Long-distance annulus backfilling of a rehabilitated sewer tunnel

The Colsman Tunnel, located in Greenwood Village and Centennial, CO, is part of the Big Dry Creek Interceptor sanitary sewer system and is owned and operated by the Southgate Sanitation District (the district). Flow from the district’s entire waste water collection system (approximately 80,000 residents) is conveyed through the tunnel. Average flows of approximately 10 million of gallons per day are constant and cannot be turned off or diverted.

The tunnel was constructed around 1977 using hand tunneling and road-header equipment. The 2,320-m (7,614-ft) long, mushroom-shaped tunnel is approximately 175 cm (69 in.) wide, 190 cm (75 in.) tall and varies in cross section and shape along its length (Fig. 1). The tunnel was constructed with a slight downward slope of 0.36 percent to the west. At its deepest point, the top of the tunnel is about 27 m (90 ft) below the existing ground surface. Access to the tunnel is provided by a portal structure on the east side of the tunnel (east portal) and a buried outlet structure on the west side of the tunnel (west portal). Four ventilation shafts (designated Vent Shaft #1 through Vent Shaft #4) used during construction are spaced approximately 487 m (1,600 ft) apart along Orchard Road (Fig. 2). Access into the tunnel is complicated by the tunnel’s depth and its location beneath Orchard Road, a heavily traveled arterial roadway adjacent to a busy shopping center and residential areas.

Tunnel ground support reportedly consisted of a variety of different lining systems according to different sources reviewed (Meurer, Serafini and Meurer, 1975; Lachel & Associates, 1991; HDR Engineering 2016). Based on as-built tunnel drawings, a manned entry condition assessment and a remote closed-circuit television (CCTV) condition assessment, several potential tunnel lining systems were utilized, including steel liner plate/welded-wire fabric/shotcrete, welded wire fabric/ shotcrete with rock bolts, and 10 cm (4 in.) wide steel sets placed at 1- to 2-m (3- to 6-ft) centers. The lining is coated in coal-tar epoxy and has experienced multiple phases of spot repairs, making visual evidence of lining type difficult to discern in CCTV inspections performed over the years. For this reason, it is unknown where the different types of reported lining types are located in the tunnel, or whether all of the lining types cited are actually present.

Tunnel condition and need for rehabilitation or replacement. Two manned entry inspections of the tunnel were carried out in 1981 and 1991 (Lachel & Associates 1991). In 2015, a multisensor robotic inspection was also performed (HDR Engineering 2016). This most recent condition assessment revealed that the tunnel needed rehabilitation and/or replacement due to deterioration of the shotcrete lining, ground water infiltration and other structural defects.

Based on the age of the tunnel, the highly corrosive waste water environment, the inability to safely perform internal spot repairs, and lack of system redundancy, the district became increasingly concerned with the ability of the tunnel to provide an additional 75 years of service life. The district engaged Burns & McDonnell Engineering Co. Inc. (Burns & McDonnell) as their owners engineer/ technical advisor to assist with the tunnel rehabilitation. Based on the complicated nature of the work, the district selected a progressive design-build approach. The team of Garney Construction (Garney), Dewberry Engineers (Dewberry) and Shannon & Wilson Inc. (Shannon & Wilson) was retained after a qualifications- and interviewbased selection process.

Tunnel rehabilitation approach and execution. After considerable analysis and evaluation, a solution consisting of tunnel rehabilitation via sliplining with high-density polyethylene pipe (HDPE) was selected. The sliplining process allowed the installation of a fully structural, completely inert, continuous pipe within the tunnel, and was completed under live-flow conditions that eliminated expensive and high-risk bypass pumping. The slipline was completed using horizontal directional drill (HDD) equipment with the capacity to pull the nearly 2,346 m (7,700 linear ft) of heavy-wall HDPE pipe and associated drill stem through the tunnel. The pipe consisted of 122- cm (48-in.) diameter SDR 13.5 IPS HDPE (PE 4710) with a wall thickness of approximately 9 cm (3.6 in.) and a weight of 218 lbs/ft.

