Member Login Menu


After a severe earthquake, a section of 48 in. diameter sewer pipeline located beneath a lake bed in Anchorage, Alaska, became buoyant and floated to the surface of the frozen lake. Upon detection of the problem, the Anchorage Water and Wastewater Utility sprang into action to temporarily stabilize the pipe, which continued to convey sewage by gravity despite having deflected vertically by as much as 14 ft. In a matter of mere weeks, the utility implemented a permanent solution while avoiding the discharge of wastewater into the sensitive aquatic environment.

On November 30, 2018, Anchorage, Alaska, experienced a magnitude 7.1 earthquake, the epicenter of which was 10 mi north of the city at a depth of 29 mi. The earthquake caused liquefaction within the bed of Campbell Lake, a local body of water created in 1957 when a dam was constructed at the mouth of Campbell Creek in west Anchorage. Unbeknownst at the time, the liquefaction caused the failure of the concrete anchors holding in place a section of sewer pipeline on the bottom of Campbell Lake. When the pipe was discovered in March 2019, the pipeline owner—the Anchorage Water and Wastewater Utility (AWWU)—acted quickly but carefully to temporarily stabilize the pipeline before conducting a permanent repair, all while maintaining the flow of sewage and avoiding any discharges into the lake.

The partially filled pipe rose vertically before being stopped by lake ice at the surface. Despite the floating pipe's dramatic reverse grade, sewage continued to flow through it by gravity. With the sewage warming the pipe and the large buoyancy force pushing against the ice, over time the pipe slowly broke through the lake ice. However, snow concealed the exposed pipe until warm spring weather revealed it in mid-March. At that time, the AWWU received a report from a resident that a large pipe was protruding through the ice on the lake.

The AWWU responded quickly and determined that the exposed pipe was a 48 in. diameter sewer trunk line made of ductile iron pipe (DIP). About 6,900 linear ft of the trunk had been buried 3 to 15 ft below the lake bed. Constructed in 1989, the sewer trunk comprised Fastite CL 52, a type of cement-lined DIP with an asphalt-coated exterior. It was manufactured by the American Cast Iron Pipe Co., of Birmingham, Alabama (see "Beyond Theoretical Limits"). At those locations along the pipeline at which the natural lake bed provided less than 3 ft of pipe cover, concrete pipe anchors had been installed and backfill had been imported and laid overtop to create pipe cover 3 ft in depth.

During the initial discovery, the AWWU determined that two 20 ft long pipe joints had protruded 8 in. above the ice, significantly deflecting the bell-and-spigot joints. The ice that encased much of the deflected joints appeared to be keeping the pipes from separating even though the joints were deflected well beyond their 4-degree maximum rating. The floating pipe continued to convey sewage, and there were no reports of backups or overflows. Minimal to no lake infiltration appeared to have occurred, based on the fact that flows within the downstream pump station remained unchanged.

The warm weather decreased the thickness of the lake ice to between 14 and 20 in. This was reason for concern because the lake ice was what was holding the pipe joint together; stable lake ice was needed to provide a steady work platform for equipment.

The AWWU's Engineering Division took the lead in managing the response effort and assembled a team of experts to implement a temporary and a permanent repair strategy. It was estimated that 140 linear ft of pipe had moved vertically, based on the length of exposed pipe, angle of joint deflection, lake depth, and invert elevations on recorded drawings. The full extent of the damage was unknown on the first day of work. The team was hopeful that only 140 linear ft had floated to the surface and no other segments were floating and still concealed by snow or lake ice.

The sewer basin served by the trunk line is home to 29,000 residents who generate an average wastewater flow of 2.3 mgd and a peak-hour flow rate of 4.7 mgd. Sewage within the trunk line flows by gravity to a 17.3 mgd pump station on the west end of the lake. If a pipe joint were to come apart the lake water would fill the pipeline, which would then discharge back into the lake; in essence, the lake would operate as the pump station's wet well! A pipe break had the potential to create a lake consisting of 6 percent wastewater after just seven days. Complicating matters further, Campbell Lake is surrounded by residential homes and is approved for use as a seaplane base by the U.S. Federal Aviation Administration. The lake also provides habitat for several species of birds and fish and is a spawning area for returning salmon.

The team conducted a sewer bypass before anything else. Performing any activity on the ice around the pipe, including efforts to stabilize it, could have potentially compromised the ice that held the pipe joints together. Ultimately, the team decided that the most logical bypass would extend 1,600 linear ft and connect the nearest on-land upstream manhole to the nearest downstream underwater manhole. Expected to take at least six days to construct, the bypass consisted of 40 ft long sections of 10 and 12 in. diameter high-density polyethylene (HDPE) pipe fused together to serve as a temporary force main and two 8 in. diameter self-priming trash pumps, one of which was on standby.

