Sigve Hovda
Norsk Hydro AS
Bergen, Norway
A systematic approach, including better planning and performance evaluation, improved the horizontal drilling efficiency of a multiwell program in the Oseberg field in the North Sea.
The field, operated by Norsk Hydro AS, is one of the largest oil fields in the North Sea and has total recoverable reserves of 1.6 billion bbl. For the past 2 years, Norsk Hydro has had a goal to increase the recoverable reserves by 10%, primarily through horizontal drilling.
The horizontal drilling program in the Oseberg field is one of the most comprehensive horizontal drilling programs in the North Sea. Drilling is conducted from two platforms: the C platform in the northern part of the field and the B platform in the southern part (Fig. 1).
The present horizontal drilling program consists of 14 oil producers from the C platform and 18 from the B platform. Total horizontal displacement varies from around 1,500 m to 5,540 m (Fig. 2). The lengths of the horizontal sections vary from 600 m to 1,500 m (Table 1).
A typical feature of these wells is the three-dimensional well trajectory. To optimize location and orientation of the horizontal section through the reservoir with respect to oil recovery, it has been necessary to turn the well path up to 1800 in some cases (for example, Well B-17 shown in Fig. 3).
A majority of the horizontal wells have production and observation objectives in more than one reservoir unit. In some cases, the uncertainty in establishing the depth of the target structure is greater than its thickness. In such cases, detailed geological steering with changes in the direction and inclination along the horizontal section has been necessary.
To optimize field production, different reservoir units are produced at different periods. Also, the production profile along the horizontal section is optimized through the use of selective perforation techniques. These requirements can only be met by using production liners cemented over the entire interval. Thus, much emphasis is placed on cementing the long horizontal liners.
WELL PLANNING
Two casing programs are used: a full casing program for the extended reach wells and a simplified casing program for the shorter wells with a sail angle less than 50 in the 12 1/4-in. section.
By having two different casing programs, one section can be skipped on the low-angle wells. Skipping this section saves about 4-5 days of rig time plus casing, cement, and mud costs.
The full casing program has an 18 5/8-in. surface casing set below the massive Utsira sand interval at approximately 1,100 m true vertical depth (TVD).
In the simplified casing program, the 24-in. top section is skipped, and instead a 17 1/2-in. top hole is drilled, and the 13 3/8-in. casing is used as surface casing (Fig. 4). The 12 1/4-in. section in the simplified program replaces the standard 17 1/2-in. section and 12 1/4-in. section. No intermediate casing is set.
The full casing program includes a 13 3/8-in. intermediate casing string (Fig. 5). In the past, the 13-3/8-in. casing was set above the "green clay" in the Eocene formation, which has historically created many drilling problems because of its reactive and unstable nature.
By setting the intermediate casing above the green clay, the problem interval was put in the top of the 12 1/4-in. section. As the 12 1/4-in. sections became longer, it became more uncomfortable to have this problem interval 2,000-3,000 m above. A clear connection between open hole exposure time and drilling problems was also seen. At present, the 13 3/8-in. casing is therefore set into the top of the Balder formation in the Rogaland group, even though this practice leads to quite long 17 1/2-in. sections.
The 12 1/4-in. section is the major transport interval. The longest 12 1/4-in. section to date is 2,636 m, but in the very near future, 12 1/4-in. sections of 4,500 m will be drilled. The production casing is a 10 3/4-in. x 9 5/8-in. tapered casing string set in the Heather formation overlying the Brent group reservoir intervals. The 9 5/8-in. setting depth is the same for both casing programs.
The 8 1/2-in. section includes the final build and turn interval up to 90 and the horizontal section. A 7-in. liner is set over the entire productive interval with a 150-m overlap into the 9 5/8-in. casing. The entire production liner is cemented.
DIRECTIONAL DRILLING
Planning a well starts once the targets are known. In the early planning phase, emphasis is put on optimizing the well trajectory, and typically, several alternative well trajectories are considered. Picking the final well trajectory involves several main considerations.
The potential drilling problems are a main consideration. For example, a well cannot be planned with a 6/30 m turn in an interval where there is a known risk of differential sticking. The angle may need to be kept as low as possible through intervals with unstable formations, such as coal stringers. Thorough engineering in the early planning phases pays off as the well trajectories become more demanding.