Garney retained Global Underground Corp. (Global) as the specialty subcontractor to perform the sliplining operation. Global utilized an American Auger DD-440T HDD machine with a 440,000 pound thrust/pullback capacity set up at the east portal (Fig. 3). A cartridge-style pipe string approach was utilized at the west portal, with fusing operations performed on 15-m (50-ft) long pipe segments (Fig. 4). The work was successfully completed in 42 days in January 2019, averaging four fusions per day for 61 m (200 ft) of pipe installation. More than 1.6 million lbs of pipe was installed.

After completion of installation, flows were redirected into the HDPE pipe, which rested on the bottom of the tunnel. There was an annular space left between the HDPE pipe and the tunnel ranging from approximately 5 to 28 cm (2 to 11 in.) horizontally and 46 to 74 cm (18 to 29 in.) vertically over the pipe centerline. Based on a preconstruction LIDAR and sonar survey of the tunnel, the volume of the annulus space was estimated by Dewberry to range from approximately 6,900 to 7,100 cu yd.

Annular space backfilling

During the design phase, the team agreed that it was desirable to backfill the annular space between the HDPE pipe and the original tunnel to reduce future risks to the pipeline. Annular space backfilling (also known as grouting) would mitigate the consequences of future tunnel lining degradation/collapse and would also provide lateral support to the HDPE pipe. The design intent was to fill the annular space as fully as is practicable, understanding that successfully backfilling 100 percent of the annulus would be challenging for several reasons.

Annulus backfilling is a common practice in the tunneling industry. A variety of materials are utilized for backfilling, including cementitious grout, conventional concrete, low-density cellular concrete or flow fill (Henn, 2003). Which backfill material is selected depends on a multitude of factors, including specific project needs, tunnel geometry and carrier pipe type, to name a few.

Placing annular space backfill is generally accomplished in one of two ways. If the carrier pipe is large enough for manned entry, backfill can be pumped into the annulus through pre-installed grout ports in the pipe. For smaller-diameter pipes, backfilling is performed from outside the pipe with the use of bulkheads. The distance over which backfill can be placed varies and is dependent on material properties and project geometry but is generally no more than 152-182 m (500-600 ft). For this reason, backfilling is typically performed in sections, and access is required at the injection points. Slicklines can be used to extend the length of backfill placement where intermediate access is not available.

For the Colsman Tunnel, the annular space backfill design was complicated by several factors. First, an external slickline was unable to be attached to the HDPE pipe prior to placement due to the damage that would occur from the rotation of the HDPE pipe during HDD pullback. A slickline was also unable to be installed after the HDPE pipe was in place, as there was insufficient room for manned entry. Second, the HDPE pipe was placed into service upon completion of installation, carrying live sewer flows that could not be disrupted. While the pipe could have been designed with pre-installed grout ports to facilitate backfilling from inside the pipe, this option was not considered viable as it would have required costly and risky bypass pumping. Instead, the HDPE pipe was installed during live sewer flows, which were then immediately channeled into the pipe once it was installed. Not only did the operating conditions preclude entry into the pipe, it increased the risks of the backfilling operation which could not damage or otherwise impact the HDPE pipe.

Most importantly, access into the tunnel for backfilling operations was very limited, consisting of two portals and four intermediate vent shafts. The spacing between these six grout injection points ranged from approximately 396 m to 518 m (1,300 ft to 1,700 ft). During the design phase, consideration was given to adding intermediate grout injection points in between the existing vent shafts by drilling into the tunnel. However, the team was concerned about rock or lining debris that could potentially fall into the tunnel during drilling, or the potential for deflecting the tunnel lining inward, either of which could complicate the pull-in of the 122-cm (48-in.) HDPE pipe. In addition, the team was unsure if drilling into the tunnel would be possible due to its depth below the existing ground surface and the uncertainty involved with the type of tunnel lining system that could be expected.

