The University Link project will extend Seattle’s light rail from the downtown Pine Street Stub Tunnel to University of Washington Station (UWS). The 5km of twin-bore, 6.4m excavated diameter tunnels are being constructed under two separate contracts, U220 and U230. The tunnels from UWS to the Capitol Hill Station (CHS) are part of the U220 Contract and use the UWS site for tunnelling operations.
One of the early challenges for the construction of the U220 project was the passage of the tunnels under a canal constructed in the 1910s to connect Lake Union and Lake Washington — the Montlake Cut. The Montlake Cut is located roughly 91m to the south of the UWS and is roughly 30m wide where the tunnels pass below. During the time of its construction, the Montlake Cut was part of a series of projects that would provide a connection between Lake Washington and the Puget Sound for purposes of commerce. Now, the canal is primarily used by recreational boaters.
Local geology
The U220 tunnels were constructed in an area where the present-day land surface reflects glacial sediments deposited during this period. The Quaternary geologic history of the Puget Sound region is dominated by a succession of at least six continental glaciations. Because of the significant ice thickness, the soils in the Puget Sound region are generally overconsolidated.
Explorations for the project encountered deposits from at least three glacial cycles and three nonglacial cycles. The deposits consist of clays, silts, sands, and gravels in various combinations, relative densities, and consistencies. In the area of the Montlake Cut, the borings showed soils that comprised glacial tills overlying a glacial lacustrine clay. The hard glacial lacustrine clay deposits are overconsolidated as a result of the glaciations. The Montlake Cut was excavated out of this clay deposit. Since the original excavation of the Montlake Cut, about 1.5m of sediment has settled on the bottom of the Montlake Cut.
Original tunnel design
The Montlake Cut has an excavated cross section that includes a flat horizontal base with walls that slope roughly at 1H:1V from the base at elevation -5.7m up to elevation 19.8m. In the center portion of the Montlake Cut, where the U220 tunnels pass under, concrete revetment walls are also placed on the slopes and are keyed into the base of the excavation. The walls extend up the slopes on each side to elevation 7m, where a 3m-wide bench creates a sidewalk a few feet above the water level around elevation 5.5m. The reminder of the slope is covered in vegetation.
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataThe preliminary engineering (PE) vertical alignment in the area of the Montlake Cut followed along a positive gradient section, under the base of the northern toe of the Montlake Cut and up to the UWS box. The distance between the top of the sediment in the Montlake Cut and the proposed crown of the tunnels was about 7m. Information available during the PE phase established that the 7m cover would enable the tunnels to be driven beneath any possible sheet pile obstructions that were suspected to be present because of known construction activity that had occurred over the years in the area of the Montlake Cut.
Historical research
Since the PE vertical alignment was being controlled by suspected obstructions due to construction activity in the area of the Cut, a full review of available information was performed during the final design phase to better define this cover requirement and avoid these obstructions. There was a series of slope stability projects along various portions of the Montlake Cut between the years after it opened in the 1920s and up to the latest project in the 1980s. This was a concern because of the soldier piles and or sheet piles present in the area of the U220 tunnels due to these projects. The historical information reviewed at the National Archives and Records Administration facility in Seattle proved helpful in determining the actual location of these possible obstructions. Based on the information available at the archives as well as additional discussions with the Army Corps of Engineers, it was determined that there was no indication that any of the follow-on projects were completed within the area of the U220 tunnels. Since the risk of hitting obstructions was now considered to be low, it was determined that the depth of the tunnels could be reduced and optimised based on the technical merits and were not bound by any obstructions.
Revised tunnel design
Based on the results of the historical data research, the alignment was evaluated to determine how much of a reduction in cover was technically feasible. The evaluation included checking the buoyancy of the tunnel liners, the slope stability of the Montlake Cut, and the predicted settlement of the structures in the area. A more detailed site investigation of the existing subsurface conditions was completed to obtain additional geotechnical parameters to perform these evaluations. Relatively undisturbed Shelby tube samples were obtained from the soil below the base of the Montlake Cut for strength and index testing to gain more information on the conditions of the soil above the crown of the bored tunnels. Based on this information, along with test results from the borings just north and south of the Montlake Cut, it was determined that the alignment could be raised to an elevation that left 4m of cover — 1.5m of the sediment that had settled on the Montlake Cut floor and 2.5m of intact glacial lacustrine clay. This allowed the UWS to be raised an equal amount, which resulted in substantial savings for the U220 project.
