Douglas Pass, Colorado, USA: Williams Gas Pipeline West, formerly Northwest Pipeline, provides natural gas to customers in eight western states. Its main pipeline traverses some 3,900 miles (6300 km) of diverse terrain and encounters many geologic hazards. To protect the pipeline, the company has implemented a comprehensive program for identifying, evaluating, monitoring, and mitigating these hazards.
The story below tells about the company's experience at Douglas Pass in 1980s. This experience was the starting point for the monitoring and mitigation program that the company uses today.
The company built its pipeline across Douglas Pass in 1955. In April, 1962, a large landslide disrupted a large section of pipe on the south side of the pass. Relocation of the pipe required 4400 feet of new pipe. In March 1963, another landslide forced relocation of 3200 feet of pipe on the north side of the pass. For the next 16 years, the area was stable, and no disruptions occurred. In 1979, however, another landslide forced relocation of another 1200 feet of pipe.
In 1980, the company authorized their geotechnical consultant, Golder Associates, to conduct geotechnical investigations to determine the magnitude of the landslide problems. In 1983, engineers from the two companies began experimenting with instrumentation with the intention of learning how to predict potential failures early enough to allow mitigation measures to be taken. During the summer of 1983, the company installed inclinometers, piezometers, and vibrating-wire strain gauges in an area that had experienced several slope failures.
Monitoring began in the fall and continued throughout winter. Only small slope deformations and changes in pipeline stress were detected. However, in the spring of 1984, rapid melting of the snowpack combined with heavy rains to reactivate the slides. Rockfalls and small slides closed the state highway for a few days, and later, a large landslide closed the highway for several months. Comparable conditions existed along the pipeline.
With the spring snowmelt, a number of piezometers showed piezometric levels above ground surface. Inclinometers measured large increases in slope deformation rates, and strain gauges showed increased pipeline stresses that correlated with slope deformations.
When strain changes accelerated during March, engineers decided to excavate a trench along the upslope side of the pipe to provide stress relief. As the trench was excavated, the pipe could be seen moving in the upslope direction. Measured strain then decreased substantially. Soon afterwards, however, saturated ground moved into the trench. The ground continued to move, and when strains increased again, a second trenching operation was begun. After this second strain relief effort, strains decreased and remained relatively constant.
By October of 1984, the total measured movement of the landslide had exceeded 10 feet, with nearly 5 feet of movement occurring in April. Without the stress relief operations, the pipeline would almost certainly have ruptured. This was a clear demonstration that monitoring could be used to predict potential failures in time for successful mitigation.
Lessons from the Monitoring Program
Approach to Monitoring
The best approach to monitoring includes identifying the hazard, evaluating the risks of the hazard, designing the monitoring program as an element of mitigation, implementing trigger levels and contingency plans, and reviewing data regularly.
Instruments should be placed where maximum changes in displacement or stress are expected. A knowledge of the local geology, such as the boundaries, depth, and relative stability of landslides, can guide the placement of instruments and also helps in the interpretation of results.
Inclinometers are useful for detecting the onset of movement. They also reveal the precise depth of the slip plane and whether multiple slip planes are present. When installed close to the pipeline, inclinometers show displacements similar to that of the pipe.
Preliminary geotechnical studies at Douglas Pass identified the boundaries and approximate depths of the individual landslides in the area. This information was used to select suitable locations and depths for the inclinometers. The bottom of the inclinometer casing must be anchored in stable ground below the slip plane of the landslide so that movements can be referenced to a stable point.
Inclinometer surveys were obtained about once a month during periods of low activity and more frequently during the spring, when wet weather reactivated the slides and increased the rate of movement.
Landslide movement eventually pinches the inclinometer casing, preventing further surveys with the inclinometer probe. Two or three inches of movement across a narrow slip plane can close the casing. However, it is possible to extend use of the casings by converting them to simple wire extensometers.
Wire extensometers are useful for monitoring large movements. Substantial movements generally occur before pipeline failure, so the reduced accuracy of these extensometers is not a serious limitation.
