When roads and railways cross on the same plane, one eventually has to give way. In Boston, Massachusetts on the Central Artery/Tunnel (CA/T) Project, it is the new highway that has given way to the sets of rail tracks that have been serving Boston's South Station railway terminal for more than 100 years. Engineers on the mega CA/T project have achieved the necessary gradient separation by going underneath and using a truly innovative solution.
Tunnel box jacking is not a new idea, nor that complex in concept, but on the scale used in Boston, that is new. It took a visit to the site to appreciate the undertaking and realise how and why the project measures up to something special.
Other less expensive and less massive methods for achieving the same end result, including other bored tunnel alternatives, were considered during the preliminary planning and design phases, but all things considered - ground conditions, location, logistics - the selected tunnel jacking method is meeting the widest range of criteria in Boston. One of the most important of these has been to promote and maintain a high degree of confidence in third party owners throughout the planning, design, construction and life-time of the facility. The third party owner and operator in this case is the Massachusetts Bay Transportation Authority (MBTA) that owns the railway facilities, and AMTRAK which maintains the tracks outside South Station on which pass several hundred trains daily.
The jacking concept is being applied in Boston to create three separate sections on the new subsurface I-90 highway and as part of the overall $13.8 billion CA/T project. Each jacked section carries two highway traffic lanes with an emergency breakdown lane, and each has an internal clearance of about 7m high x 20m wide. Two of the tunnel jacks take the eastbound and westbound carriageways of the I-90 highway under the rail tracks while the third is a two-lane I-90 westbound off ramp passing under the tracks.
The tunnel jacking application in Boston is one of the largest uses of the technique in the world and one of the first applications of the method in the United States. The three tunnel sections comprise eight separate precast concrete elements of between 20m to 50m long, 11.5m high, and up to 24m wide, with heavily reinforced concrete floors and roof decks of 2m thick and walls of 1.8m thick.
The off-ramp tunnel is 51m long and comprises two precast jacked elements, while the two carriageway tunnels are 79m and 116m long and comprise three elements each. Each element for the tunnel crossings is cast in large slurry wall supported jacking pits, a major construction undertaking in itself, and is being jacked into position some 2.5m to 9m beneath the rail tracks. At all times the tracks must remain in service. Any interruption to train services would result in heavy penalties to the designers, contractors and highway project owners.
In preliminary planning and design, the rail underpass was envisaged as a staged cut-and-cover operation moving each set of rail tracks one-by-one and building the highway beneath in short increments. The non-disruptive tunnel jacking alternative proposed by Hatch Mott MacDonald was approved by CA/T project designers Bechtel/Parsons Brinckerhoff, and it was on this proposal that contract documents were released.
In early 1997, the Slattery Skanska/Interbeton/J F White/Perini Corp JV was awarded Contract C09A4 as the lowest of three bids at $398 million. "In addition to the jacked tunnels, the contract includes extensive cut-and-cover tunnels, boat sections and elevated structures which comprise the huge I-90/I-93 highway interchange," explained Gary Almeraris, Operations Manager for the JV. "This is the largest contract on the CA/T project, and all its construction activity must be managed and undertaken within the confines of a very restricted project site. The site is hemmed in on each side by the existing I-90 and I-93 highways, the Fort Point Channel, South Station and its railway tracks, as well as within close proximity to seven neighbouring CA/T construction contracts. It is a very challenging project."
After gaining access to the sites, the first task in the tunnel jacking section was to excavate the three large casting and jacking pits adjacent to the rail tracks. These pits are 15m to 18m deep, and are supported by slurry walls. A massive reinforced concrete slab of up to 1.5m thick provides a base to the pits and a suitable foundation on which to cast and then jack the elements of each tunnel section. A 6mm thick steel base plate provides the interface between the concrete of the base slab and the underside of the element to be moved. The back walls of the pits are designed to withstand thrust forces of 11,000 tonne to 12,000 tonne needed to push each element forward and ultimate loads of 18,000 tonne to 19,000 tonne.
