This paper illustrates an innovative design of openings in existing cast iron tunnels and describes the challenges during their installation. These openings were formed within the existing station tunnels served by the Piccadilly and Northern Lines at King’s Cross St. Pancras Underground Station.
The “flat arch” theory has been used as the conceptual basis for this design. Flat arches rely on the transfer of vertical loads through a thrust force which leads to large compressive stresses but minimal tensile and shear forces. The immediate benefits of this solution are the optimization of the size and depth of the segments, with no complicated connections and allowing simplified and staged installation.
The load-bearing elements of the opening frames including lintels, sills and jambs have been sized to fit within the width of an existing ring. The choice of this method facilitated the gradual installation of the frames in order to ensure minimal impact on the stability and structural integrity of the station tunnel rings. A significant benefit was the avoidance of temporary internal props that would have severely affected the platform space for passenger usage and therefore led to platform closures.
The project
King’s Cross St. Pancras Underground Station is one of the busiest stations of the London Underground network and is served by 6 tube lines and acts as an interchange for St. Pancras and King’s Cross Mainline railway stations. Its redevelopment was planned in conjunction with the renovation of St Pancras Station for the Channel Tunnel Rail Link (CTRL) and comprises improvement to the existing facilities of the station, two new ticket halls, the refurbishment of the existing ticket hall and new tunnelled passageways to the deep tube lines to relieve congestion due to the projected increase in usage by 2012.
New routes to the Northern and Piccadilly lines are provided by a network of pedestrian tunnels connecting one of the new ticket halls to the existing station tunnels. Two pairs of openings within the existing cast iron rings were formed to provide access from the existing platforms to new concourse areas. The size of the existing station tunnels is 6464mm i.d. The required new structural opening is 2.6m wide and 2.8m high.
Traditionally, opening sets are built by fitting deep beams and jambs within the station tunnel rings or bolting the beams at the back of the rings after excavating and exposing the external tunnel face. These require significant temporary support including the use of vertical props within the station tunnels. This is because a large number of segments need to be dismantled or left laterally unsupported compromising the structural integrity of the cast iron rings and leading to significant deformations of the lining. The props would impinge on the available space within the platform hindering passenger flow and safety leading to a closure of the station for several weeks.
Due to the key importance of this underground station for the operation of the London Underground network, all works needed to be designed and staged to maintain the operation of the tube lines at all times. The only acceptable disruption to the operation of the line was the closure of the platform tunnels to the public during a limited number of ‘52 hour’ weekend possessions. While the work was carried out within the station tunnel, safe clearances had to be maintained to the track to allow London Underground to run through passenger and engineering trains.
Design
In order to safeguard the integrity of the existing lining and to maintain operational services a different approach was considered for the design of the openings in the existing station tunnels. With this approach, the opening frames were formed by progressively removing existing segments within the lining and replacing them with new segments intended to support the hoop load through the opening. These segments, bolted to all adjacent existing and newly replaced segments, formed the lintel, jambs and sill of the opening set.
Simple beam theory would have led to a considerable depth of section in order to provide sufficient structural capacity and allow for full bending moment connections capable of maintaining the structural continuity of the lintel and sill.
To minimize the depth of the section and to reduce the complexity of the segment connections, the lintel and sill beams were designed as flat arches. The inbuilt hoop load within the affected tunnel rings is redistributed to the segmented lintel and sill beams through the thrust force created by the arching action within their web plates. This eliminates tensile stresses as all the forces are resolved into compressive stresses. By using the arch theory, significant loads can be carried by relatively shallow elements since the compressive stresses hold the segments together in a state of equilibrium. Therefore, the internal forces transferred between the bolted segments are mainly compressive stresses and the bolted connections need to carry nominal temporary loads. The segments were designed in Spheroidal Graphite Iron (SGI) due to its very high compressive characteristic strength.
The new segments (figure 2), were installed systematically to ensure minimal impact on present load paths. Works were carried out on a maximum of two rings of the existing tunnel lining at any time and the new segments were grouted into place immediately after installation. Sacrificial segments were used to infill the space of the actual opening to provide continuity of the lining until the opening set is constructed and capable of carrying the loads.
