibration if a suitable flexible solution is found.
Extending technology. Even the established technology being used on the manifold is being required to fulfill more demanding roles, in some cases roles that are quite different from their designed use. Heading that list is the use of large-diameter valves, which in the manifold must control a high-pressure, high-velocity sand slurry. There is little historical precedent for this type of use for these valves; inspection in the field is needed to verify the design.
Logistics. In such a remote environment as the Horn River Basin, gathering, moving and assembling large quantities of equipment and ensuring 24/7 availability of trained service personnel are major challenges that require careful planning. In many ways, the logistics were closer to those of support for an offshore project than for any surface project outside of the North Slope.
DESIGN DETAILS
The manifold method to fracture stimulation has been approached from three points of view: 1) a single-run manifold intended for use on multiple pads with low service requirements and a long life expectancy; 2) a single-run manifold intended for use on a single pad; and 3) a smaller double-run manifold intended for use on multiple pads. Each of these approaches has advantages. They also share certain characteristics. The basic concept is the use of a stationary pump station and manifold design to deliver 750,000 to 1 million gallons of a frac mix at about 10,000 psi, at a rate of 100–110 bbl/min., more efficiently and cost-effectively than the use of mobile frac units that travel from well to well. The tradeoffs analyzed to judge these three approaches were initial cost versus total cost and proven technology versus leading-edge technology.
Spools and valves. The double-run design was chosen, and two nearly identical trunk lines were designed, with many of the spool and valve components interchangeable. This interchangeability would extend the overall life of the manifold by allowing parts subject to higher wear to be moved to lower-wear locations in the manifold. A 5-in. nominal ID was selected as a balance of initial cost, serviceability and overall life. The basic building blocks of the manifold are approximately 8-ft spool pieces with heavy wall sections for long life. The spool pieces were manufactured with integral flanges to minimize turbulence and increase life. High-hardness steel was selected with an eye toward low wear and long life.
The major problem with spool design has been leaks in the flange seals, which were found to result largely from inconsistent support for the manifold and jumper lines and from the pulsation vibration initiated by the high-pressure frac equipment. There was very little wear on the spool pieces on either side. Sizing, design and the dual-line concept work very well in this regard.
After spools, the second major piece of a manifold system is the valves. In this case, valves are used to isolate the left half of each string from the right half. Figure 4, a screen shot of the supervisory control and data acquisition (SCADA) system used to monitor and control the manifold system’s operation, shows all the valves for one of the manifold’s two sides. A central collection block is used to interface the frac pumps to the manifold. A primary valve and a redundant valve are placed on each side of the collection block to isolate the two sides. Moving away from the collection block, a block tee is placed to align with each frac tree. Downstream of the tree, a valve is positioned to isolate any frac tree not being subjected to frac activity. This first valve is the manual valve, used for backup and redundancy. The primary frac tree isolation valve is located on the frac tree mixing block.
 |
|
Fig. 4. A screen shot of the SCADA system showing valve positions in the trunk line, jumper lines and frac trees on one side of the manifold.
|
|
This same configuration is repeated on the left and right side of each string, resulting in 32 primary valves and 32 redundant valves. Additionally, four primary valves and four redundant valves are used to isolate right from left on the manifold trunk lines. All totaled, the manifold has 72 valves. The primary valves are all electrically actuated, both because low temperatures at the pad location make hydraulic actuation problematic and because the manifold is intended to move every few months, which is easier to do with electric power supplies and controls than with hydraulic plumbing.
With 72 valves in the manifold system, maintenance is critical to performance. An inspection and maintenance cycle was established to regularly function test and lubricate all valves in the system.
SCADA system. The decision to use a manifold had major safety implications for the project. Typically, site safety is centered on the highest risk present: well control. In the case of this site, all potential sources of high flow or high pressure have the potential to act at any location on the manifold. These include, but are not limited to, perforated wells, stimulated wells, frac pumpers, coiled tubing pumpers and pumpdown plug and perf equipment. These risks, while not really different in scope from ordinary site risks, are now distributed through the site by the manifold and require specific attention. It was clear that an additional method of monitoring the status of the valves and various other locations within the manifold was an absolute necessity.
A SCADA system was developed to provide for remote control of valves and pressure monitoring within various parts of the system. Other data collected and functions controlled include the flowback and testing system and the frac water supply system. This information proved to be very valuable in the course of the project. A further enhancement on the next pad will be the use of position indicators on all the manual valves.
Valve operation. To allow crews to work on one end of the manifold, rigging up equipment, it is absolutely critical to ensure that the crews are isolated from potential danger in the rest of the manifold. Therefore, a double barrier and a bleed-off policy were adopted, and the actuated valves were locked out at the SCADA panel and password protected. The actuation of a valve involves a notification, check and response protocol for the crews working near the manifold and valves. The system’s 72 valves may seem like a lot, but in fact work out to only 4½ per well. With two lines and two valves per line required to lock out for service, that leaves only eight valves in the system of 16 wells outside of those required for safe work—just half a valve per well.