The team selected low-density cellular concrete as the most appropriate backfill material for several reasons. Cementitious grout and the more traditional backfill materials have a relatively high density when compared to the HDPE pipe and would result in pipe floatation and/ or collapse during placement. Furthermore, cementitious materials have a higher heat of hydration, which would cause a significant degradation the structural properties of the HDPE pipe. However, low density cellular concrete made with hydraulic cement, water and preformed foam has a lightweight density and significantly lower heat of hydration. Due to the risks associated with drilling into the tunnel, the design-build team decided to attempt to fully grout the annular space using only the existing vent shafts and portals as grout injection points. While this approach was considered achievable, it was understood to be at the upper limits of the technology due to the large distances between injection points and the relatively flat slope of the tunnel.

Garney retained Cematrix Cellular Concrete Solutions (Cematrix) as the specialty subcontractor to perform the annulus backfilling, after receiving and evaluating bids from multiple specialty subcontractors. For simplicity, “placement of annular space backfill” and “low-density cellular concrete” are generally referred to as “grouting” and “grout,” respectively, for the remainder of this paper.

Initial grouting plan and performance

Grouting plan and equipment setup. The original grouting plan was to inject grout from the upstream end of the tunnel at the east portal and progress westerly down the tunnel to each of the four vent shafts for subsequent placement, terminating at the west portal. Two lifts of grouting would be performed from each location. The first lift would consist of approximately 382 to 535 m3 (500 to 700 cu yd), and would be terminated before the placement location was grouted off to permit the second lift to follow on the next day. The second lift was planned to top off the remaining volume to be placed in that reach and would close out that access point. The planned grouting reaches, including distances between injection points, depths of the vent shafts and estimated theoretical volumes of grout, are presented in Table 1. During grouting, Garney planned to lower a camera into the adjacent vent shaft to look for visual confirmation of grout. Not all of the vent shafts were installed at the crown of the tunnel; therefore, visual monitoring would be complicated by limited line of sight at some locations. The volume of grout placed would also be compared against the theoretical volume to determine if the reach was effectively grouted.

Cematrix developed the following grout mix design to produce a minimum 28-day compressive strength of 100 pounds per square inch (psi) and unit weight of 30 pcf, per project specifications: 500 pounds Envirocore IL (10) MS cement; 250 lbs potable water (0.50 to 0.65 water to cement ratio); and 58 lbs preformed foam produced with Provoton foam agent with 3.5 percent water.

The grouting operation included a dry cement silo and mixing trailer (Fig. 5). Cement was delivered via truck and placed into the silo; approximately 10-12 trucks were required per day to produce approximately 765 to 917 m3 (1,000 to 1,200 cu yd) of grout. The grout was produced in an on-site automated batch plant where cement and water were combined to form slurry. Water and foam concentrate were mixed, then compressed air was added to create the preformed foam. The rates of slurry production and preformed foam production were linked via a central control panel to create the desired density of the finished cellular grout material. The grout was pumped via a 10-cm (4-in.) hose, from the mixing trailer to the injection point. The entire grouting equipment setup was portable and was moved to each vent shaft and portal location for grout placement. Due to the limited space available along Orchard Road, utilizing fly ash in the mix design was not considered to avoid the need for two silos and truck deliveries of two components.

During grouting, Cematrix performed quality control (QC) testing of the grout. Shannon & Wilson provided a full-time engineering technician onsite to perform quality assurance (QA) construction materials testing of the cement slurry, along with part-time engineering oversight (Fig. 6). Testing included Marsh funnel viscosity and unit weight measurements of the cement slurry on at least an hourly basis. In addition, both Cematrix and Shannon & Wilson collected a set (four each) of 15-cm by 8-cm (6-in. x 3-in.) cylinders of grout for compressive strength testing at the same frequency. The hourly testing frequency corresponds to one sample for approximately every 130 cu yd of grout placed, based on the anticipated rate of grout placement.