A flat arch depends on the rigidity of the lateral elements, hereafter named abutments, which need to absorb the considerable lateral thrust force and maintain any resulting lateral movement to a minimum. Figure 3 indicates the forces that are created within the lintel/sill and the reaction required by the abutments. Significant longitudinal movements would lead to large vertical displacements of the lintel and sill beams and corresponding failure to comply with serviceability deflection requirements. As the arch finds a new equilibrium after these movements, structural failure of the SGI opening set is not expected. However, large vertical movements would compromise the structural integrity of the existing lining supported by the opening lintel and sill causing cracks in the cast iron segments.
The stiffness of the lateral support of the opening sets may be affected by several factors. Defects in the original construction (dated 1900), presence of soft packing such as deteriorated timber or voids and possible weak material properties of the cast iron may lead to unexpected movements and distortions of the adjacent cast iron segments. Significant movements may also lead to undesirable cracking.
This risk has been mitigated by injecting high-strength grout in the joints between the adjacent 5 existing cast iron tunnels. To reduce the stresses in the adjacent cast iron rings, the abutments were strengthened by replacing the segments next to the lintel/sill with stiffer and more ductile SGI segments. These additional segments allowed for the redistribution of the compressive stresses over a larger width of the first affected existing cast iron segments.
To ensure the lateral load capacity and stiffness of the supports were consistent with the design requirements, the Designer proposed a simulation of the arch thrust force transferred from the abutment to the lateral support provided by the cast iron segments. This force was induced by means of 4 hydraulic jacks. The jacks were placed in specifically designed recesses at the centre of the opening set lintel and sill, below the central key segment. The jacks were pressurised to impart a large compressive force in the segments. This force was transferred to the abutments where the design thrust force is expected to act when the opening sets are engaged.
The benefit of this simulation was to force the opening set elements to move and impart all the possible non-elastic deformations due to imprecision in assembling the new opening sets, soft packing/voids and existing defects in the build. As the load was increased to the expected serviceability value, a bespoke instrumentation system monitored the elastic behaviour of the opening set elements and the adjacent existing segments to validate the structural ability of the system.
Plastic deformations were compensated by the provision of steel shims installed in the gaps formed between the key and the adjacent segments. Elastic deformations were released but they were small and accounted for in the design.
Installation of new segments
The installation of the new segments was carried out during five ‘52 hour’ weekend possessions. To safely replace the old segments with the new SGI segments, the work was carried out within a hoarding protection to maintain safe working clearances to allow passenger and engineering trains through.
The preparatory works ahead of the segment replacement works included stripping of finishes, accurate survey of the existing lining, circle joint packing and ring roll, diversion of services around the opening areas above and below the platforms. To install the sill beams, the existing platform was modified in proximity of the openings. The concrete platform structure was replaced by a steelwork solution with prefabricated steel panels so that it was possible to temporarily remove the panels during the installation works to gain access to the invert of the tunnel. The panels were then restored to reinstate the platform for public use the next day.
Figure 4 shows the sequence of the replacement of the old cast iron segments with new SGI segments. During each weekend possession, 8 segments within a pair of rings per opening were replaced. The segments were replaced from the abutments towards the centre, fit tight to the far side in order to minimise any gap between segments. The last segments to be installed were the key segments. Any gap formed between the key and the adjacent segments was filled with shims. The rings 1 and 5 are replacements to the existing lining and provide a stiffer support at the abutments and facilitate the dispersion of load onto the adjacent cast iron segments. Some smaller segments were designed and installed between the lintel/sill and the existing segments within rings 1 to 9 to account for the different roll of the rings. The segments marked I1 are infill segments that temporarily reinstate the continuity of the rings.
The existing segments were removed by grinding two strips in the middle of the segment and pulling out the two remaining sides. The new segments were installed with winches. Once the segments were in place, the site engineer reviewed the orientation and the contact with the adjacent segments. Spirit level and strings were employed to ensure the segments were orientated in the design position in order to build the lintel segments on a straight line. Where required, shims or wedges ensured full contact between the flange plates of the segments in both circumferential and radial joints.
