DIRECTIONAL DRILLING SOLVES UNSTABLE SLOPE ALONG CANADIAN PIPELINE

R. P. Boivin, P. C. Cavanagh, R. G. Clarke
NOVA Corp. of Alberta
Calgary

Directional drilling has been used to relocate two high-pressure gas pipelines operated by NOVA Corp.'s Alberta Gas Transmission below an unstable slope in a remote area of northeast Alberta (Fig. 1).

Using directional drilling to place the pipelines below the failure zone was the least expensive, lowest risk option, compared to other alternatives considered.

The failure zone had been satisfactorily delineated and modeled during several years of field monitoring and office reviews.

NOVA Corp. of Alberta is a widely held company operating internationally from Calgary.

The company's Alberta Gas Transmission pipeline network is the main system for moving natural gas from Alberta's processing facilities to delivery points within the province and to border stations for export.

With assets of more than $3 billion (Canadian), the pipeline system includes 11,400 miles of pipeline, 2 to 48 in. OD, 44 compressor stations, and 960 major receipt and delivery points.

Total deliveries for 1992 were 3.4 tcf, or more than 10% of the natural gas produced annually in North America.

SITE ANALYSIS

The important events related to the pipelines' construction and operation and the slope stability study consist of the following:

January-March 1981: 10-in. Liege Lateral was constructed across the tributary to Loon Creek Valley (approximately 560 m wide and 40 m deep).
January-March 1988: 16-in. Liege Lateral Loop was constructed,
Difficulties encountered during construction at this creek crossing included caving of the ditch walls at the south sagbend and sloughing of the creek banks. In addition, old failure scarps were noted to be reactivated on the south slope.
March 1988-November 1990: Geotechnical investigations began with the installation of numerous slope indicators and piezometers in the south slope. Slope indicators monitor horizontal slope movements at many elevations within the slope. Piezometers measure pore pressures (pneumatic) and groundwater elevations (standpipes) within the slopes.

The installations were replaced periodically because of the significant (up to 65 mm/year) and, at times, rapid slope movements. The pipelines were stress relieved with more than 300 mm of rebound noted in both pipes.

Geotechnical investigations were extended to the north slope. Relatively small movements were detected there (in the order of 12 mm/year).

A detailed slope-stability study was begun in 1988 following the rapid failure of slope indicators (Fig. 2) in the south slope.

In addition to field monitoring, the program was expanded to include review of regional geology, aerial-photograph interpretation, field investigations, and instrumentation installations and monitoring, laboratory testing, and slope-stability analyses.

REGIONAL GEOLOGY

Bedrock in the site area consists of six interbedded Cretaceous Age sandstone and shale formations overlying a Devonian erosional surface. The two formations of concern were an upper approximately 60-m thick shale and an immediately Underlying 10-15 m thick sandstone. The shale and sandstone dip to the southwest at about 4 to 5 m/km in the area of the site. The shales are more prone to instability than the sandstones.

These latter, however, are subject to piping and block toppling as they provide for regional Groundwater flow. This may also influence stability within the shales. 2

The upper reaches of Loon Creek and its tributaries, which include the Liege Crossing, are incised in this uppermost shale formation. This formation consists of thinly bedded, dark grey to black bentonitic and concretionary shale, together with thin beds of grey siltstone, argibaceous sandstone, bentonite, and coal.

Additionally, the shale is highly unstable and prone to two general types of instability: relatively, shallow rotational sliding and more deep-seated transnational failures.

The rotational sliding is often associated with localized over-steepening of slopes, while transnational failures are more associated with regional factors such as ground water or with existence of weak layers within the strata. Both are reading apparent on aerial photographs of this valley.

The surface deposits above the bedrock are of quaternary age comprising gravels, sands, silts and clays including glacial tills. The thickness is quite variable, in the order of 10 m to 60 m.

Surface topography consists of flatlying valley uplands with saturated muskeg deposits and sparse vegetation. Larger tree cover exists immediately adjacent to and on the valley's slopes. The crossing area slopes are undulating with slope angles at about 10 to 12 off horizontal.

A distinct pattern of individual ridges and scarps on the slopes that roughly parallel the creek suggests the slopes are subject to past and ongoing sliding in a relatively deep transnational mode.

Numerous springs, disrupted drainage paths, and areas of ponded water lie along the creek banks. Spring flooding causes extensive toe erosion of the valley slopes.

FIELD PROGRAMS, RESULTS

Field drilling programs began at the site in 1988 following completion of the Liege Lateral loop. The drilling was to obtain stratigraphic information and soil and bedrock samples for laboratory testing, as well as to install slope monitoring instrumentation.

Initially, the investigations concentrated on the south valley slope where instability seemed more obvious. But in 1990 they were expanded to include the north slope as well. Many of the slope indicators were replaced annually because of rapid movement at some locations.

Instruments were monitored approximately four times per year. A summary of instrumentation locations and results is presented in Fig. 2. Results from a typical slope indicator are shown in Fig. 3.

Laboratory testing conducted on selected soil and rock samples from the site consisted of soil-density testing, grain-size analyses, Atterberg Limit tests, and strength testing. Table 1 presents a summary of selected laboratory test results.