Initial grout placement. Grouting operations began at the east portal on March 15, 2019. The intent was to backfill between the east portal and Vent Shaft #4 (a distance of 437 m (1,436 ft) in two lifts. The calculated volume of grout in this reach was estimated to range from 878 to 891 m3 (1,149 cu yd to 1,166 cu yd). Cematrix stopped grouting the first lift after placement of 217 cu yd, because the grout had begun to back up into the portal and they did not want to risk grouting off access for the planned second lift. At the request of Cematrix, Garney poured a concrete bulkhead at the east portal outfitted with a grout injection nozzle and air vent to accommodate the remainder of grouting. Grouting resumed on March 18, 2019. As the bulkhead allowed for grout injection to be performed under low pressure (less than 5 psi), Cematrix was able to place 937 m3 (1,225 cu yd) of grout in the second lift. A total of 1,102 m3 (1,442 cu yd) of grout was placed from the east portal, which exceeded the theoretical annular space volume, but no grout was visually observed with the camera setup in the adjacent Vent Shaft #4. It is normal for cellular grouts to experience in-place yield loss due to the dissipation/consolidation (bubble popping) of the air bubbles as the grout travels. The longer the grout travel distance in an annular space, the greater the yield loss. This observation led Cematrix and the design-build team to suspect that the grout may not be performing as intended.

The grouting operation was moved to Vent Shaft #4 on March 21, 2019. Vent Shaft #4 was approximately 46 cm (18 in.) in diameter, and penetrated the tunnel crown near the northern edge of the tunnel, offset from the centerline. The vent shaft was approximately 13 m (44 ft) deep. The 10-cm (4-in.) grouting hose was setup over the vent shaft with a steel frame, which allowed the grout to be injected vertically into the shaft without any applied pressure or use of a tremie pipe (Fig. 7). During the first 115 m3 (150 cu yd) placed, the grout was observed to be flowing both upstream and downstream from the vent shaft. After that, the grout began flowing downstream only, leading Cematrix to believe that it had filled to the leading edge of the grout placed from the east portal. After only another approximately 76 m3 (100 cu yd) of placement, the grout began to back up from downstream and was in danger of closing out the vent shaft. The low slope of the tunnel was effectively flat, and the very lightweight grout may have been mounding up rather than flowing. Cematrix terminated grouting after placing only 204 m3 (267 cu yd) that day.

Cematrix concluded that the flowability of the grout was not as they had expected; the grout was not flowing as far as planned and was gelling faster than expected. Although the slope of the tunnel was relatively flat, Cematrix had reasonably expected that gravity would provide more assistance in moving the grout downhill. For the grout that was placed during this initial operation, Cematrix estimated a volume loss or yield of approximately 11 percent, which was higher than the expected yield of 5 percent. While some loss is inevitable due to the nature of the low-density cellular concrete, as the grout is pumped longer and longer distances it is subject to more mechanical stresses that can affect the structure of the air voids and lead to unacceptable performance. The design-build team elected to halt grouting operations to revisit the current approach.

Re-evaluation of the grouting plan and exploration of alternatives

Although it was considered to be at the upper limits of the technology, the design-build team had anticipated that it would be possible to fill the annular space using only the tunnel portals and the four existing vent shafts as grout injection points. However, based upon the demonstrated flowability of the grout during the first three grout placement days, the team concluded that a different approach was needed. Multiple methods were considered and weighed, and two alternatives were ultimately considered by the project team: 1) allowing for only partial grouting of the annulus and 2) adding additional grout injection points by drilling into the tunnel. Increasing the flowability of the grout by using a heavier mix design was considered but was ultimately rejected due to the increased risk of pipe flotation and damage due to high heat of hydration.

Partial grouting of the annulus. The team carefully considered allowing the grouting operation to continue as is, recognizing that the entire annulus would not be backfilled. If Cematrix were to continue placing grout from the remaining injection points, an unknown volume of grout would be placed at each vent shaft, forming a grout plug. These localized grout plugs would serve to anchor the pipe in the tunnel, thus preventing movement during any surge events. In addition, the grout plugs would isolate the tunnel annular space from corrosive atmospheric conditions, as the space would eventually fill up with ground water seeping into the tunnel. However, the ground water seepage would also serve to float the pipe due to buoyancy, that would result in low spots and sags along the pipe which would reduce its capacity. In addition, future degradation of the tunnel lining and/ or tunnel collapse could damage the pipe in areas where backfilling wasn’t completely performed. For these reasons, the team concluded that leaving the annulus only partially filled was not acceptable to the long-term performance of the tunnel, and that modifications to the grouting plan were needed to ensure the service life of the sewer.