At the completion, the 3D movements of the existing segments around the opening frame were reviewed. Apart from a segment, which moved 7mm for what is believed to be a mechanical reason (load imparted from a winch connected to one of the bolt holes), all segment movements were maximum 3mm – within the allowable 7mm displacement.
Instrumentation and monitoring
The critical assumption made during the design is that, to act as a flat arch, the opening sets must be longitudinally restrained by the existing rings to the immediate right and left of the newly installed segments. Construction was staged to create an opportunity to validate this assumption through the analysis of two fundamental characteristics during the application of a force simulating the thrust action: Rigid body movement and internal force distribution. This was achieved through an instrumentation and measurement scheme designed to provide stresses and movements of the segments at critical locations, and changes in value over time.
The monitoring of the stresses, via strain readings, within the newly installed SGI segments and the existing cast iron rings was carried out by means of Vibration Wire Strain Gauges (VWSG) applied directly to the exposed surface of the tunnel segment skin plate (figure 5). Real-time data providing a full understanding of the compressive stresses within the lining were recorded at a frequency of 10 minutes. The monitoring of the movement of the segments was undertaken by three methods:
• The overall lateral movement (longitudinally) of the abutments was monitored through 3D prism targets mounted on the external rings and in the centre of the pair of openings. The targets were read every 10 minutes by a theodolite installed in the station tunnels
• To backup data from the theodolite, the movement between abutments was monitored through tape measurements using a laser meter. This was carried out at every step of the jacking procedure to record longitudinal movements
• The out of plane movement of the segment (radial, towards the centre of the tunnel) was monitored through some dial gauges fixed on a build bar. The device is a spring connected to a bar which records relative movements with a 0.1mm precision. The built bar, a few centimetres offset from the opening set components and independent from the opening set, provided a reference point to the location of the segments. Movements were recorded at every increment of jack pressure
The recorded movements were checked against pre-defined control trigger levels. The trigger levels were defined as a control measure to avoid damage of any affected element and provide the designer with hold points to suspend the procedure and review the collected data in the event of unforeseen behaviour of the segments. The control trigger levels for the longitudinal movement of the two abutments at the far sides of both opening sets based on the laser measurements were a Green Level of 5mm, an Amber Level of 15mm and a Red Level of 25mm, the design allowable movement of the opening set components.
The pressure in the hydraulic jacks was monitored through gauges included between the jacks and the pumps that inject the fluid that operates the jacks. This precaution was necessary to ensure there was no unexpected drop in pressure and therefore load in the hydraulic jacks due to unexpected movements of the set.
Validation – jacking procedure
Four hydraulic jacks per pair of openings were employed to validate the design. The jacks were some 450mm deep with a 360mm diameter piston. The design load of the jacks was 500 tons, the value of the thrust force expected in the deepest opening sets in service. Prior to jacking, several preliminary checks were carried out. The opening sets were inspected from the cross passage connected to the new lower concourse. The purpose was to avoid any gap between the lintel/sill and the newly built concrete structure forming the cross adit. Any gap was filled with high strength grout to provide a lateral restraint to outwards movements of the jacked segments.
The jack’s position and shimming were inspected prior to the start of jacking. To minimise any risk of inwards movement of the segments, packing between the jacks and the segments connected to the jacks included wedges to impart a slight eccentric load to force the segments towards the external cross passage structure.
Predefined bolts, the top and bottom of the sill, lintel and abutment pieces, were loosened. This avoided overstressing the bolted connection in the event of any significant movement of the opening set elements. The bolt holes were slotted to allow for a maximum lateral movement of 25mm. The key bolts were replaced with longer bolts and loosened at one side to allow for the longitudinal movement of the abutments and any gap opening between the key and the adjacent segments.
Jacking of the opening sets was programmed over 3 stages to monitor behaviour of the opening sets with different levels of stress and review internal stresses and movements over a period of time with the jacks in a position and time period to maintain stresses within the segments. At the first stage the jack pressure was increased to 1000kN. The difference in pressure between each jack was maintained within 100kN. The expectation was to impart sufficient load to correct any imprecision in the build of the new segment and start viewing any other plastic deformation. To monitor these effects, the load was applied at small increments of 50kN per jack, about 1% of the target force. Measurements of the rigid movements were carried out at each stage. Following completion of these increments, the jack were locked and left in position for a minimum of 3 days.