To convert inclinometer casings to wire extensometers, company engineers attached a wire to an anchor and pushed the anchor into the casing to a depth below the slip plane. The other end of the wire exits the top of the casing. Landslide movement is monitored by measuring the length of wire that is pulled into the casing. As reported above, one of the wire extensometers measured movement of 10 feet.
The depth of the slip plane is determined by the inclinometer, so it is best to start with inclinometer surveys and then convert the casing to an extensometer as required. When large deformations are anticipated, it is advisable to convert the inclinometer to an extensometer soon after the slip plane is located.
Standpipe piezometers are used to monitor pore-water pressure and the effectiveness of drainage measures. Rising pore-water pressure is a useful indicator of impending landslide activity.
Standpipe piezometer measurements are obtained with a water level indicator or with a VW wire pressure transducer. Although the pressure transducer is more expensive, it can be connected to a data acquisition system for unattended readings.
Surface surveys are used at the most active sites to monitor movement of monuments installed in the ground or survey targets welded to the buried pipe. Survey costs can be high and it is not always easy to relate surface movements to pipeline integrity.
When the pipeline crosses a landslide, it is subjected to shearing forces at the lateral edges of the slide. This can result in bending and subsequent rupture of the pipe. When the pipeline is aligned with the landslide, it will be subjected to compressive and tensile stresses by the downward moving soil. Compressive stresses cause buckling and rupture. Tensile stresses rarely cause failures.
VW strain gauges are used to monitor strain in the pipe. At each monitored location, three strain gauges are mounted 120 degrees apart and oriented to measure strain in the longitudinal axis of the pipe. The distribution of tensile and compressive strains reported by the three gauges reveals how the pipe is being deformed.
Strain gauges should be installed on new pipe or pipe that has been effectively stress-relieved by excavation or cutting. This provides baseline readings for later calculations of stress in the pipe. The maximum strain in the pipe is calculated from the three measurements and can be substantially higher than the individual strain readings.
Trigger levels for stress relief or other mitigation measures are based on allowable longitudinal stress. This provides a safety margin to account for such uncertainties as initial stress condition, future deformation of the pipe, corrosion effects, and the fact that the strain gauges may not be located at the point of maximum strain.
Good applications for strain gauges are monitoring pipelines affected by slow-moving landslides and for monitoring the effectiveness of mitigation measures. Strain gauges are not appropriate for monitoring fast-moving hazards such as mudflows or rockfalls.
Mitigation by Trenching
A variety of measures are available to protect the pipeline from a landslide. The most expensive and time consuming measures are stabilizing the landslide or relocating the pipeline. The size and number of landslides at Douglas Pass made such measures unfeasible.
Trenching is a reliable mitigation measure that can be completed quickly at modest cost. It involves excavating a trench along the upslope side of the pipe, when the alignment of the pipeline is perpendicular to the direction of ground movement. Trenching provides rapid relief, as documented by strain gauge monitoring. The effect is also visible: the pipeline shifts upslope in the trench by several feet immediately after trenching.
Trenching is less effective when the pipeline is aligned with the direction of landslide movement, and it is not effective at all for rapid, catastrophic landslides. However, landslides of this character are always preceded by significant increases in the rate of deformation. If the area is properly instrumented and monitored, the landslide can be predicted, and emergency measures can be undertaken prior to the failure.
The monitoring and mitigation techniques implemented by Williams Gas Pipeline West at Douglas Pass were successful in preventing pipeline failures under exceptionally unfavorable geologic and climatic conditions. These techniques are still being used today as one component of the company's comprehensive program for identifying, evaluating, monitoring, and mitigating geologic hazards to its pipeline.
Credits and References
Thanks to Graeme Major of Golder Associates for providing the information used for this story. Mr Major has been monitoring pipelines in the western US for more than 20 years and is a co-author of the two papers listed below. Thanks also to Jill Braun of Williams Gas Pipeline West, and to Oil and Gas Journal.
Braun, Major, West, and Bukovansky, "Geologic hazards evaluation boosts risk-management program for Western US pipeline," Oil And Gas Journal, November 9, 1998, pp 73-79.
Greenwood, Bukovansky, and Major, "Line-monitoring instruments prove effective for western US areas subject to landslides," Oil and Gas Journal, February 17, 1986, pp 69-73.