Drilling additional grout injection ports. The designbuild team determined that an additional nine grout injection points located between the west portal and Vent Shaft #4 would be required to reduce the length of the grouting reaches to a maximum distance of 180 m (590 ft) (Fig. 8). Increasing the number of injection points would result in a grouting program more consistent with the conventional 152 to 182 m (500 to 600 ft) flow range and a 306 to 459 m3 (400 cu yd to 600 cu yd) placement range with pressure grouting.

However, there were considerable risks that needed to be addressed before proceeding with drilling into the tunnel, primarily related to two key questions: 1) could injection points be drilled into the tunnel effectively; and 2) could injection points be drilled into the tunnel without damaging the HDPE pipe which was carrying live sewer flows? As previously discussed, the team was unsure if drilling into the tunnel was feasible. There were a variety of tunnel lining systems that were reportedly utilized during construction, but as-built drawings do not indicate where the different materials were utilized. In addition, the team did not know the thickness or condition of the steel liner plate, or if the welded wire fabric and steel sets themselves would be difficult or impossible to drill through or would only partially deflect (which could damage the HDPE pipe).

There were additional risks related to the drilling process itself, including the vertical tolerances of the drilling operation and how much control the drilling subcontractor would have when advancing through the different types of material. The depth to the top of the tunnel was known at the existing vent shafts, but was interpolated between these locations based on the preconstruction LIDAR and sonar survey of the tunnel. For this reason, the targeted depth was only approximate, leading to some uncertainty. Would the driller be able to tell by drilling action when the top of the tunnel was reached? Or would the drill bit readily puncture the lining and keep on drilling into the tunnel, potentially damaging the HDPE pipe?

After considerable discussion between the designbuild team and multiple drilling companies, an air-hammer drilling approach was selected to advance the grout injection points. Based on a combination of suitability of equipment, approach to the work, and availability, Xtreme Drilling (Xtreme) was retained as the drilling subcontractor. Xtreme utilizes small, compact drill rigs with a small working footprint. The compact drill rigs have relatively precise downhole and rotational speed control, and are light enough for the driller to have a good degree of feel that allows him to know when drilling behavior changes. Xtreme recommended using a scratcher bit to drill through overburden soils and bedrock. Once the top of the tunnel was near, Xtreme would change their drilling operation to utilize an air-hammer bit to penetrate through the remaining bedrock and through the lining. The air hammer needs to be in contact with the bottom of the borehole to operate and will immediately stop operation if it were to hit a void as it would lose air circulation. This fail-safe would prevent the drill from advancing into the HDPE pipe.

Drilling field demonstrations. To evaluate the condition of the lining, Garney made an exploratory excavation near the west portal to observe the composition of the tunnel liner. The tunnel liner was found to consist of weak degraded shotcrete without the presence of welded wire fabric or steel sets. While this field demonstration alleviated the concern that the drilling operation would not be able to penetrate the lining, it increased the concern that conventional drilling equipment would not be able to distinguish between the soil/rock and the tunnel.

Garney mobilized Xtreme to the east portal to test the effectiveness of the air-hammer approach. A section of HDPE pipe was placed inside a section of concrete portal roof that had been removed from the tunnel. The annulus between the pipe and the concreate was approximately 30.4 cm (12 in.). The purpose of this test was to verify the reaction of the hammer drill to penetrating the tunnel roof and encountering the void space before impacting the pipe. The hammering action halted by itself as soon as the void space was encountered and the pipe remained untouched, exactly as anticipated. The drill stem was then purposely lowered onto the HDPE pipe and allowed to hammer for approximately one minute, resulting in little more than surface scratching on the pipe.