Excavation of the pits gave the JV immediate experience of the geological conditions expected ahead of the tunnel jacking. Historical records and preliminary site investigation data in the contract documents had already predicted complex, mixed and unpredictable conditions. The rail tracks are built on reclaimed land under which are buried the granite blocks of the original sea front retaining wall and other timber pile-supported structures. The fill is primarily granular but contains various debris including chunks of concrete, steel, wood and abandoned rail track.
Below the fill are extensive continuous deposits of organic silt with fine sand and some peat. Underlying this, and comprising the thickest soil deposit at tunnel alignment, is marine clay with a section of softer clay beneath a deposit of stronger and less compressible clay at the top. The groundwater level is within 3m and 1.8m of the surface.
A major concern during design was stabilisation of these weak soils. Excavation of the material ahead of the elements was not a problem, but controlling the loss of ground at the 278m2 tunnel faces was critical, not only to limit surface settlement under the tracks, but also to maintain line and level accuracy during jacking.
Initially, the baseline design anticipated soil stabilisation using chemical grouting and dewatering in the fill materials, horizontal jet grouting in the organic deposits, and soil nailing in the marine clay. The JV, however, was concerned about the risks associated with these methods and suggested ground freezing as a value engineering alternative.
"It took about a year to progress from impossible to approval," said Peter van Dijk, Tunnel Manager for the JV, "but the advantages in the end were convincing. Ground freezing provided a complete treatment of the soil mass prior to the start of tunnel excavation, it lowered significantly the risk of ground losses through 'windows' in the original treatment methods, and it improved stability measures by encapsulating large obstructions until cut up and removed. Freezing was therefore approved, and permission was granted by the railway authorities to gain access to the track areas for installation of freeze-pipes and brine-circulation systems well ahead of excavation activity."
Specialist sub-consultant Mueser Rutledge of New York designed the freeze programme, and ground freezing subcontractor Freezewall was engaged to undertake the installation. More than 1,700 freeze pipes were installed vertically and between the rail tracks. Calcium chloride (brine) at -30°C maintained the freeze. Being so close to Boston Harbor and the Fort Point Channel, the ground water is salty and -1°C is its freeze temperature. Once established, the frozen ground was at about -10°C and had an in-situ ground strength of about 5N/mm2 to 10N/mm2, about the same strength as lean concrete. It took about 3 to 4 months to establish the freeze and about 6 months to thaw completely once the freeze is discontinued.
Project: The Central Artery/Tunnel (CA/T)
relocation highway project, Boston, Massachusetts.
Contract tender value: $398 million
Client: The Massachusetts Turnpike Authority
Overall project construction management/design and supervision: Bechtel/Parsons Brinckerhoff JV
Section designer: Maguire/Fredrick R Harris JV
Specialist tunnel designer: Hatch Mott MacDonald
Contractor: Slattery Skanska (New York)/Interbeton (The Netherlands)/ J F White and Perini Corp (Massachusetts) and with specialist advisors Edmund Nuttall and John Ropkins Ltd of the UK
Ground freezing consultant: Mueser Rutledge Consulting Engineers
Ground freezing subcontractor: Freezewall, a division of Moretrench American Corporation
The freeze pipes are installed on a 2m to 2.5m grid, and the frozen zones extend to about 2m beyond the side walls of the tunnel elements, to about 1m below ground surface, and about a meter above the base of the elements. "We left it above the tunnel invert because we didn't want to alter the properties of the clay," explained van Dijk. "Freezing followed by a thaw would induce clay degradation, and the clay in its natural state is a favourable foundation for the heavy tunnel elements once pushed into place."
Expansion of these particular ground conditions when frozen is about 120mm on all sides, or about 1% to 2% by volume, and beneath the rail tracks a maximum heave of 175mm was specified.
"During the planning and design stage, many of the concerns put forward by the track owners and service operators (MBTA and AMTRAK) hinged on how the freeze has to be installed, what are the consequences of a freeze operation, and the presentation of an acceptable contingency plan, should one be needed to slow down the freeze," explained Phil Rice, Resident Engineer on the contract for Bechtel/Parsons Brinckerhoff. "Together with MBTA and AMTRAK engineers and managers, in what has become a successful partnering relationship, these concerns were thoroughly investigated and managed to the acceptance of both the highway and the railway authorities."
