Use of a Digger Shield to Successfully Complete Tunnel after Ground Conditions Proved Too Adverse for a TBM

Mark Havekost Jacobs Associates

Christa Overby City of Portland

Jim Kabat Michels Corporation

Patty Nelson City of Portland

ABSTRACT: Construction of the 1,829 m long (6,000 ft) Portsmouth Force Main Tunnel in Portland, Oregon, met with difficulties after the tunnel boring machine (TBM) was unable to advance through running soils containing cobbles and boulders. Significant ground loss occurred during efforts to break up and pass the boulders through the cutterhead. Efforts to stabilize the ground and advance the machine were unsuccessful, so the tunneling method was re-evaluated. This paper discusses construction challenges and efforts undertaken to resume tunneling. A plan was implemented to remove the TBM and replace it with a digger shield. TBM retrieval was through a jacked steel casing enveloping tunnel support system and machine.

INTRODUCTION

The Portsmouth Force Main (PFM) Project includes approximately 4,877 m (16,000 ft) of a single 1.68 mm (66 in.) force main that conveys up to 454 million liters (120 million gallons) per day of combined sewage from the Swan Island CSO Pump Station (SIPS) to the existing 1.83 m diameter (72 in.) Portsmouth Tunnel in North Portland. The project was one of the last major components of the City of Portland, Bureau of Environmental Services (BES) program to minimize combined sewer overflows (CSO) into the Willamette River. The project was divided into two segments, Segment 1 and Segment 2, as shown in Figure 1.

Figure 1 Project Alignment

Segment 1 consisted of 2,134 m (7,000 ft) of open-cut construction, and 914 m (3,000 ft) of large-diameter, 2.13 m (84 in.) microtunnel. Segment 2 consisted of 1,829 m (6,000 ft) of conventionally excavated soft ground tunnel through the highland bluff that borders the east bank of the Willamette River. The tunnel contains a 1,676 mm ID (66 in.) fiberglass reinforced polymer (Hobas) force main pipe. The tunnel extends up to 42.7 m deep (140 ft) through sandy Catastrophic Glacial Flood Deposits and Troutdale Formation gravel between the South Portal Shaft and the North Connection Shaft. The tunnel was mined and lined from a shaft at the south end of the alignment. Figure 2 shows the South Portal Shaft configuration and the railroad runs between the shaft and the base of the bluff.

Figure 2. Aerial View of South Portal Shaft and railroad crossing

Construction of the Segment 2 tunnel met with difficulties after the tunnel boring machine (TBM) was

unable to advance through running soils containing cobbles and boulders. Significant ground loss occurred during efforts to break up and pass the boulders through the TBM. Efforts to stabilize the ground and advance the machine were unsuccessful, so the tunneling method was re-evaluated. A plan was implemented to remove the TBM and replace it with a digger shield. TBM retrieval was through a jacked steel casing enveloping tunnel support system and machine. Figure 3 indicates the general timeline for the project elements discussed herein.

Figure 3. Project time line SUBSURFACE CONDITIONS

Geologic conditions along the alignment vary dramatically between the lowland and bluff highland

intervals, as discussed below. During tunnel design, six mud rotary and four rotosonic borings were advanced to evaluate soil engineering properties and develop a geologic profile for the tunnel alignment (shown in Figure 4). Boring locations varied slightly from the final tunnel alignment due to traffic control concerns. The rotosonic boring on the south end of the tunnel alignment is offset approximately 250 feet from the alignment. Although geologic contacts were interpreted between borehole locations, the geologic profile served as the baseline for geologic conditions within tunnel and shaft excavations. Groundwater was expected to be below the tunnel for the entire alignment.

Figure 4. Interpretive geologic profile along tunnel alignment

The Mocks Bottom lowlands in the vicinity of the South Portal Shaft are underlain by fine-grained alluvium that was deposited in an abandoned meander of the Willamette River. Artificial fill, predominantly consisting of sandy dredge spoils, was placed across Mocks Bottom starting in the 1930s to raise the ground surface above the 100-year flood elevation.

Although the bluff slope adjacent to the South Portal Shaft and above the railroad crossing has shown no historic evidence of instability, it was considered marginally stable. Historic and active slope failures in the sandy bluff slope have been identified to the east and west of the tunnel alignment.