The main strength result obtained was that the residual friction angle within the shale was approximately 8..

The subsurface conditions encountered in the boreholes were consistent with the regional geology for the area. The overburden varied 2.1-0.6 m in thickness and consisted of brown sandy silts and clays.

The bedrock consisted of grey to black clay shale, noticeably more fractured and containing slickensides greater than the level of movement noted in the slope indicators.

The water level readings indicated a perched condition in the colluvium and overburden. Piezometers located in the bedrock suggested little or no excess pore pressure in this zone.

That the piezometers were situated in locations where pore pressure fluctuations were undetected was improbable because of the many piezometers installed.

The south slope indicators showed a roughly horizontal slip surface, about 5 m below creek level (at elevation 505 m) for the lower two-thirds of the slope. The upper one-third of the slip surface rose more steeply before exiting (nearly vertical) at the slope crest.

The north slope had a slip surface, based on limited data, at approximately elevation 502 m.

South slope movements were in a north-northeastern direction, or about 50-80 degrees off pipe centerline, and toward a meander loop in the creek immediately east of the pipeline right of way (ROW).

Ongoing, significant erosion was evident on the creek banks at this location. Movement along the slip plane was in the order of 65 mm/year with about 40 mm occurring during the late spring and summer months.

North slope movements were roughly parallel to the pipeline. Movement along the slip plane was in the order of 12 mm/year based on limited readings.

SLOPE STABILITY ANALYSES

Slope stability analyses were carried out to help assess site conditions. Bishop's method was used to carry out the analyses which consisted of both random composite (circular/planar) and fully specified slip surfaces. Back analysis was used to confirm the stratigraphic model and material parameters before analysis of various stabilization alternatives was carried out.

Results of all the slope indicator data and slope-stability analyses implied the failure profile shown in Fig. 4.

The failure profile was used to assess the effectiveness of several potential slope-stabilization alternatives. The review made clear that solutions involving slope stabilization through dewatering would be inappropriate for two reasons:

Drainage of the perched water tables in the upper colluvial layers would have little effect upon the slide, which is seated much deeper in the bedrock.
Drainage of the bedrock would likely be difficult because there is little excess pore pressure to relieve and the low-permeability materials would take considerable time to drain.

OPTIONS EVALUATED

Several options were considered as solutions to address the ongoing risk to the pipelines as a result of slope movement (Table 2).

One group of solutions was aimed at increasing the stability of the slopes. Logistical concerns, however, because of the remoteness of the site, difficulties associated with winter work, and potential environmental problems made stabilization alternatives unattractive (Table 2).

A long reroute around the valley (Table 2, Reroute D) and directional drill options were viewed as the low risk, low-maintenance options. The major problem was their relatively high cost.

In addition to those options presented in Table 2, monitoring and periodic strain relief of the lines were considered.

The procedure, carried out once in March 1990, consisted of taking the lines out of service and excavating from the slope crest down to the overbend, allowing the pipes to rebound, and then backfilling the pipelines again. This was not considered viable long-term for the following reasons:

It requires several days of down time for each line during peak demand periods (winter).
the site is remote and ground accessible only in the winter; even then construction of long winter access roads is required.
Frequent strain-relieving operations would be required because of fairly high annual slope movements.
It was ultimately decided that the most appropriate solution would be to reroute the pipelines with directional drilling by installing new pipe beneath the unstable zones within the slope.
Significant considerations in arriving at this solution included the following:
Because of the remoteness of the site and access being generally restricted to the winter when the pipelines normally operate at or near capacity, it was assessed that solutions relying on monitoring, ongoing maintenance or having a higher risk of being unsuccessful would be less desirable than even expensive low-risk reroute alternatives.
Because of the particular situation of the Liege Lateral system, where it is operating near capacity most of the time, it would be desirable to implement a solution that does not reduce the capacity or impose any operational restrictions on the existing system (a long reroute would be a constriction on the system).
It was considered that pipe inspection tools would be used on the Liege system within the next 10 years.
In order to avoid future costs of installing additional sending and receiving equipment, it would be desirable to maintain, as close as possible, the current piping configuration through the crossing without having to install additional values.
Soil conditions were appropriate for directional drilling.
Slip planes were well defined and at a relatively shallow depth relative to the valley bottom. Therefore, directionally drilled holes could be easily located below the planes of instability for both sides of the valley (Fig. 4).
The solution was less expensive than the two pipe reroute.

CONSTRUCTION

The directional drilling work was undertaken in early 1992. The project consisted of building and maintaining an 84-km winter access road including eight ice bridges, setting up and running a 50 person winter camp, preparing the drill rig set-up site and the pipe weld-up area, and replacing approximately 1,300 m (1,000 m of which by directional drilling) each of the existing 10 and 16-in. pipelines with two new pipelines of the same diameter.

Grade 359 pipe with 7.1 mm and 9.4 nun W.T. was used for the 10 and 16 in. pipes, respectively. The pipe was externally coated with 0.50 mm to 0.61 mm of fusion-bonded epoxy (FBE). The specified minimum coating thickness was approximately 25% greater than NOVA's typical minimum.