"In fact," Rice continued, "jet grouting is being used extensively on other parts of the CA/T project and it has been noted that grouting can cause heave much more rapidly than freezing. Heave caused by grout injection can be 50mm in minutes, and there is little control on where the grout will go. With freezing, the change in level is about 2-3mm/day and the area of influence is well defined. There are a total of 1,200 survey points in the project area, including across the railway tracks. About 300 of these are read daily and the data is analysed by computer, held in the overall project database, and printed out as readily appreciated information. Crews of experienced railway maintenance workers are on 24h duty to undertake any necessary tamping of the ballast to correct rail tolerances."
Meanwhile, the two elements of the shortest of the three tunnels were cast at the same time in its jacking pit and these were ready to start jacking in October 1999.
Within the JV, a considerable degree of experience and technical know-how was contributed by Edmund Nuttall Ltd of the UK, a sister company of Interbeton within the HBG Group of the Netherlands. Edmond Nuttall has completed five tunnel jacking projects in the UK working in close co-operation with John Ropkins, a consulting engineer in the UK. The most recent of the five projects was the 50m long x 23m wide x 9.5m high culvert under the main railway near Maidenhead. This is believed to be the first tunnel jacking operation to include ground freezing for stabilisation. Through various agreements, Edmund Nuttall and John Ropkins are working to promote and execute tunnel box jacking world-wide through the HBG Group. Technical input at tender stage by Nuttall Ropkins to develop the freezing technique, the excavation methods and the open-face shield concept and design, in conjunction with box jacking experience, secured the contract for the JV and, through an ongoing commitment, has contributed to its success.
A major contribution to the project was the innovative anti-drag system (ADS) developed and patented by Ropkins. The concept is based on creating a 'carpet' of closely spaced, 19mm diameter steel cables across the top and bottom of the boxes to create a stationary separation layer between the moving box and the overlying and underlying ground. These cables, about 1,000 across the top and another 1,000 across the bottom, are anchored using crimped-on copper sleeves to plates attached to the thrust pit headwall bracing beams, and are paid out from inside the element as the tunnel is jacked forward. The free end lengths of the cables are stored on the inside of the tunnel elements, on spools for the bottom cables and in brackets in the ceiling for the top sets. It is via slots cast into the leading edges of the shield at the front of each tunnel box that the cables are freely paid out as the boxes move forward. Grease is fed continuously to the cables through lubrication ducts cast into the concrete, to prevent snagging and further reduce drag.
Enough steel cable to see the tunnel boxes through to their final positions is needed in the cable reserves at the start of each jacking operation. A total of more than 400km, or 1.5 million feet, of steel cable is needed to complete the three tunnel lengths. Once the tunnels are in position, the steel cable will be left in place and sacrificed. It would be a difficult, time-consuming and uneconomical operation to retrieve the buried cables.
As well as preventing drag, the carpet of cables on the underside of the elements prevents the ground being compacted and produces the effect of 'tracks' on which the box can slide. The cables across the top deck are designed in addition to transfer friction forces between the roof concrete surface and the greased cables to the cable anchors. The exact load on these upper cables is difficult to predict, but based on past experience and a conservative calculation, the load was estimated to be a maximum of 1.7 tonne/cable for the shorter ramp tunnel and 2.7 and 3.4 tonne/cable for the two I-90 carriageway tunnels.
To excavate the 278m2 face of the tunnels, the leading element of each is preceded by a 3.3m long concrete shield which is fitted with a steel cutting edge and is divided into 12 excavation 'cells' by vertical walls and a temporary mezzanine working deck.
The frozen ground is excavated for the most part by four high-powered purpose-designed Webster Schaeff roadheaders from the UK, two working on the invert and two on the mezzanine. A Brokk hydraulic jack hammer unit is available on both levels to break up granite boulders and harder obstacles. Excavated material falls from the mezzanine to the invert, where a Gradall excavator and a CAT 906 loader gather excavated material into a central muck pile. From here a 42yd3 Wagner scoop tram transfers muck to a 4m3 skip at the back of the thrust pit which, when full, is lifted by crane for discharge on the surface.