The bluff and highlands are underlain by fine-grained flood deposits. These flood deposits generally consist of interbedded sand and sand with silt, with isolated lenses of small gravel. The deposits are on the order of 30.5 to 43.6 m (100140 ft) in the alignment corridor. Ice-rafted boulders are known to have been deposited along with the fine-grained flood deposits during flood events. These boulders consist of igneous or metamorphic rock (granite, gneiss, or quartzite). Boulders were not encountered in the project borings, but one boulder was removed from the North Shaft excavation.

Troutdale Formation gravel deposits and interbedded sand lenses underlie the fine-grained flood deposits. The upper contact of the Troutdale Formation has been eroded, forming an irregular geologic contact. Cobbles and nested cobbles are present within the Troutdale Formation. A probable 0.6 m diameter (2 ft) boulder was encountered in the Troutdale Formation below the tunnel zone in one project borehole. Troutdale Formation gravel is commonly cemented, rarely contains boulders, and has good stand-up time. In contrast, more recent Coarse-grained Catastrophic Flood Deposits, which underlie Mocks Bottom lowlands, are rarely cemented, contain frequent boulders, and have poor stand-up time.

Based on the mud rotary borings, the geology in the bluff was interpreted to be fine-grained catastrophic flood deposits over older Troutdale Formation gravel deposits for the entire alignment; however, standard penetration test (SPT) sampling is not ideal for identifying Troutdale gravel. Gravel deposits were not encountered in the design rotosonic borings, which are best for identifying Troutdale gravel, with the exception of the north shaft boring (below the tunnel).

In hindsight, the south end of the tunnel likely encountered Coarse-grained Catastrophic Flood Deposits; and not Troutdale gravel. This became apparent only after the tunnel encountered un-cemented, bouldery deposits and was ultimately confirmed when four additional rotosonic borings were drilled along the south end of the tunnel alignment during construction. DESIGN REQUIREMENTS

Because of the presence of predominantly sandy soils along the tunnel alignment, ground settlement resulting from ground loss was a major concern. The tunnel alignment crossed under a steep bluff and under a major roadway leading to only entrance to the University of Portland. A 305 mm diameter (12 in.) cast iron water line runs the entire length of the tunnel alignment, and a jet fuel line parallels and crosses the central portion of the alignment. These structures cannot tolerate settlement. The specifications provided two settlement thresholds, referred to as the Action Trigger Level and the Maximum Allowable Movement.

The contractor was responsible for preventing settlement and was required to submit corrective measures, taken when thresholds were exceeded and, and verify that these corrective measures were effective.

A geotechnical instrumentation and monitoring program was incorporated into the project. Instrumentation included surface settlement control point arrays. Each array consisted of three points that are centered above tunnel centerline. At selected locations, the central settlement control point was replaced with a settlement casing to monitor ground settlement below the surface that could propagate to the surface. The first combined surface/subsurface monitoring array was located near the start of the tunnel.

In specifying requirements for the tunneling operations, there were numerous discussions on how prescriptive to be on the machine requirements. In the end, it was decided to allow flexibility for the contractor to select the machine it felt best suited the conditions. Both a digger shield and an earth pressure balance (EPB) TBM were allowed; however, provisions for face control were required for both options.

To help manage risk, the project included a geotechnical baseline report that described the materials expected to be encountered and included baselines for the ground behavior for the tunnel excavation. The anticipated ground behavior through the fine-grained flood deposits was predominantly slow raveling to fast raveling. Boulders were baselined for the Troutdale Formation and in mixed face conditions along geologic contacts.

The GBR also described the successful construction of the existing 2,179 m long (7,149 ft), horseshoe-shaped Portsmouth Tunnel. This tunnel was constructed between 1966 and 1967 using an 2.4 m diameter (8 ft) open-face pneumatic shield in the fine grained flood deposits. Figure 5 shows the shield that was used. Tunnel excavation took eight months to complete, and five men worked at the heading during each shift. This past experience confirmed that the fine grained flood deposits are favorable for tunneling.