As the pressure was applied to the jacks, the affected lintel and sill beams were visually inspected. Minor movements, some 1mm, were noted at the first increments. The recorded inwards movements of the segments were very small. After expected movement due to the build, the segments were stable. Figure 6 shows the behaviour of one of the abutments of the opening sets. Three strain gauges records are attached. At stage one, the load imparted was 20% of the total. The stresses peaked at 25MPa, reducing to less than 20MPa when the jack was locked. To lock the jack, the pressure was released maintaining the piston in position. Although movements of the segments were still restrained, since the elastic strains were very small, the lack of pressure in the jacks led to a slight drop in strains that translated in a reduction in stress within the monitored segments.
At the second stage the jack pressure was increased to 3000kN. After every increment of 100kN of all jacks, measurements of the longitudinal and lateral movements were recorded. The jacks were then locked and left in position for a minimum of 3 days.
The force at this stage simulated about 60% of the thrust force produced by the hoop load at serviceability state. The effects of this force produced some recorded movement of the segments. This was sufficient to mobilise the abutments. The maximum movements were in the region of 2mm. The stresses recorded by the strain gauges increased linearly as expected to a value in the region of 65MPa to then slightly reduce after the jacks were locked. Similarly at the third stage, increments of 100kN were applied to the jacks until the load reached the back-calculated maximum force of 5000kN. The maximum lateral movements were small also at this stage, between 1 and 2mm. The stresses reached a maximum 102MPa before stabilising to between 70 and 80MPa. During the 5 days the jacks were left in position, the stresses in the new segments decayed only 5%. This proved the opening set was stable and capable of carrying a relevant percentage of the imparted load over a length of time with minor expected losses in strain.
The jack pressure was then reduced to 1000kN to reduce any elastic deformation in the opening set while maintaining any plastic deformation. Shims were applied in the gap between the key and adjacent segments. Now the infill segments could be removed and the opening set signed off.
Conclusions
The design and construction works proved it was possible to form an opening in a tunnel using the flat arch theory, without any delay to the service. Although the procedure had no precedent, every stage was completed as planned. The installation and jacking imparted movements in the region of 3mm in longitudinal and radial direction – within the control measure defined ahead of the works.
The service load expected by the set components was imparted without causing any major deformation or sign of distress in the existing lining. The recorded strains and stresses in the lintel segments confirmed the Designer’s expectations.
This solution can be very efficient especially in large tunnels (i.d.>5m). The benefits are a structural system within the thickness of the lining with resulting internal forces transferred in pure compression and not in bending and shear. The segmented lintel and sill also make possible a gradual installation with minimal impact and deformations to the tunnel.
Other openings in King’s Cross have been designed to be bolted onto the extrados as described. However these have been on smaller tunnels with lower hoop forces. The deformations produced by this method if proportioned to the tunnel size were much larger. The ‘bolted at the back’ solution on a station tunnel would have required larger beams and significant excavated volumes behind the lining to form sufficient space to install the lintel and sill beams. The extensive excavation and the time required to fix the lintel beam would have significantly increased the risks of undesirable deformations to the station tunnels.
The mitigations provided by the flat arch approach to the risks of damaging the lining of a cast iron tunnel were a critical benefit for a tunnel with historical deformations and reduced capacity to tolerate further deformations due to excavation works. T&TThis paper illustrates an innovative design of openings in existing cast iron tunnels and describes the challenges during their installation. These openings were formed within the existing station tunnels served by the Piccadilly and Northern Lines at King’s Cross St. Pancras Underground Station.
The “flat arch” theory has been used as the conceptual basis for this design. Flat arches rely on the transfer of vertical loads through a thrust force which leads to large compressive stresses but minimal tensile and shear forces. The immediate benefits of this solution are the optimization of the size and depth of the segments, with no complicated connections and allowing simplified and staged installation.