Because this was mainly a directional drill project, there was considerable discussion on whether to have the directional drill contractor as the prime contractor or as a subcontractor to a separate pipeline prime contractor.

It was decided to have the drilling firm as a subcontractor to the prime because of the long access road required; construction required to clear, prepare and clean up the campsite, drill site and lay-up areas; the 600 m of open trench conventional pipeline required at the tie-in locations; and the prime contractor's familiarity with NOVA's contract documents and procedures.

It was thought that these positives outweighed the negative of the increased cost to NOVA because of the prime's increased risk exposure.

The project was bid as a separate guaranteed lumpsum price for each pipeline. The successful bidders were Waschuk Equipment Rentals Ltd. (Waschuk), with Smith International Canada Ltd. (Smith) as the directional drilling subcontractor.

The contract was awarded on Nov. 18, 1991. Winter road construction started on No-,-. 30 and was completed on Dec. 8.

The camp and rig set-up sites were ready by Dec. 12 and 16, respectively, while the camp and drill rig were ready by Dec. 18, 1991, and Jan. 2, 1992, respectively.

Fig. 5 shows the drilling area; Fig. 6, the rig itself and first pilot hole.

The two new pipelines would be placed in a new parallel right-of-way west of the existing lines.

Water supply for the drilling operations turned out to be more difficult than anticipated. With insufficient water available from the creek, two water trucks were required to haul in water from another source, a one-way distance of 13 km.

Waschuk and Smith commenced with the 10-in. pipeline proceeded continuously on two 12- r shifts because all drilling had to be completed before spring breakup with the subsequent loss of the access road.

The 9-1/8-in. pilot hole drilling began on Jan. 3, 1992, at an entry point 200 m south of south valley crest. The design hole profile was well below the unstable zone, about 30 m below the creek at the valley bottom, with an exit point 100 m north of the north valley crest.

Entry and exit angles were approximately 12.

Pilot-hole progress was monitored with a downhole wireline steering package located behind the mud motor.

Additionally, a surface "True Tracker" magnetic grid was set up at both the entrance and exit points. The 200m long grid verified the entry profile and also gave an accurate exit profile before "break through." This ensured the safety of the nearby existing, fully operational gas lines. '

Pilot hole drilling went well except for the problem of circulation loss due to mud rings. Frequently, the trip out of several joints was required to clear the mud rings. This problem persisted even with the use of various mud mixtures and additives.

FIRST PILOT HOLE

The first pilot hole took 5 days to complete and exited 3 m short and 1.5 m right of the target point. Rearning to 17.5 in. was done in one pass and was completed Jan. 12.

The 1,000 m of welded and tested 10-in. pipe was pulled through in one continuous pull. The pipe string had been positioned partially on rollers and partially on layers of built-up snow in order to facilitate pipe pull through (Fig. 8). Pulling the pipe back through the hole took less than a day (Fig. 9).

Drilling the 9-1/8-in. pilot hole for the 16-in. line began Jan. 17 and was completed Jan. 23.

Because rearning for the 10-in. hole had gone well, the contractors decided to ream in one pass to 26 in. The fly cutter stuck in the hole, however, and several attempts to free it failed.

Ultimately, the drill-stem pipe fractured during a pull-out attempt and the fly cutter and several hundred meters of drill stem pipe were lost. The hole was sealed with concrete at both ends and abandoned.

Additional right-of-way was acquired and the drill rig was moved.

Drilling the third pilot hole began Feb. 3 and was completed Feb. 8. The exit point was 10.8 m left of target but on target for length. The hole was reamed to 17.5 in. in one day.

Problems were still encountered during rearning to 26 in., however, as a "hard spot" resulted in another pipe twist off. Pipe was not lost, but several days were required with a 12-1/4-in., a 17-1/2-in., and finally the 26-in. opener to get through the spot.

Reaming was finally completed Feb. 19. The pipe was again pulled in one section and again it took less than a day.

The pull ends of both pipelines were excavated and inspected following pull through. The coating and pipe were observed to be in excellent condition.

Both pipelines were also hydrostatically tested and pigged after installation. No leaks, off-specification roundness or dents were noted.

Final tie-ins were completed on Mar. 5 with contractor clean up completed Mar. 6. Drilling mud was disposed of on-site after testing verified the fluids were environmentally benign.

The project was completed within NOVA's schedule. Total project costs were $5.5 million (Canadian).

Pre and post-construction monitoring has indicated that the directionally drilled pipelines have been isolated from the unstable portion of the slope.

REFERENCES

Savigny, K. W., Harris, M.C., Campbell, J.W.M., and White, J.M. "The Engineering Geology of Soft Upper Cretaceous Bedrock in the Athabasca River Valley near Fort McMurray, Alberta." IV International Symposium of Landslides, Toronto, 1984, pp. 185-190.
Lytiuk, A.T., Ozoray, G., "Hydrogeology of the Pelican-Algar Lake Area, Alberta," Alberta Research Council Earth Sciences Report 801, 1980,
Bishop, A. W., "The Use of Slip Circles in the Stability Analysis of Slopes," Geotechnique, Vol. 5, 1955, pp. 7-17,