Figure 5. Portsmouth Tunnel open-face pneumatic shield (front view) in 1967 photograph RAILROAD CROSSING

A cased crossing was required by Union Pacific Railroad (UPRR) for the force main construction beneath its tracks adjacent to the South Portal Shaft. A minimum of 1.4 m (4.5 ft) of vertical cover between the top of the track tie to the top of the casing was required. The original crossing permit envisioned concrete casing pipe being installed using pipe jacking methods from the South shaft to the northern edge of the (UPRR) right-of-way. The contractor requested a modification to the crossing permit to install a 3,048 mm diameter (120 in.) steel casing beneath the tracks using open-cut methods. This was approved by UPRR because traffic on the track consisted of a single train, no more than once per day Monday through Friday. The steel casing pipe was backfilled in the trench with controlled-density fill (CDF). UPRR personnel replaced the ties, ballast, and track. Figure 6 shows the crossing configuration. The sizing of the casing proved to be an essential criterion for being able to remove the TBM, as discussed later.

Figure 6. Railroad crossing profile for turning under the highland bluff TUNNEL EXCAVATION

The TBM was launched through the casing in September 2009. The contractor selected a Lovat MP104PJ, Series 11200 TBM to excavate a 2,642 mm diameter (l04 in.) tunnel. The TBM consisted of a conventional open face cutterhead with closure doors, internal pressure regulated gates (muck ring) and a conveyor system for the transport of spoil from the face to the muck cars. Figure 7 shows the cutterhead face of the machine looking out through the casing. The tunnel was supported by steel ribs and steel lagging for about 3 m (10 ft) beyond the end of the 3,048 mm (120 in.) steel casing, at which point wood lagging was utilized. A double track switch was constructed inside the casing and through to the shaft, where an empty train of muck cars and flat car with materials, was stationed for transport into the tunnel.

Figure 7. Lovat TBM in South Portal Shaft

The TBM was run in open mode above groundwater with the ability to control fast raveling to running

ground with the muck ring and pressure regulating gates. Temporary support consisted of steel ribs and timber lagging. Although extensive research was conducted to identify Segment 2 buried obstructions, unidentified cobbles and boulders were encountered beneath the bluff slope at the start of tunneling.

TUNNELING DIFFICULTIES Coarse grained gravel deposits containing loose cobbles and boulders were encountered during the initiation of tunneling, leading to significant difficulties for the TBM. These deposits are not typical of the Troutdale Formation and were not identified in project design boreholes. In addition, the boulders were greater in size and greater in number than described in the GBR for the Troutdale Formation. The TBM cutting tools were not configured to break boulders, so the tunneling method consisted solely of pulling boulders through the face of the machine. Early on, the muck ring was damaged and the pressure regulating gates were removed to provide better face access to remove cobbles and boulders.

Several of the boulders encountered were too large to fit through the doors of the head of the TBM and had to be broken up by hand in front of the machine. Boulders that were small enough to fit through the doors also impacted the excavation, since to bring them into the TBM the doors had to be opened as wide as possible, allowing the fine-grained soils to run into the heading. The TBM cutterhead direction also had to be alternated to try to catch the boulders in the cutterhead opening and bring them into a position where they could be broken by hand. These efforts resulted in significant ground disturbance and overexcavation. Figure 8 shows a typical boulder in the cutterhead and the void that formed in front of the machine. Several large sinkholes formed in the bluff slope above the tunnel, as shown in Figure 9.

Figure 8. Boulder in cutterhead opening with ground loss void due to running ground

Figure 9. Sinkhole in bluff slope above tunnel

To keep the TBM going, a remedial grouting program, using polyurethane and cement bentonite grout,

was implemented through horizontal holes in the bluff slope and through the tunnel face to help stabilize the fine grained soils. The process was marginally effective and resulted in extremely slow progress and numerous delays. The tunnel only advanced about 56 m (184 ft) in 8 months of active mining.

Midway through the remedial grouting efforts it was determined that grouting from the ground surface would be a more effective method for stabilizing the soils above the tunnel and preventing additional sinkholes. A surface grouting program was initiated in March 2010. The program utilized sodium silicate grout injected through vertical and battered sleeve port pipes to stabilize the granular soils (Harkins, 2012). The pipes were installed from a hiking trail located mid slope above the TBM. Figure 10 shows the grouting plan and section that was implemented. The grouting stabilized an approximately 21 m long (70 ft) zone behind and in front of the TBM. Upon completion of the grouting, the TBM was advanced an additional 12 m (40 ft) into the grout stabilized zone so that the TBM could be inspected and tunneling methods reevaluated by the construction team.