Another very important consideration is the maximum surface torque for each alternative trajectory. The torque applied to the drillstring is limited to the maximum torque output from the top drive.
To determine the surface torque prior to drilling, extensive torque simulations are conducted. The drillstring configuration is one of the most important inputs for the torque simulation because each drillstring configuration can give very significant differences in surface torque.
Once the drillstring configuration is reasonably determined, the hydraulic calculations are carried out. In high angle and horizontal wells, high pump rates are important in order to avoid hole cleaning problems. One detrimental effect of poor hole cleaning is an abnormal increase in torque.
In planning a horizontal well, one has to be sure that a sufficient pump rate can be maintained all the way to total depth without exceeding the pressure rating of the mud pumps. The high pump rate will in most cases be the maximum flow rate of the mud motor. In the 8 1/2-in. horizontal sections in the Oseberg field, 6 3/4-in., high-flow motors are used. These motors have a maximum flow rate of 2,300 l./min. Table 2 shows drill pipe sizes and maximum flow rates for all sections.
The second important inputs into the torque-and-drag model are the friction factors. The friction factors are a function of the type of drilling fluid, the type of drill pipe (hardbanding, etc.), and formation characteristics.
The friction factors can vary significantly from one field to another. Thus, to obtain results with a high confidence level from the torque and drag simulations, the friction factors have to come from the same field and preferably from the same platform.
Torque data have been collected during drilling in the Oseberg field for several years and placed in a data base. By back calculating from real torque tests, the real friction factors are found. This systematic work developed friction factors that are based on a large amount of data from the same field. The result is torque and drag calculations that correspond very well to actual observed values. These data are very helpful tools in the planning of new, demanding wells.
DRILLING PROBLEMS
Drilling problems experienced so far in the Oseberg horizontal drilling program can be divided into three areas:
- Problems with unstable coal layers in the Ness formation
- Problems in orienting the horizontal section using polycrystalline diamond compact (PDC) bits
- TVD control in long horizontal sections.
COAL LAYERS
Coal layers are present in the Oseberg field in the Ness formation within the Brent group. The Ness overlies the Oseberg formation, which is the main reservoir body.
Besides coal layers, the Ness consists of deltaic plain sandstones, claystone/shales, and siltstones. The sand is represented by fluvial channel fills. These sands are an important contributor to the extra reserves in the field.
The vertical thickness of the coal layers is 0.5-1.5 m. They are not laterally continuous, making correlation between wells difficult.
During normal vertical and deviated drilling, the coal does not present any problems because of its limited vertical thickness, and hence, short well bore exposure. In horizontal drilling, the coal thickness can be problematic, however.
A coal layer of just 1 m vertical thickness could have a well bore exposure of 20-30 m, depending on the dip angle. Experience from several wells indicates that coal exposure of this magnitude leads to severe hole problems. Such problems include tight hole during drillstring trips and even more severe tight hole/pack-off problems during liner running. The main problems here so far include one stuck 7-in. liner and two sidetracks.
The cure for the coal problem has been a redesigned well path in the 8 1/2-in. section. Instead of drilling more or less horizontal through these layers, the inclination has been increased to give a relative angle between the well path and the coal layers of approximately 20. The challenge from a drilling perspective is that this relative angle leads to inclinations of 108-111 in parts of the horizontal section.
The most obvious benefit is reduced well bore exposure. An additional benefit is probably rock mechanics related.
This solution was first used on Well B-05, where coal layers directly caused two sidetracks. The Ness formation was finally penetrated with an inclination of 110 in this well, giving a relative angle between the coal layers and the well path of approximately 23 (Fig. 6). The effect was positive; the well was drilled to total depth through the entire Ness formation and into the overlying Heather formation without problems. Later, a second well was completed with success using the same concept.
ORIENTING
The goal for the horizontal sections is to drill the entire section and the buildup interval in one run. So far, this goal has not been achieved because of problems orienting the drillstring while using PDC bits.
The basic problem is the resistance to sliding the drillstring in the orienting mode. To slide the drillstring, force is applied to the drillstring by slacking off the drillstring weight. Once this force is large enough to exceed the static friction force, a sudden advance of the drillstring occurs. This sudden advance forces the PDC cutters into the formation, and the reactive torque created turns the drillstring. The driller then has difficulty keeping a constant tool face.