Of utmost importance to the safe operation of the actuated valves in the manifold and frac trees was the development of a valve operation philosophy. This was undertaken to ensure a clear understanding of what critical information would be available from the SCADA and the site in general, what responses would be required and how that would translate into valve operation.
Under typical conditions, the frac tree valves will be used as follows:
The lower manual master valve remains open during operations. It is operated on maintenance days as per the service company’s maintenance procedures. The valve may be boroscoped when closed, then reopened and greased to complete the maintenance cycle.
The upper actuated master valve remains open during operations. It may be closed in emergency well control situations only. It is operated on maintenance days as per the service company’s maintenance procedures.
The upper manual master valve is used by wireline or coiled tubing crews.
The actuated frac valves on the jumper lines are the main valves controlling the frac. They are opened at the start of the frac job and closed after the job.
The manual frac valves for the two lines remain open during normal frac operations. They are used to isolate the frac tree or the actuated valves in the case of an actuated valve failure. They are also used to provide a double barrier during concurrent operations.
The pumpdown valve is to be used for pumpdowns and acid spearheads for the fracturing treatments or any specialized pumping.
The casing flowback valve is used to flow back wells up casing.
At any time during the use of the manifold under pressure, the pressure at all 16 frac trees is monitored from the frac data center on the pad site. During the final pressure test, before pumping commences, it will be verified that no parts of the manifold or frac trees are pressurized other than those planned. After pumping begins, the rest of the pressure transducers on the frac trees will be monitored.
If at any time pressure increases on any part of the manifold or frac trees, procedures call for the immediate closing of the redundant manual valves between the pumping side and the area of the pressure increase (i.e., either the manifold trunk lines or the frac tree lines). This is to be done by the frac crew iron hand who is physically monitoring the frac pumper manifold. The pressure should be continuously monitored, and pumping operations continued as long as the pressure does not exceed a value equivalent to the initial shut-in pressure defined for offset fracs.
This pressure value would indicate communication at surface. If it occurs, the frac should be swept over to fresh water and shut down to troubleshoot for a leak.
Erosion control. Erosion is a significant problem in high-rate, high-volume fracture treatments. In the past, temporary plumbing for frac treating has not been treated consistently with velocity standards and inspections. As the desire to increase the size and number of frac treatments crystallized, a plan was implemented to design for, inspect for and mitigate erosion. The design and field operation of the manifold and frac tree use accepted standard erosional velocity safe limits of 40 ft/s for flow with abrasive particles and 70 ft/s for clean fluid.
The team identified three main controllable variables affecting erosion in the ductile steel piping system. In order of importance, these are velocity, piping system shape and total amount of sand pumped. The critical variable is the shape, because of how it influences slurry velocity changes within the system. Significant wear will tend to occur immediately downstream of shape changes within the piping system. A fair amount can be learned from previous experience in slurry system design, but ultimately any new shape would have to be modeled and/or tested in the field to develop an understanding about how the slurry system, with flowrate and density changes, imparts erosion. This is certainly true for cases approaching or exceeding the rule- of-thumb velocity limits. On rare occasions, flow at significantly less than the velocity limits may also cause erosional wear where imperfections and surface irregularities exist. At Apache’s Horn River project, hardfacing, which increases wear characteristics, was planned for use only to repair worn areas.
Both critical and noncritical components are exposed to erosion. Critical components are those that, if they fail, either affect well control or cannot be replaced. The lower master valve is considered critical, while the rest of manifold is considered noncritical.
Design of the piping system aimed to minimize the severity of shape changes, and thus slurry velocities. Where critical components were predicted to experience velocities above the safe limit, an inspection program was established with acceptance criteria and history match where possible. It was decided that components would be taken out of service if they missed an inspection or their materials failed to meet standards.
Noncritical components were considered most vulnerable to erosion in areas experiencing velocity changes. Therefore, it was decided that newly shaped components of the piping system, even in areas that experience velocities below the erosional guidelines, would be inspected periodically until wear characteristics could be determined. The components most affected by this are the collection blocks and elbows, and the piping sections immediately downstream of these. Also of interest are valves, especially those that may have experienced flow while partly open.
One other technique was used to mitigate erosion: a frac tree wear sleeve that reduced the size of slurry velocity changes and increased the distance between velocity changes for the top of the casing. The wear sleeves have produced generally excellent results.
Hoarding. A unique challenge of using a rigged-in manifold in Arctic conditions was keeping the water from freezing in the unused parts of the manifold. Typically in cold-weather pumping operations, the lines are laid out and kept empty until pumping commences. The jobs are pumped as continuously as possible, and the lines are drained upon job completion or significant delay. Breaking integrity on the manifold to drain it and taking the time to re-pressure test the reassembled connections for each job is not desirable. Insulation and heat tracing with electric lines or a fluid line were attempted but were unable to keep up with cooling due to the winter weather. In addition, any maintenance required on the manifold or the SCADA control system would expose the technician to the elements, reducing efficiency. Instead, a hoarding system was designed to enclose the frac tree and manifold area and warm them with forced air heat. It also provided a heated work area around the equipment for maintenance.