"It is a machinery intensive operation," said Steve Leius, Tunnel Superintendent for the JV, "and the fleet also includes a man lift on each deck and a 15 tonne cherry picker on the lower deck. With the machinery we have 28 miners, operators and electricians in each tunnelling shift and work progresses on a two 10hr shift/day, 6 day/week schedule. Progress is based on completing one 3ft cycle/day, excavating a 3ft round out ahead of the face first, and then jacking the box forward before starting the next excavation sequence at the beginning of the next day."
As excavation advances, individual circuits in the freezing regime are decommissioned, and the vertical freeze pipes are cut out using thermic lances as they are encountered.
Jacking forces to push the tunnel boxes forward are applied only at the invert, with a main jacking station at the rear in the jacking pit, and an intermediate jacking station between the elements in each two-or three-element string. The jacks used on the job have the same capacity of 485 tonne each at 42 MPa working pressure and 808 tonne each at the maximum allowable 70 MPa pressure. The thrust cylinders are 520mm diameter and have a maximum stroke of 1,067mm for the 25 jacks in the rear jacking station, and 419mm for those in the 28 to 32 jacks in the intermediate jacking stations. All the jacks are grouped together in clusters of two or four with manifolds and one set of hydraulic hoses from the hydraulic pump set to each cluster. Jacking forces of up to 15,000 tonne total are transmitted from the base slab into the post-tensioned slurry walls of each pit.
Each jacking station is activated in sequence, caterpillar style, pushing each element forward at a barely perceivable rate of about 30min/900mm stroke. After each 900mm stroke, spacers are inserted at the rear jacking station and are replaced with 3,600mm permanent spaces as the jacking progresses. The maximum 12,000 tonne pushing force available in the 25 jacks in the rear jacking station is sufficient to push the two or three-element strings for about the first 50m without engaging the intermediate jacks. This helps maintain alignment, as do the guidewalls. The elements are cast within 1.2m high guidewalls cast into the jacking pits with foam filling of only about 25mm in the very tight tolerance gap between the wall and the moving box. Friction on the side-walls during the jacking process can be reduced using a polymer gel lubricant. Heating elements cast into the concrete walls prevent the gel freezing when in contact with the frozen ground. Jacking is a delicate balance of keeping the ground frozen and the lubricant liquid.
Having started in mid-October 1999, the ramp tunnel jacking was completed in December 1999. At the end of the process, survey measurements indicated that the leading edge of the two elements of the tunnel came to rest within 80mm of their nominal vertical alignment, and within 60mm from their nominal horizontal location, well within the ±150mm tolerance specified. Tests of the anti-drag system cables in the roof revealed that at the end of the 41m push, maximum load on a single cable was 1 tonne, or 60% of the design value.
It was agreed during the TunnelTalk site visit that the tunnel jacking alternative is an expensive option but that in urban areas; under very shallow cover; beneath vital infrastructure which must be kept in operation throughout the construction period; and through soft, non-homogeneous ground conditions where dewatering is unacceptable and settlement a major concern; the method, in association with ground freezing for soil stabilisation, was the best method. After the Boston success, and supported by successes elsewhere in the world, several similar projects were said to be on the drawing board for other rail and road underpasses in the United States. At a presentation about the technique to the British Tunnelling Society in January 2000, Doug Allenby of Edmund Nuttall and John Ropkins suggested that the future could see jacks of longer boxes with larger cross section areas; jacks for single radius curved tunnels; and jacks through a greater variety of ground conditions including rock. The future looks bright for tunnel jacking.
The June 2016 meeting of the British Tunnelling Society by Andrew Robinson, Christopher Howe and James Thomson of Jacked Structure Ltd provided an update of recent developments in the jacked-concept of urban shallow tunnels with potential for using jacked concepts for creating large underground structures and caverns in soft soils and possible application of the jacked concept on planned projects in the UK including the High-Speed 2 railway project and the proposed Hammersmith Flyunder highway project in West London.