These problems are not common with tricone bits. The resistance to sliding the drillstring is not influenced by the type of bit. Rather, because of a different and less aggressive cutting action on the tricone bits, the reactive torque generated as the bit is forced towards the formation is much less than that of PDC bits.
The problem with tricone bits, however, is their limited on-bottom life. Consequently, much time is spent on bit trips.
To solve this problem, the Oseberg drilling team worked closely with a few selected bit manufacturers to develop less aggressive PDC models. Since the start of this project about a year ago, several concepts have been tested: smaller cutters, impact arrestors, increased backrake angle, etc.
Progress has been made during this period, but a final breakthrough occurred on the last well drilled from the B platform. A prototype bit model was used together with a new variable top stabilizer on the motor. This concept made it possible to orient in an extent previously not possible. This combination saved several days of rig time on only one section. The bit featured a quite radical cutter geometry compared to standard bits.
TVD CONTROL
A major concern on the first horizontal wells was a significant difference in TVD depth between the measurement-while-drilling (MWD) survey measurements and the depth found from the detection of the oil/water contact on the MWD resistivity sensors. The divergence was later confirmed by high-accuracy hydrostatic pressure surveys taken in conjunction with the production logging.
The serious consequence of this divergence between measured and actual TVD depth was that the well path came too close to the oil/water contact and in some cases actually dipped into the water. Even if the driller started to orient towards the high side once the water was seen on the resistivity curves (approximately 1.5 m from the free water level), it was not possible to avoid the interval coming too close to the water by that time. These intervals will produce too much water; consequently, they are lost as productive intervals.
An investigation found an angle between the drillstring axis at the location of the directional sensor and the well bore axis. This angle was generated because the directional sensor in the MWD tool was placed behind the upper stabilizer in the drillstring (Fig. 7). The average magnitude of this angle was 0.19, based on bending analysis of the different drillstrings.
This number was also verified by physical measurements. The directional drilling contractor provided an algorithm that corrected for this "sagging" effect. This algorithm is now used routinely at the well site.
Directional measurements from the previously drilled wells were also corrected. The corrected surveys showed good agreement on the two wells where pressure surveys were available (Fig. 8).
HORIZONTAL LINER CEMENT
The most important activity in the planning phase of a horizontal cement job is the computer simulation work. These simulations are run to find the optimum pump rates for the different phases of the cement job.
The optimum rate is the highest possible rate that the well bore can withstand without fracturing the formation (lose returns).
Oil well cementing involves removing one fluid already in the annulus (mud) and replacing it with a different fluid (cement). The more effective this replacement, the better the results will be.
The ultimate efficiency is complete removal of the original fluid. To reach this goal, both the spacer system and the cement slurry are designed to be pumped at rates resulting in a turbulent flow regime in the annular space between the liner/casing and the open hole. Turbulent flow improves the displacement efficiency and ensures uniform distribution of cement in the annulus, even in intervals where the degree of casing centralization falls short of the design value.
Additionally, the different fluids in a cement job can contribute to the total result both by hydraulic effects and chemical effects.
SPACER SYSTEM
The spacer system has three primary functions:
- Initiate turbulent flow
- Separate incompatible fluids (oil-based drilling mud and cement slurry)
- Leave the pipe and formation water wet.
In the Oseberg field, a two-component nonweighted system is used. The first component consists of base oil, and the second component is drill water.
The oil component is compatible with the drilling mud; thus, the two fluids can be mixed without creating any gelation or settling effects. Any mixing of the base oil and the mud merely has a thinning effect on the mud, a desirable effect that helps in mud removal.
Similarly, the water component is fully compatible with the cement slurry. Any mixing of the two fluids only results in a diluted cement slurry. Pumping the fluids under turbulent conditions also tends to minimize this mixing zone because the annular velocity profile in a turbulent flow regime is fairly flat and could be considered to move as a "front."
Both the oil component and the water component contain surfactants. The oil component also contains a mutual solvent that allows for mixing of the hydrocarbon and aqueous phases.
The total spacer volume is typically 20 cu m, and the oil component constitutes one fourth the total volume. This volume is chosen to give a minimum contact time for the water phase (15 cu m) of 10 min.