The load-bearing elements of the opening frames including lintels, sills and jambs have been sized to fit within the width of an existing ring. The choice of this method facilitated the gradual installation of the frames in order to ensure minimal impact on the stability and structural integrity of the station tunnel rings. A significant benefit was the avoidance of temporary internal props that would have severely affected the platform space for passenger usage and therefore led to platform closures.
The project
King’s Cross St. Pancras Underground Station is one of the busiest stations of the London Underground network and is served by 6 tube lines and acts as an interchange for St. Pancras and King’s Cross Mainline railway stations. Its redevelopment was planned in conjunction with the renovation of St Pancras Station for the Channel Tunnel Rail Link (CTRL) and comprises improvement to the existing facilities of the station, two new ticket halls, the refurbishment of the existing ticket hall and new tunnelled passageways to the deep tube lines to relieve congestion due to the projected increase in usage by 2012.
New routes to the Northern and Piccadilly lines are provided by a network of pedestrian tunnels connecting one of the new ticket halls to the existing station tunnels. Two pairs of openings within the existing cast iron rings were formed to provide access from the existing platforms to new concourse areas. The size of the existing station tunnels is 6464mm i.d. The required new structural opening is 2.6m wide and 2.8m high.
Traditionally, opening sets are built by fitting deep beams and jambs within the station tunnel rings or bolting the beams at the back of the rings after excavating and exposing the external tunnel face. These require significant temporary support including the use of vertical props within the station tunnels. This is because a large number of segments need to be dismantled or left laterally unsupported compromising the structural integrity of the cast iron rings and leading to significant deformations of the lining. The props would impinge on the available space within the platform hindering passenger flow and safety leading to a closure of the station for several weeks.
Due to the key importance of this underground station for the operation of the London Underground network, all works needed to be designed and staged to maintain the operation of the tube lines at all times. The only acceptable disruption to the operation of the line was the closure of the platform tunnels to the public during a limited number of ‘52 hour’ weekend possessions. While the work was carried out within the station tunnel, safe clearances had to be maintained to the track to allow London Underground to run through passenger and engineering trains.
Design
In order to safeguard the integrity of the existing lining and to maintain operational services a different approach was considered for the design of the openings in the existing station tunnels. With this approach, the opening frames were formed by progressively removing existing segments within the lining and replacing them with new segments intended to support the hoop load through the opening. These segments, bolted to all adjacent existing and newly replaced segments, formed the lintel, jambs and sill of the opening set.
Simple beam theory would have led to a considerable depth of section in order to provide sufficient structural capacity and allow for full bending moment connections capable of maintaining the structural continuity of the lintel and sill.
To minimize the depth of the section and to reduce the complexity of the segment connections, the lintel and sill beams were designed as flat arches. The inbuilt hoop load within the affected tunnel rings is redistributed to the segmented lintel and sill beams through the thrust force created by the arching action within their web plates. This eliminates tensile stresses as all the forces are resolved into compressive stresses. By using the arch theory, significant loads can be carried by relatively shallow elements since the compressive stresses hold the segments together in a state of equilibrium. Therefore, the internal forces transferred between the bolted segments are mainly compressive stresses and the bolted connections need to carry nominal temporary loads. The segments were designed in Spheroidal Graphite Iron (SGI) due to its very high compressive characteristic strength.
The new segments (figure 2), were installed systematically to ensure minimal impact on present load paths. Works were carried out on a maximum of two rings of the existing tunnel lining at any time and the new segments were grouted into place immediately after installation. Sacrificial segments were used to infill the space of the actual opening to provide continuity of the lining until the opening set is constructed and capable of carrying the loads.
A flat arch depends on the rigidity of the lateral elements, hereafter named abutments, which need to absorb the considerable lateral thrust force and maintain any resulting lateral movement to a minimum. Figure 3 indicates the forces that are created within the lintel/sill and the reaction required by the abutments. Significant longitudinal movements would lead to large vertical displacements of the lintel and sill beams and corresponding failure to comply with serviceability deflection requirements. As the arch finds a new equilibrium after these movements, structural failure of the SGI opening set is not expected. However, large vertical movements would compromise the structural integrity of the existing lining supported by the opening lintel and sill causing cracks in the cast iron segments.