Figure 10. Grout port array to stabilize ground in front of TBM (Harkins, 2012)

Concurrent with the grouting program, additional rotosonic borings were drilled along the south end of the tunnel alignment in order to investigate the nature and lateral extent of the coarse gravel deposits. The results of the additional borings indicated that the gravel deposits and the upper contact with the Troutdale Formation could extend for another 610 m (2,000 ft). It was concluded that boulders could continue to cause excavation difficulties. The revised geologic profile is shown in Figure 11.

Figure 11. Revised tunnel profile based on additional investigations

TBM EXTRACTION

Several options were considered, including a rescue shaft, retrieval of the TBM through the portal,

retrofitting the TBM, changing the tunnel grade, and underground conversion of the TBM into a simple shield. summarizes the options that were considered. Underground conversion and retrofit options were eliminated because of the extensive ground improvement that would be necessary to stabilize the ground for working on the machine and continuing with the TBM. Options that involved a rescue shaft were eliminated because of cost and adverse impacts to the University of Portland, located at the top of the bluff. The selected plan involved TBM retrieval through a jacked steel casing that enveloped the tunnel support system and the TBM. It was decided that the TBM would be replaced with a digger shield of similar diameter. This option would be least cost and could be executed quicker compared to the other options. The option was feasible

because the 3,048 mm (120 in.)casing size selected by the contractor was large enough to accommodate an inner casing that was larger than the machine.

Once the tunnel and casing were stripped of all materials and equipment, including the tunnel rails, a 2,997 mm (118 in.) diameter casing was jacked through the 3,048 mm casing and along the outside of the tunnel from the South Portal Shaft. The casing encapsulated the initial ground support and the majority of the TBM. The portion of ribs and lagging that was covered by the new casing was removed at the end of each advance. Since the tunnel began to curve in the area were the TBM was located, the casing joints were designed to float using connection tabs to allow the casing to negotiate the curve. Bentonite injection ports were included in the casing for lubrication.

The TBM was placed on rollers and moved to the beginning of the horizontal curve section using the TBM thrust rams to pull the TBM. From there the TBM was removed from the casing using a crane and pulleys to tug on a cable attached to the back of the machine. After one of the TBM rollers broke 10 feet from the portal, the thrust cylinders on the pipe jacking frame was used to complete the extraction.

After the TBM was pulled back from the face, the casing was advanced to within 1 foot of the tunnel face. Figure 12 illustrates the three-step process of casing installation (Step 1), TBM retrieval (Step 2), and casing advancement to the tunnel face followed by the installation of the new digger shield (Step 3). Figure 13 shows the TBM exiting the casing in the South Portal Shaft.

Figure 12. General sequence for TBM removal and digger shield launch

Figure 13. TBM after extraction from casing

DIGGER SHIELD EXCAVATION

Used and new digger shields were considered. The key objective was to find a shield that could fit

through the casing and that would be compatible with the steel ribs and lagging support materials already procured for the project. A Herrenknecht DA2640-MH2S tunneling shield was ordered in July 2010. The digger shield was equipped with a hood and an excavator arm and a hydraulic belt conveyor. Upgrade options were also selected, including a sand table, stabilizing fins, and a compact drilling system. The shield and appurtenant equipment cost around $2.4 million. The digger shield arrived at the project site at the end of October and was underway in December.

Figure 14. Herrenknecht open face digger shield with excavator arm

Initially, the digger shield excavation progressed slowly. Concerns about overexcavation and sinkholes, carried over from the previous TBM experience, caused the operators to implement a procedure that required them to stop mining as soon as running and raveling ground was evident and perform remedial grouting from the face. Although the grout provided temporary relief from ground loss, it also bound up the digger shield, requiring significant effortssteering and thrustto break it free and continue mining. This resulted in an extremely low tunnel advance rate. Since running and raveling ground was anticipated throughout the entire alignment, these digger shield operational procedures were not sustainable. The start-stop procedures also made accurately monitoring muck volumes difficult.