The second and third days of the work involved pipe inspection. To understand the full extent of the trunk failure, the AWWU mobilized a dive team to survey the pipe damage by cutting an access point into the ice. However, the silt content in the lake precluded visibility. As a result, the divers had to walk along the side of the pipe to take the necessary measurements. Based on their findings, the team determined that 100 ft of pipe was above the lake bed and the highest point of the pipe had floated upward by 14 ft, with minimal horizontal movement. One of the joints had deflected 14 degrees, well beyond the manufacturer's design maximum.

A remotely operated vehicle owned by the AWWU was used to inspect the underwater downstream manhole to ensure it was not compromised and was in fact a good candidate for receiving the flows from the sewer bypass. The manhole was 6 ft below the lake ice, presenting many logistical challenges. The first was determining the best way to access it. One option involved draining the lake, which would have taken considerable time, posed safety issues, and caused significant property damage. Because lake ice is supported by the water below, removing the water would also have left large chunks of unsupported ice. Private docks along the lake's edge would have been damaged as attached ice lost its support.

Another option involved building a temporary access road from the shore to the manhole and extending the lid of the manhole above the water; the team pursued this option. Because the design required a quick turnaround, the AWWU retained several engineering firms to work separately but simultaneously on design solutions. The engineering firms were also tasked with developing a design to temporarily stabilize the floating pipe once the bypass was operational. Hiring several engineering firms in this manner proved extremely effective. The AWWU received different solutions and recommendations in short order from the firms and synthesized their ideas to move forward.

Because the floating pipe had the potential to separate in the lake at any time, developing a contingency plan was critical. The intent of the plan was to reduce the negative effects associated with a pipe break occurring before the bypass pumping system became operational. Because the surrounding ice was melting, the plan would also apply if additional sections of floating pipe were discovered and found to be separated.

The team sought a way to quickly isolate the compromised underwater pipe from the rest of the collection basin. Downstream of the lake, the trunk line crosses Campbell Creek by means of an inverted siphon comprising a vault with three sluice gates. The vault was not intended to be used as a means to isolate the trunk; rather, it was designed to have at least one gate open at all times.

The AWWU team analyzed the vault to ensure it could withstand the static pressure of the lake. The team determined that the vault would be fully capable of doing so with minor modifications. The ability to isolate the trunk downstream of the lake from the rest of the collection system was valuable because it reduced the amount of sewage in the basin that was at risk from the amount generated by 29,000 residents to the amount generated by 11,000 residents, or roughly 1 mgd. Isolation at this location effectively separated the pump station and several watertight manholes from the lake. If not for the isolation, the watertight manhole lids could have become pressurized up to 10 psi.

With a means to reduce the size of the affected sewer basin, the team turned its attention to mitigating the effects of the remaining 1 mgd of wastewater at risk. Flat hose and a pump were staged upstream at the on-land manhole to divert 500 gal./minute to the adjacent sewer basin across the lake, which was served by an 8 in. diameter collector pipe that had 500 gal./minute in available capacity.

Five small lateral collector pipes fed directly into the upstream pipe on the north side of the lake, each generating a small amount of flow. A 3,500 gal. tanker truck was stationed at each of these locations to collect the low flows, while another tanker truck was stationed at the trunk manhole upstream of the lake to assist in collecting the larger flows. The trucks discharged the sewage into a manhole on a large trunk connected to a separate sewer basin about 2 mi away. The upstream pipe had plenty of storage capacity to attenuate the flow during the daily peak hour. All septage haulers in the Anchorage area were on notice to be available to assist with collection and discharge of wastewater.

The utility was now able to temporarily collect most of the flow in case the affected pipeline separated before the bypass was complete. If additional segments within the lake were found to be floating, the lake would need to be drained immediately by removing two large sluice gates at the dam using an excavator. As a precautionary measure, an excavator was staged at the dam.

The design to access the underwater manhole consisted of constructing a temporary access road to the manhole. First, sections of lake ice were removed by means of a chain saw and an excavator. The road was then constructed by placing geotextile fabric on the lake bottom and placing imported gravel until it extended above the water surface. In this way, the road was constructed in segments, providing the excavator a solid platform from which to remove the next section of ice and build the next section of access road.

To extend the manhole above the water surface, a 7 ft diameter, 3/8 in. thick steel caisson was fabricated. The caisson included a 4 in. wide steel flange on one end. A 4 in. thick rubber gasket was placed on the flange to ensure a watertight seal against the rough concrete manhole structure. A diver cleaned debris off the top of the manhole structure with an air knife, and the caisson was lowered onto the manhole. A large pump was placed inside the caisson to pump out the water. Anchors were installed to connect the caisson flange to the top of the concrete manhole, making a watertight seal. Gravel was placed around the caisson to protect it from large floating ice sheets.