The stiffness of the lateral support of the opening sets may be affected by several factors. Defects in the original construction (dated 1900), presence of soft packing such as deteriorated timber or voids and possible weak material properties of the cast iron may lead to unexpected movements and distortions of the adjacent cast iron segments. Significant movements may also lead to undesirable cracking.
This risk has been mitigated by injecting high-strength grout in the joints between the adjacent 5 existing cast iron tunnels. To reduce the stresses in the adjacent cast iron rings, the abutments were strengthened by replacing the segments next to the lintel/sill with stiffer and more ductile SGI segments. These additional segments allowed for the redistribution of the compressive stresses over a larger width of the first affected existing cast iron segments.
To ensure the lateral load capacity and stiffness of the supports were consistent with the design requirements, the Designer proposed a simulation of the arch thrust force transferred from the abutment to the lateral support provided by the cast iron segments. This force was induced by means of 4 hydraulic jacks. The jacks were placed in specifically designed recesses at the centre of the opening set lintel and sill, below the central key segment. The jacks were pressurised to impart a large compressive force in the segments. This force was transferred to the abutments where the design thrust force is expected to act when the opening sets are engaged.
The benefit of this simulation was to force the opening set elements to move and impart all the possible non-elastic deformations due to imprecision in assembling the new opening sets, soft packing/voids and existing defects in the build. As the load was increased to the expected serviceability value, a bespoke instrumentation system monitored the elastic behaviour of the opening set elements and the adjacent existing segments to validate the structural ability of the system.
Plastic deformations were compensated by the provision of steel shims installed in the gaps formed between the key and the adjacent segments. Elastic deformations were released but they were small and accounted for in the design.
Installation of new segments
The installation of the new segments was carried out during five ‘52 hour’ weekend possessions. To safely replace the old segments with the new SGI segments, the work was carried out within a hoarding protection to maintain safe working clearances to allow passenger and engineering trains through.
The preparatory works ahead of the segment replacement works included stripping of finishes, accurate survey of the existing lining, circle joint packing and ring roll, diversion of services around the opening areas above and below the platforms. To install the sill beams, the existing platform was modified in proximity of the openings. The concrete platform structure was replaced by a steelwork solution with prefabricated steel panels so that it was possible to temporarily remove the panels during the installation works to gain access to the invert of the tunnel. The panels were then restored to reinstate the platform for public use the next day.
Figure 4 shows the sequence of the replacement of the old cast iron segments with new SGI segments. During each weekend possession, 8 segments within a pair of rings per opening were replaced. The segments were replaced from the abutments towards the centre, fit tight to the far side in order to minimise any gap between segments. The last segments to be installed were the key segments. Any gap formed between the key and the adjacent segments was filled with shims. The rings 1 and 5 are replacements to the existing lining and provide a stiffer support at the abutments and facilitate the dispersion of load onto the adjacent cast iron segments. Some smaller segments were designed and installed between the lintel/sill and the existing segments within rings 1 to 9 to account for the different roll of the rings. The segments marked I1 are infill segments that temporarily reinstate the continuity of the rings.
The existing segments were removed by grinding two strips in the middle of the segment and pulling out the two remaining sides. The new segments were installed with winches. Once the segments were in place, the site engineer reviewed the orientation and the contact with the adjacent segments. Spirit level and strings were employed to ensure the segments were orientated in the design position in order to build the lintel segments on a straight line. Where required, shims or wedges ensured full contact between the flange plates of the segments in both circumferential and radial joints.
At the completion, the 3D movements of the existing segments around the opening frame were reviewed. Apart from a segment, which moved 7mm for what is believed to be a mechanical reason (load imparted from a winch connected to one of the bolt holes), all segment movements were maximum 3mm – within the allowable 7mm displacement.