The hoarding system was put together in conjunction with a wellhead shelter and work platform. The shelters were designed for a static load and a fall arrest load in case any equipment was dropped while rigging up equipment above the wellhead. The shelters had removable inner walls so that when they were in line with other shelters they could be forced-air heated in conjunction with the adjacent wellhead shelter. The shelter acted as the main frame for the hoarding as well as a conduit for heat. It is planned that future wellhead shelter designs will also provide support for additional work surfaces within the shelter to assist maintenance technicians accessing manifold equipment that is out of reach from the ground.
RESULTS
From an equipment manufacturer’s point of view, Apache’s shale gas project in the Horn River Basin has been a very challenging project; this is a new approach to frac operations, requiring near-continuous service. As a result, some of the design parameters were not well-known, such as vibration, high-flow erosion/corrosion, the northern British Columbia environment, and logistics related to the remote location. Despite the steep learning curve, the first pad has been rated a success.
The main benefit of using the manifold was the improvement in the availability of wells for frac treatment, which saved downtime and therefore reduced the key controllable costs: personnel and equipment charges. In the Horn River Shale, frac spreads can range from 20 to 40 units including all pumping horsepower, pumpdown equipment, blenders, chemical mixing equipment and spare units for 24-hr operations. Use of the manifold system prevents the idling of all this equipment and associated personnel in case of a problem in a wellbore that prevents or delays a frac treatment.
An additional benefit was faster turnaround from frac to frac. In this project, as in many other shale gas projects, after the frac, a bridge plug and a perforating gun or guns are pumped down to plug off and perforate the next zone to make it ready to frac. In a single-well operation this takes roughly the same time it takes to pump a frac, and most of the frac equipment and personnel sit idle until the plug-and-perf job is complete. At the Apache Horn River site, the pumpdown is done on the previously fraced well concurrently with the frac treatment on the next well. In addition, other well activities such as intervention with coiled tubing or jointed pipe, wireline logging, inspection or maintenance may all be done concurrently with frac stimulation.
The manifold, the custom-designed, closely spaced frac head system and the SCADA system required a substantial investment demanding a substantial improvement in frac treatment frequency to show a positive payout for the equipment. Given the size of the fracs and the conditions, it was estimated that without the manifold, the maximum frac frequency achievable on a monthly basis would be 1.75 fracs per day. The target rate for the first use of the manifold was 2.5 fracs per day. Assuming the costs of individual frac treatments are held constant and a site burn rate of US$250,000 per day, that equals savings of over $10 million over the duration of the frac campaign. If the long-term goal of three fracs per day is met, the savings increase to $16.3 million per frac campaign. This is not including the added benefit of accelerated production.
Through the use of the manifold and additional resources, NPT time on the critical path was greatly reduced, although the effects of the manifold system are difficult to isolate. The project went from less than one frac per day to 2.5 fracs per day, surpassing the previous best performance of 1.75 fracs/day achieved in the Horn River Shale. Furthermore, that previous record had been achieved in summer conditions, whereas this project began in January, in the coldest weather of the year.
There are other indications of the manifold system’s success: The key NPT events on the previous project were wireline pumpdown plug and perf failures, which averaged 8% of total runs. Depending on the issue, the cost to the critical path of a pumpdown failure ranges from $50,000 for a pull pack to surface and reload to $250,000 for a pump-off/fishing operation. Failures off the critical path with a manifold system cost less than $25,000 per occurrence. Without changing the failure rate on wireline pumpdowns significantly, there was virtually no impact on the critical path timing (several hours in 111 days) and a minimal impact on costs.
The quest for improved fracture stimulation performance is critical in plays like the Horn River Shale. The use of the frac manifold improved frac cycle time by avoiding critical path NPT and always provided a practical place for the crews to continue to do productive work. The operations motto “16 wells, no waiting” delivered work efficiency superior to the previous Horn River pace setter. The frac manifold, a key component in the improvement of frac efficiency, will continue to be developed to push toward a least-cost/best-value solution to the problem of fracture stimulation in high-cost environments. 
ACKNOWLEDGMENT
The authors thank Apache Corporation for the permission to publish this article.
|
THE AUTHORS
|
 |
Karl DeMong is the Completions and Well Services Manager for Apache Canada Ltd. He began his oilfield career in 1986 and worked as a drilling and completions engineer for several operators in Canada. For eight years he worked internationally in a major service company’s multilateral technology group. He has published technical papers and articles and filed patents in the areas of multilateral technology, expandable tubulars and shale gas technology. Mr. DeMong earned a BSc degree in mechanical engineering from the University of Saskatchewan in 1991.
|
|
|
 |
Kendall E. Keene is the Senior Wellhead and Production Engineering Manager for T3 Energy Services. He began his oilfield career in 1968, and has worked as an engineer for manufacturers of both drilling and completion equipment. He has published papers on fire resistance equipment, high-performance choke trim and team engineering. Mr. Keene holds patents for drilling equipment and seal technology, and has patents pending in seal technology and slurry manifold design. He earned a BS degree in mechanical engineering from the University of Wyoming in 1972. |
|