Grout injection point drilling. Based on these demonstration tests, Xtreme mobilized to the site to perform the injection point drilling. Xtreme was able to drill the additional nine grout injection points in less than a week without incident. Drilling proceeded relatively quickly through the overburden and bedrock until the drill bit was to within about 1.5 m (5 ft) of the targeted top of the tunnel liner, where Xtreme switched to the air-hammer bit. The top of the tunnel was typically encountered within one or two feet of the estimated depth. Drilling through the liner proceeded with relative ease, as no steel sets, liner plates or welded wire fabric were encountered.

Execution of the revised grouting plan

Based upon the behavior of the grout during the first three grout placements, the revised grout plan included pumping the grout under pressure rather than to place grout by gravity and depend upon unpressurized flow. Pressurized flow was achieved by the use of a 7-cm (3-in.) steel casing pipe along the entire length of the newly drilled grout injection points. Pressurized flow was achieved at the existing vent shafts by installing a sewer plug with a bypass hole. The grouting pressure was monitored at the ground surface utilizing a pressure gauge on the injection stack, and the allowable pressure was calculated for each injection point based upon the vertical depth (head) to the HDPE pipe.

Cematrix field-verified grout placement by closely monitoring installation pressures and by using specially developed closed-circuit cameras. The cameras were lowered into adjacent grout injection locations to visually verify when grout had reached the next installation port (Fig. 10). Once the grout was visually confirmed to have reached the next injection location, grouting operations stopped to avoid overfilling and inadvertently sealing off the subsequent grout injection point.

During grout installation, a barometric loop was installed on the downstream end of the tunnel. The barometric loop served to raise the hydraulic grade line within the HDPE sewer pipe so that the entire line was fully submerged (flooded). The submerged line served two primary purposes. First, the filled pipe helped to ballast and weigh down the buoyant HDPE so that it wouldn’t float during grouting operations. Second, the high water in the pipe helped to transfer away the heat of hydration developed during grout curing. Removal of the heat during curing was important to protect the physical characteristics of the new plastic pipe.

Using the additional grout injection points, Cematrix moved numerically from the upstream end of the tunnel toward the west portal. Cematrix set up its operation at four locations along Orchard Road. From these four locations, all 13 grouting locations were able to be reached. Approximately 7,400 cu yd of LDCC was placed and pumped. Records and audits were performed with regard to expected voids to fill and actual quantity of grout installed. Physical properties of the grout were tested and grout cylinders were collected for every 99 m3 (130 cu yd) of grout placed. All of the grout placed met the intent of the specifications with regard to unit weight and compressive strength. At the completion of grouting, the quantity of grout placed was within 5 percent of the predicted volume of void space to be filled. Based on a combination of the visual verification of grout placement using the downhole cameras and the 5 percent variance on estimated grout volume placed, the team was able to conclude that backfilling of the annulus had been substantially completed to the degree practical. Grout loss was estimated to be less than 5 percent.

Conclusion

The Colsman Tunnel project was successfully completed through the initiative, creativity and exceptional problem-solving skills of everyone involved. The teamwork and cooperation among multiple organizations allowed the project to be completed under budget and at a fraction of the cost of other feasible alternatives. The design-build process was without a doubt the most effective procurement method for a project with so many complexities and risks, and Burns & McDonnell and the district were well-served by the process.

References

Dewberry Engineers. 2018. Tunnel Slip Line Plan and Profile. Construction Plan.

Lachel & Associates, Inc. 1991. Colsman Tunnel, Tunnel Condition Assessment Report. Report: Denver, Colorado.

HDR Engineering, Inc. 2016. Condition Assessment & Rehabilitation Program Technical Memorandum for Southgate Sanitation District’s Big Dry Creek Interceptor and Colsman Tunnel. Technical Memorandum: Denver, Colorado.

Henn, R. W., ed. 2003. AUA Guidelines for Backfilling and Contact Grouting of Tunnels and Shafts. London: Thomas Telford Publishing Meurer, Serafini and Meurer. 1975. Orchard Tunnel. Construction Plans.