Instrumentation and monitoring
The critical assumption made during the design is that, to act as a flat arch, the opening sets must be longitudinally restrained by the existing rings to the immediate right and left of the newly installed segments. Construction was staged to create an opportunity to validate this assumption through the analysis of two fundamental characteristics during the application of a force simulating the thrust action: Rigid body movement and internal force distribution. This was achieved through an instrumentation and measurement scheme designed to provide stresses and movements of the segments at critical locations, and changes in value over time.
The monitoring of the stresses, via strain readings, within the newly installed SGI segments and the existing cast iron rings was carried out by means of Vibration Wire Strain Gauges (VWSG) applied directly to the exposed surface of the tunnel segment skin plate (figure 5). Real-time data providing a full understanding of the compressive stresses within the lining were recorded at a frequency of 10 minutes. The monitoring of the movement of the segments was undertaken by three methods:
• The overall lateral movement (longitudinally) of the abutments was monitored through 3D prism targets mounted on the external rings and in the centre of the pair of openings. The targets were read every 10 minutes by a theodolite installed in the station tunnels
• To backup data from the theodolite, the movement between abutments was monitored through tape measurements using a laser meter. This was carried out at every step of the jacking procedure to record longitudinal movements
• The out of plane movement of the segment (radial, towards the centre of the tunnel) was monitored through some dial gauges fixed on a build bar. The device is a spring connected to a bar which records relative movements with a 0.1mm precision. The built bar, a few centimetres offset from the opening set components and independent from the opening set, provided a reference point to the location of the segments. Movements were recorded at every increment of jack pressure
The recorded movements were checked against pre-defined control trigger levels. The trigger levels were defined as a control measure to avoid damage of any affected element and provide the designer with hold points to suspend the procedure and review the collected data in the event of unforeseen behaviour of the segments. The control trigger levels for the longitudinal movement of the two abutments at the far sides of both opening sets based on the laser measurements were a Green Level of 5mm, an Amber Level of 15mm and a Red Level of 25mm, the design allowable movement of the opening set components.
The pressure in the hydraulic jacks was monitored through gauges included between the jacks and the pumps that inject the fluid that operates the jacks. This precaution was necessary to ensure there was no unexpected drop in pressure and therefore load in the hydraulic jacks due to unexpected movements of the set.
Validation – jacking procedure
Four hydraulic jacks per pair of openings were employed to validate the design. The jacks were some 450mm deep with a 360mm diameter piston. The design load of the jacks was 500 tons, the value of the thrust force expected in the deepest opening sets in service. Prior to jacking, several preliminary checks were carried out. The opening sets were inspected from the cross passage connected to the new lower concourse. The purpose was to avoid any gap between the lintel/sill and the newly built concrete structure forming the cross adit. Any gap was filled with high strength grout to provide a lateral restraint to outwards movements of the jacked segments.
The jack’s position and shimming were inspected prior to the start of jacking. To minimise any risk of inwards movement of the segments, packing between the jacks and the segments connected to the jacks included wedges to impart a slight eccentric load to force the segments towards the external cross passage structure.
Predefined bolts, the top and bottom of the sill, lintel and abutment pieces, were loosened. This avoided overstressing the bolted connection in the event of any significant movement of the opening set elements. The bolt holes were slotted to allow for a maximum lateral movement of 25mm. The key bolts were replaced with longer bolts and loosened at one side to allow for the longitudinal movement of the abutments and any gap opening between the key and the adjacent segments.
Jacking of the opening sets was programmed over 3 stages to monitor behaviour of the opening sets with different levels of stress and review internal stresses and movements over a period of time with the jacks in a position and time period to maintain stresses within the segments. At the first stage the jack pressure was increased to 1000kN. The difference in pressure between each jack was maintained within 100kN. The expectation was to impart sufficient load to correct any imprecision in the build of the new segment and start viewing any other plastic deformation. To monitor these effects, the load was applied at small increments of 50kN per jack, about 1% of the target force. Measurements of the rigid movements were carried out at each stage. Following completion of these increments, the jack were locked and left in position for a minimum of 3 days.