The operational procedures were modified to place more emphasis on forward thrust and controlled muck removal, followed by systematic void filling behind the shield. This change required that the operators become confident in their abilities to operate the digger shield in ground that is not improved with grout. The sinkholes developed quickly because of shallow ground cover and significant overexcavation by the original TBM. In the deeper ground it took several days for settlement to propagate enough to be observed in a multilevel settlement monitoring point after an overexcavation event with the digger shield. This indicated that the material above the tunnel started to effectively bridge/arch and settlement risk could be significantly reduced by quickly filling (within 2 or 3 shifts) the overexcavated annulus or void above the tunnels initial support. The deeper ground cover and improving ground behavior as the tunnel got into more uniform fine grained flood deposits increased the operators confidence balancing muck removal and digger shield advance rate. The advance rate improved and settlement was not observed for the remainder of the tunnel.

The contractor used a VMT system to consistently log the location of the digger shield and provide data daily. As a quality control check, the contractor bored a 152 mm (6 in.) hole from the ground surface through the tunnel to ensure the tunnel was on the correct alignment prior to the last curve into the North Shaft for retrieval. The alignment of the tunnel was surveyed using this down hole, and minor adjustments were made to the machine heading prior to reaching the North Shaft. As a result, the machine entered the North Shaft on the correct alignment within the break-in window of the liner plate shaft wall.

The North Shaft was an elliptical shape supported by ring steel and steel lagging. The shaft was constructed over and around the existing Portsmouth Tunnel. The elliptical shape allowed for clearance to remove the tunnel shield without disturbing the exposed existing tunnel and thus eliminated a sewer bypass.

The contractor broke through the North Shaft on April 29, 2011. The advance rate ended up averaging 14.1 m (46.4 ft) per day, even considering the slow start and operating difficulties early in the tunnel drive. The total construction cost for the tunnel, force main, and shaft structures was around $30 million.

Figure 15. Digger shield after hole through at North Shaft CONCLUSIONS

Good communication and partnering with the contractor helped the parties find an innovative solution to resolve the Differing Site Condition encountered during the tunneling. When it became evident that boulders were being encountered well above the conditions described in the GBR and the boulders could not be effectively extracted through the machine face without causing loss of ground in the surrounding soils, the project team agreed that the best way to retrieve the TBM from under the hillside was to grout a zone around the machine, then jack a steel casing into the hillside around the existing tunnel and pull the machine out through the South Portal Shaft. This method proved very effective, and the machine was successfully retrieved and replaced with the digger shield, a more appropriate tunneling machine and method for the boulders encountered. This method was successful because the contractor selected a starter casing with sufficient size to jack an inner casing and because the TBM had only traveled a short distance.

There were several lessons learned on this project. Since boulders were anticipated, it would have been beneficial to provide more specific requirements for the tunnel machine to handle the boulders and/or require to the contractor to submit a specific work plan for pulling boulders through the face in advance of tunneling, including measures for stabilizing the ground around the boulder as it is being removed. This in turn could have assisted with procuring a more appropriate machine for dealing with the boulders or for at least having a plan in place to reduce the time to implement the ground stabilization measures required to advance and ultimately retrieve the machine. Other considerations for projects facing similar ground conditions include the following:

Rotosonic drilling methods with continuous soil core recovery are effective for characterizing coarse grained deposits, including identifying the size and quantity of cobbles and boulders.

Select alignment with options for surface grouting. Install settlement casings early in drives and after change in ground to monitor for ground loss during

the initial tunnel excavation learning curve. Consider aggressive cutterheads for EPB tunneling in ground with cobbles.

REFERENCES Harkins, B.M., and Parmantier, D.M. 2012. Stabilization of running granular soils in TBM alignment with

sodium silicate grout. In Proceedings of the 4th International Conference on Grouting and Deep Mixing. American Society of Civil Engineers.

IntroductionSubsurface ConditionsDesign RequirementsRailroad CrossingTunnel ExcavationTunneling DifficultiesTBM ExtractionDigger Shield ExcavationConclusionsReferences