Although this revised alignment was determined to be technically feasible from a design standpoint, there were additional measures taken in the contract documents to ensure that the contractor was aware of the sensitive nature of this portion of the tunnel drive. A key item was a required pre- Montlake Cut meeting to ensure that there were not any issues outstanding prior to advancing under Montlake Cut.
TBM’s and tunnelling excavation
For excavation of the U220 tunnels, the contractor, Traylor/Frontier- Kemper (TFK), chose to use two EPBMs and trailing gear manufactured by Herrenknecht, of Schwanau, Germany. The EPB tunnelling method uses active face support to prevent ground settlement by using a unique combination of earth pressure transducers, thrust jacks, screw conveyors, material scales, ground conditioning and backfill grout.
Pressure transducers
The pressure transducers, or EPB cells, are the direct feedback to the operator of the actual pressure in the TBM plenum. There are six cells located in the excavation chamber and two on each screw — one at the forward side and one at the rear.
Thrust Jacks and Screws
There are 16 pairs of thrust jacks on the TBM. The operator uses the thrust jacks to advance the machine (up to 100mm/ min). Thrust jacks are hydraulic devices, and by adjusting the hydraulic pressure in the cylinder, the operator can control the force that the cylinder is applying. Using these two controls, the operator is able to control thrust force and the speed of the thrust jacks.
The 16 pairs of thrust jacks have individual pressure controls to enable steering. There are linear displacement devices fitted within four of the thrust jacks to provide extension measurements of the jacks for TBM advance rate calculation and to determine the orientation and lead of the ring in the tail shield for ring building purposes.
Screw conveyors are used to extract muck from the excavation chamber in a controlled manner. Since the screws act as a sealed mechanism, the muck extraction rate (and thus the pressure in the excavation chamber) can be precisely controlled. The machines for U220 use three screw conveyors. The first screw extends from the bottom of the excavation chamber up to the top of the first gantry. The number one screw can extend and retract 0.7m into the excavation chamber to help facilitate muck pickup and repairs to the screw if required. The second screw is connected to the first screw via a universal joint and conveys muck to the third screw. Using multiple screws, the earth pressure is stepped down so that upon exit of the third screw conveyor the pressure energy of the material matches atmospheric conditions. To control the flow of muck through the screws, the operator reads values from the EPB cells located on the screw conveyors and adjusts the speeds of each screw individually. If not controlled properly, the screw could quickly be jammed with muck.
Belt conveyors and Muck Boxes
Because the TBMs were launched in a semishort mode with only a portion of the backup gantries installed, muck boxes were used to remove the material from the shaft and a load cell from the crane was used to confirm that proper weights were being achieved for each tunnel push. The weights from the load cell were recorded, and the amount of soil conditioner had to be subtracted in order to obtain actual muck weights. The crane load cell was independently verified using and addition load cell under the hook to known calibration weights and rigging.
Backfill grout
The backfill grout is injected into the annulus between the excavated ground and the extrados of the segmental lining. The accelerated grout is made of two components, A-liquid and B-liquid. The B-liquid is sodium silicate in liquid form and acts as the accelerator. During machine advance, the PLC controls both the flow and pressure of the A and B liquids, which are pumped from their respective tanks on the backup gantries to a mixing packer in the shield and out through grout lines that are built within the tail shield.
Tunnel excavation
The typical approach to EPB tunnelling is to use the theoretical combined earth and water pressure along the tunnel alignment to establish upper and lower limits at which the machine will operate. However, it has been found that in some types of ground, the theoretical earth and water pressures are not realised, and operating at the theoretical pressure may cause severe packing and clogging of material in the cutterhead and screws. It was determined that the glacially overconsolidated clay, in which the machines would pass through while mining under Montlake Cut, would be the type of material where the pressures might not be realised. Therefore, to limit the risk of lost ground and prevent packing and clogging of the excavation chamber and screws, the TBMs were operated at actual earth and water pressures. This was particularly important in the area of the Montlake Cut since, because of the low cover, there was an apparent risk of heave if the machine was operated above the actual pressures.