The temporary 12 in. diameter HDPE force main was placed on top of the ice. With the spring thaw, the force main floated on the water because HDPE is less dense than water. Steel piles, 4 in. in diameter, were driven into the lake bed on either side of the force main to protect it from large ice sheets. (Forces generated by the ice sheets could have readily pulled the force main out of the manhole.) Completing construction of the temporary force main took seven days and represented the first big milestone of the pipeline repair effort.

With the sewer bypass pumping system in operation, the next step was to temporarily stabilize the floating pipe before breaking up the ice. If it had not been stabilized before the ice melted, the pipe could have separated and discharged into the lake the approximately 150,000 gal. of raw sewage that remained in the floating pipeline. The AWWU team elected to fabricate a system that would not only stabilize the pipe but also safely push it down to an elevation sufficient to restab the joints-that is, to rejoin the pipe sections properly using their bell-and-spigot joints.

Additional Insight: Beyond Theoretical Limits

Stabilizing the pipe in this fashion allowed it to temporarily convey sewage via gravity, and this was critical. The AWWU required permission from the Alaska Department of Fish and Game to drain the lake for permanent repairs. However, when this would be allowed remained uncertain. Potentially, it could have been as late as October, meaning the temporary bypass pumping system would have needed to remain operational for up to five months. Because the bypass pumping system needed to be continuously monitored, the temporary system would be costly to operate and would create a noise nuisance issue for the surrounding residential area.

Several design firms worked on concepts for temporary stabilization. The selected design was developed by CRW Engineering Group LLC, which has offices in Anchorage and Palmer, Alaska. The design entailed straddling the pipe with four pairs of 12 in. diameter steel piles. The piles were driven 25 ft below the lake bed to provide sufficient frictional force to resist the buoyant forces.

Most of the pile-driving equipment large enough to drive a 12 in. diameter pile was too heavy for the thinning ice. However, the team found tracked pile-driving equipment weighing only 25,000 lb. Crane mats were used to further distribute the load. While the piles were being driven, the lake ice flexed downward, displacing lake water through openings onto the top of the ice. This water was pumped back into the lake at a different location to reduce the overall weight on the ice in the work area.

Once the piles were in place, I-beam cross members with collars on each end were placed on each pair of piles. The collars were oversized so that the cross members could slide vertically, even if the two opposing piles were not plumb. Dual cables were attached to the cross member above the pipe and wrapped around the pipe to secure it to the cross member. Two side-by-side I-beam cross members were placed across the top of the pair of piles. Two thick-walled 20 ft long pipes were placed vertically on top of the cross member above the pipe and in between the two side-by-side cross members to act as a guide. An I-beam cross member was placed on top of the vertical pipes, and two chain hoists were extended from this upper cross member to the side-by-side I-beams. As the workers operated the chain hoists, the vertical pipes pushed down the sewer trunk. To prevent the sewer trunk from moving downward on its own, a chain hoist was placed between the cross member above the pipe to the two side-by-side cross members on top of the piles. (See the figure.)

In this way, the support system was designed to enable the team to push the pipe in a controlled manner while stabilizing it in all directions. The ice around the pipe resisted the buoyancy force, which was as high as 90,000 lb. Once the team began to push the pipe down, the support system counter acted the buoyancy forces.

To reduce the buoyancy forces and the consequence of failure, the isolated pipe was flushed clean with water and then filled half way with lake water by means of 48 in. plugs that were installed in the pipeline upstream and downstream of the damaged section after the temporary bypass had been established. In their centers, the plugs contained a 6 in. diameter pipe with a valve. While the valve of the downstream plug was opened to enable sewage to drain from the pipe, lake water was pumped into the valve on the upstream plug. Workers added a vent to the high point to remove trapped air.

The team determined a sequence for the pipe-pushing procedure, which included placing steel rods that were visible from the lake surface every 15 ft along each side of the floating pipe. This step enabled the team to monitor any horizontal alignment changes while the pipe was being pushed down. If the team observed deflection in the rods, work would stop and efforts would be reevaluated. The team was able to push the pipe 6 ft downward in this manner.

The team placed a monitoring system upstream of the compromised pipe to evaluate the surcharge level continuously. Even though the trunk was 8 ft higher than normal, sewage could safely be conveyed via gravity without backing up into streets or structures.

The adjustable items on the pipe support system were welded in place to eliminate the risk of movement. The pipe was then inspected via closed-circuit television. No leaks were observed, and the pipe was determined to be stable and in sound condition. The team elected to place the pipe temporarily back in service until the final repair could be made. The bypass pumping system remained on standby and would be used again during the final repair.