As the pressure was applied to the jacks, the affected lintel and sill beams were visually inspected. Minor movements, some 1mm, were noted at the first increments. The recorded inwards movements of the segments were very small. After expected movement due to the build, the segments were stable. Figure 6 shows the behaviour of one of the abutments of the opening sets. Three strain gauges records are attached. At stage one, the load imparted was 20% of the total. The stresses peaked at 25MPa, reducing to less than 20MPa when the jack was locked. To lock the jack, the pressure was released maintaining the piston in position. Although movements of the segments were still restrained, since the elastic strains were very small, the lack of pressure in the jacks led to a slight drop in strains that translated in a reduction in stress within the monitored segments.
At the second stage the jack pressure was increased to 3000kN. After every increment of 100kN of all jacks, measurements of the longitudinal and lateral movements were recorded. The jacks were then locked and left in position for a minimum of 3 days.
The force at this stage simulated about 60% of the thrust force produced by the hoop load at serviceability state. The effects of this force produced some recorded movement of the segments. This was sufficient to mobilise the abutments. The maximum movements were in the region of 2mm. The stresses recorded by the strain gauges increased linearly as expected to a value in the region of 65MPa to then slightly reduce after the jacks were locked. Similarly at the third stage, increments of 100kN were applied to the jacks until the load reached the back-calculated maximum force of 5000kN. The maximum lateral movements were small also at this stage, between 1 and 2mm. The stresses reached a maximum 102MPa before stabilising to between 70 and 80MPa. During the 5 days the jacks were left in position, the stresses in the new segments decayed only 5%. This proved the opening set was stable and capable of carrying a relevant percentage of the imparted load over a length of time with minor expected losses in strain.
The jack pressure was then reduced to 1000kN to reduce any elastic deformation in the opening set while maintaining any plastic deformation. Shims were applied in the gap between the key and adjacent segments. Now the infill segments could be removed and the opening set signed off.
Conclusions
The design and construction works proved it was possible to form an opening in a tunnel using the flat arch theory, without any delay to the service. Although the procedure had no precedent, every stage was completed as planned. The installation and jacking imparted movements in the region of 3mm in longitudinal and radial direction – within the control measure defined ahead of the works.
The service load expected by the set components was imparted without causing any major deformation or sign of distress in the existing lining. The recorded strains and stresses in the lintel segments confirmed the Designer’s expectations.
This solution can be very efficient especially in large tunnels (i.d.>5m). The benefits are a structural system within the thickness of the lining with resulting internal forces transferred in pure compression and not in bending and shear. The segmented lintel and sill also make possible a gradual installation with minimal impact and deformations to the tunnel.
Other openings in King’s Cross have been designed to be bolted onto the extrados as described. However these have been on smaller tunnels with lower hoop forces. The deformations produced by this method if proportioned to the tunnel size were much larger. The ‘bolted at the back’ solution on a station tunnel would have required larger beams and significant excavated volumes behind the lining to form sufficient space to install the lintel and sill beams. The extensive excavation and the time required to fix the lintel beam would have significantly increased the risks of undesirable deformations to the station tunnels.
The mitigations provided by the flat arch approach to the risks of damaging the lining of a cast iron tunnel were a critical benefit for a tunnel with historical deformations and reduced capacity to tolerate further deformations due to excavation works. T&T
Layout of the redeveloped King’s Cross St. Pancras underground station Fig 1 Cross section and
long section of a segmental
opening set
Fig 2 Arching action on lintel beam and abutment location Fig 3 Sequence and replaced segments Fig 4 Position of strain gauges around a pair of opening sets. The red marks (72No.) shown indicate strain gauges aligned horizontally. The blue marks (4No.) indicate strain gauges aligned vertically. The 4 hydraulic jacks are shown in brown within the central Fig 5 Movement of abutments (NLS-VEC-04B6B, 04C6B) and stresses calculated from the strain gauge readings in the vicinity of an abutment Fid 6 Acknowledgements
The Author wishes to thank London Underground Limited for their permission to publish this article. In addition he wishes to acknowledge the teamwork of Morgan Est/Beton und Monierbau Joint Venture, Metronet Rail SSL, London Underground, Soldata and his colleagues in Arup in delivering this innovative solution and the successful completion of these major tunnelling works at King’s Cross St Pancras