Montlake cut readiness review
Prior to tunnelling under the Montlake Cut, a readiness review meeting was held to discuss a back analysis for tunnel statistics associated with the initial portion of the tunnel drive. In the review session, the data for muck weights, grout takes/ proof grouting, and EPB pressures was reviewed to ensure the machine was operating at an acceptable level.
Since the tunnel muck was being removed with muck boxes, special attention was given to the box weights for each ring. The weights would be recorded from the load cells and after the soil conditioners were subtracted, a three ring average was reviewed to determine if they were within theoretical limits. A three ring average was used due to the storage within the TBM excavation chamber and the three screws and the variation of tunnel pushes, individual rings could produce a higher amount of variation.
During the review session, the backfill grout amounts were reviewed along with the proof grout takes for each of these rings. The theoretical volume of the grout for each ring, which is based on the TBM cutterhead bucket lip diameter, the outside diameter of the precast segmental linging and the length of the ring, is 4,553L. The average for the first 58 rings was 4,771L, a difference of less than 5 per cent.
And finally, the EPB pressures were also examined during the review session. These pressures were lower than the theoretical due to the overconsolidated condition of the material and ultimately, the cutterhead was able to be inspected in free air prior to tunnelling activities under the Montlake Cut.
Tunnelling under Montlake Cut
At the completion of the readiness review and the TBM cutterhead maintenance, tunnelling under Montlake Cut started with the southbound machine. The approach to tunnelling under the Montlake Cut was to maintain the same operations that had successfully taken place in the previous rings before the Cut as well as to monitor the pressures closely to ensure no heave occurred in this low cover. During this period, there was also a heightened level of supervision over EPB pressures, thrust and grout.
During the course of mining, clear bubbles started to appear within the waterway of the Montlake Cut. As the southbound machine passed under the Montlake Cut, small pea-sized bubbles drifted to the surface at certain locations. It is unknown if these bubbles were from the air/soil conditioner being injected by the TBM escaping through the low cover or from disturbing the organic-ridden sediment that creates the 1.5m-thick layer overlying the bottom of the Montlake Cut. The bubbles remained clear for the entire duration of the drive, and dissipated once the machine passed under the south end of the Cut. The bubbles reappeared during the northbound drive and dissipated in the same fashion.
The southbound machine was able to cross the Montlake Cut in six days with an average daily footage of 14m/day. The northbound machine successfully crossed over a five-day period.
Instrumentation data
Along with the settlement points along the alignment, the stability of the Montlake Cut during tunnel excavation was monitored using two inclinometers: inclinometer I-391 on the north side of the Montlake Cut, and inclinometer I-390 on the south side of the Cut.
Both inclinometers are located in between the northbound and southbound tunnels on a concrete pathway along Montlake Cut.
There was movement towards Montlake Cut reported for both inclinometers during tunnel boring. The displacements started near the tunnel invert and reached a maximum just above the tunnel crown.
The movement began as each TBM cutterhead passed the inclinometers, and continued while each tail shield passed the inclinometers.
The movement continued for a couple of days after the second (northbound) TBM tail shield passed the inclinometers, but since then there has been very little reported movement. The TBM tail shield on the northbound alignment passed I-391 on August 6, 2011, but the first inclinometer reading after August 5 was August 8. The maximum lateral displacement reported at I-391 (8mm) was slightly less than the maximum lateral displacement reported at I-390 (5mm) because the depth of cover was less at I-391 and the TBM was moving towards the Montlake Cut at I-391, whereas it was moving away from the Montlake Cut at I-390.
Overall, the stability of the slopes along the Montlake Cut was maintained during tunnel construction, and little surface settlement resulted from the tunnel construction.
The settlement reported for points within 7.6m of either tunnel centreline was in general less than that which was originally predicted