While one crew worked on the bypass pumping system and the pipe stabilization, a second crew completed inspection of the remaining 5,500 linear ft of pipe underneath the lake. This task began by staking the pipe alignment on the ice using existing GPS points of the manholes.

Every 40 ft along the alignment, workers used a chain saw to cut a slit in the ice, perpendicular to the alignment. A steel pole was placed in the 48 by 3 in. slit and pushed down through the lake bed using a small impact hammer until the pipe was located, and the pole was pushed down several times at each slit until the crown of the pipe was found. The pipe crown was then surveyed every 40 ft, and these survey points were analyzed to check for reverse grade or abnormalities as compared with the drawings. Based on the information collected, the team determined that the pipe had not moved significantly at any other locations. This measurement method proved effective in rapidly obtaining the information needed.

With the pipe stabilized and the extent of damage determined, the team turned its attention toward the permanent repair. On March 27, the team determined that construction would need to be completed before May 10 or held until after fall to meet permit requirements intended to protect spawning salmon. To reduce negative effects to area residents and decrease project costs significantly, the team elected to perform the permanent repair before that date.

However, a design needed to be completed and ready to bid in one week to meet the accelerated schedule. The design for the final repair included:

  • draining the lake
  • constructing an access road from shore to the floating pipe
  • removing 180 linear ft of the 48 in. diameter DIP
  • reinstalling the existing pipe with new gaskets at the proper grade

Several new sections of pipe were ordered just in case any sections became damaged. Of the 180 ft of pipe that was removed, 120 ft was determined to be in good enough condition to reinstall with new gaskets. New, larger concrete anchors were cast in place every 10 ft on center, and 3 ft of cover was placed over the reinstalled pipe.

All the standard permits were needed, including a habitat permit from the Alaska Department of Fish and Game, which required filling the lake no later than May 10. The permit also required that, while draining the lake, biologists transport any stranded fish back to the main creek channel and that any invasive fish be removed.

The following measures were taken to meet the short schedule:

  • The bid advertisement lasted only one week. Despite this, three competitive bids were received.
  • The selected contractor would have only two weeks to drain the lake, complete the repair, and refill the lake.
  • The contractor used a thick-walled boat to break up lake ice so that the lake could be drained earlier.
  • Because the two gravity-dam drains could not empty the lake fast enough, a large pumping system was constructed to continuously pump 25,000 gal./minute for the duration of the project.

Even with the pumping system, the desired water level could not be reached, so a dike was built to protect the work site.

Despite the challenges, the selected contractor—the Frawner Corp., of Anchorage—was able to successfully complete the project as planned and fill the lake in time for the returning spawning salmon run.

This project presented challenges at every step. The probability of failure was high, and the consequences of failure would have been severe. A failure could have resulted in millions of gal. of sewage contaminating a pristine Alaska lake that provides a habitat for numerous animal species along with recreation and public enjoyment. At the same time, a breach in the system could have had catastrophic results downstream, inundating a pump station and overwhelming the wastewater collection system. The response effort and repair work received significant attention from the local media, which increased the pressure to complete the project without incident.

The project succeeded by bringing various contractors and consultants together to generate innovative ideas in a short time. The project's success can be further measured in that there were no injuries, no sewage spilled, and no environmental damage. Additionally, the overall project cost was low, at $2.3 million. Furthermore, the community and permitting agencies agreed that the response effort was completed successfully.

The project also offered several key lessons:

  • After a large earthquake, check manholes for surcharging, which could be the result of a pipe grade change downstream. In the case of the Campbell Lake pipeline, the floating pipe would have been discovered much earlier if this had been done.
  • In an emergency situation, hire several engineering firms to work simultaneously but independently on the same problem. Although this approach may seem wasteful to some, having multiple alternatives added value for the utility in this case. And the ultimate solution may involve a hybrid of ideas developed by multiple firms.
  • Develop a project contingency plan as soon as possible. This can calm nerves and create a productive environment that enables the response team to address the problem under less pressure.
  • When the project is complete, place the contingency plan into a formal emergency response plan. In this way, if the same event or a similar one happens again, the utility has the information on hand to devise a response.

Willie O'Malley, P.E., is a project manager in the Engineering Division of the Anchorage Water and Wastewater Utility. This article is based on a paper that will be presented at ASCE's UESI Pipelines 2020 Conference, which at press time was scheduled to be held August 9-12 in San Antonio.

PROJECT CREDITS Owner Anchorage Water and Wastewater Utility Design engineers CRW Engineering Group LLC, Anchorage and Palmer, Alaska; the Boutet Company Inc., Anchorage and Wasilla, Alaska; VEI Consultants, Anchorage Contractor Frawner Corp., Anchorage

Civil Engineering , May/June 2020, © American Society Of Civil Engineers. All Rights Reserved


Read Civil Engineering magazine on your smart device: download our apps.

app store play store