CEMENT SLURRY
The cement slurry is designed to be pumped in the turbulent flow regime, to have good fluid loss control, and to be stable and nonsettling.
The turbulent flow improves the displacement mechanics, improves fluid loss control to prevent dehydration during slurry placement across long sections of permeable formation, and prevents pressure loss and possible gas invasion. Slurry stability is required to prevent channels or pockets from forming on the high side because of free water or settling.
The slurry design is a careful compromise of these requirements. The design is verified by laboratory testing using rig samples before each job.
Table 3 lists a typical slurry composition used for the horizontal liners. Table 4 lists the typical slurry properties.
JOB EXECUTION
Extensive simulations are run in the job planning phase to ensure achieving high pump rates without fracturing the formation (Fig. 9). The output from these simulations is a detailed pump-rate plan for the entire job, starting when the liner is at setting depth and ending when the plug is bumped.
The most critical phase concerning lost returns occurs as circulation is established after the liner is landed. It is extremely important to be careful in this phase. The mud has typically not been circulated for 30-40 hr at this point and is therefore gelled up. Even though there is a good margin between the mud weight and the fracture gradient in the Oseberg field, typically 50-70 bar surface pressure is required just to break circulation at this stage.
The circulation is brought up to the planned maximum displacement rate prior to starting the cement job. For example, if the maximum displacement rate is planned to be 1,800 l./ min, the well is circulated at 1,800 l./min prior to pumping the cement. This rule of thumb for the Oseberg field applies not only for the liners, but for every casing job.
The point is that if the hole can withstand a certain pump rate with mud without losses, it will be able to withstand the same rate during the cement displacement (the very last part of the displacement will in some cases make an exception to this rule). Another critical phase is the very last part of the displacement where the heavy slurry starts to build hydrostatic head in the annulus as it travels up into the build-up interval. Again, the simulator proved a very useful tool here. If critical, the displacement rate is slightly reduced in this phase to compensate for the hydrostatic effect (constant bottom hole pressure).
Because all liquid additives are used, mud pits are not required to premix or prehydrate dry additives. Both platforms are equipped with batch mixers, which historically have been used for liner and plug slurries.
In this drilling program, the batch mixer is used to prepare the drill water and fluid loss material required for the job. During the job, this mixture is taken directly to the displacement tanks where the remaining materials are added using the liquid additive proportioning system. Typically, the slurry is mixed in the automatic density control mode, where all parameters such as required density, water requirement, and slurry yield are preprogrammed. The mixing is automated and needs only minimal input from the operator.
RESULTS
Since the start of horizontal drilling in the Oseberg field in May 1992, nine horizontal liners have been run and cemented. Total liner lengths have varied from 1,200 m to 1,950 m. All liners were cemented without using stage tools, external packers, etc.
With the exception of two liners where other problems rendered a normal cement job impossible, the results have been good. The cement bond logs (CBL) have been good to exceptional. The latest cemented liner on the Oseberg B platform had a CBL response varying from 0.5 m V to 1.5 m V (Fig. 10).
The following factors were of great importance:
- Turbulent flow regime for all fluids
- Good pipe centralization (minimum American Petroleum Institute standard)
- Thin and stable cement slurry
- A clean well prior to cementing
- Water wet formation and pipe.
DRILLING EFFICIENCY
Norsk Hydro has developed a system for drilling performance evaluation. The system has two levels:
- A detailed level with every section of a well broken down into six main operations.
The detailed level is used for estimating time and for establishing budget days for upcoming wells. Budget days for new wells are based on the average number of days spent on the five best wells in the area. The system is typically updated twice a year. New, quickly drilled wells will therefore continuously lift the standard.
- The other level expresses the efficiency for a whole well in terms of meters per day.
This second level is used to compare a well with other wells and for benchmarking with other operators (Fig. 11). This level is also very practical for finding long-term trends.
An analysis of the Oseberg drilling program over the past 2 years shows that Norsk Hydro now drills horizontal wells more efficiently (lower cost) than it drilled standard 40-65 deviated wells before. The main reasons for this positive trend include better planning, a focus on performance evaluation, systematic work to avoid major disasters (stuck pipe), and better equipment.
Copyright 1994 Oil & Gas Journal. All Rights Reserved.
