This article provides guidelines and considerations for designing remedial measures for recovering collapsed large-diameter tunnels, in soft ground, mined using sequential excavation and support (SES) methods. The guidelines were developed by the authors as a result of a number of observed tunnel collapses that occurred during excavation where partial or complete failure of the outer lining occurred. Owing to the studied tunnel’s sizes of up to approximately 150m² and the strength of the ground (soil and weak sedimentary rock), excavation was typically carried out in three stages – top heading, bench, and invert.
Each collapsed tunnel was typically being constructed through a mixture of poorly cemented dense or hard soils and weak Tertiary sedimentary rocks, locally jointed, folded and faulted, and contain moderate or high water tables, excavated using hydraulic excavators and breakers.
During or immediately after each excavation stage, initial support, herein termed the “outer lining”, had been installed to support and stabilise the excavated cross-section prior to installation of a cast-in-place reinforced concrete inner lining. The outer lining typically consists of lattice girders, shotcrete reinforced with welded wire fabric, and rock dowels. Depending on the ground quality, pre-support of the tunnel crown zone ahead of the excavation face had been required and comprised of forepoling with various combinations and lengths of steel bars, pipes, sheets, pipe roofing and grouting. Typically the tunnel drive dewatering was carried out using face drainage probes, vacuum well points from within the drive and deep wells drilled from the surface.
Design management
The principle goal of collapse recovery is to re-excavate the drive through the failed ground without undue delay or risk of subsequent collapses. Remedial measures are typically designed to allow use of the same excavation method (SES), the same equipment and the same personnel to mitigate delays to the programme.
Generally the objective is to stabilise the ground around the collapse. For smaller collapses in granular material this may be achieved by grouting with OPC, polyurethane or silica grout from both the face and above the collapsed zone, supplemented with the installation of limited pre-support. Larger collapses may warrant complete isolation of the drive before re-excavation. This is to ensure that new stress fields caused by the collapse cannot breach other areas of the tunnel lining. This latter design typically results in a canopy over the opening to produce a reinforced arch supported on competent material, which would subsequently support the majority of the loads of the failed ground during the re-excavation.
The supporting arches utilised on the recovery designs may typically consist of either: i) fully grouted double or triple pipe roof canopy depending on the overburden, extending to the bench or ii) jet-grouted column canopy extending over the arch and supported on a jet-grouted footings (Figure 1).
The design and implementation typically involves the following steps:
Site investigation principles
A geotechnical model of the collapse area is required to develop a numerical analysis model for designing the remedial measures. The primary objectives of the detailed investigations are: i) to determine the collapse geometry; ii) to determine the ground’s geotechnical properties in and around the collapse; iii) to determine the location and extent of obstructions resulting from the collapsed lining that may impede the recovery effort; and iv) to detect any voids caused by the collapse.
The extent of a collapse in a tunnel can usually be observed or reasonably assumed. However, the extent of the collapse outside/around the tunnel is often unknown, and can only be speculated about based on the volume of collapsed material and known geological features. Furthermore, the collapse zone may be expanding or migrating toward the surface. Disturbance outside the collapsed zone requires assessment to allow the selection of an appropriate recovery design.
Observation and survey should be carried out immediately, so that no useful data is missed. Inspect the surface to see if the collapse has holed through and survey sink holes and tension cracks and monitor them for enlargement, or the appearance of new ones. There may be other signs of disturbance, such as the disappearance of surface streams, or the severance of services in built up areas. If the collapse has not holed through, it may have created a surface subsidence trough or one may still be developing. Look for and monitor new cracks in structures and roads. Be alert for other telltale signs such as inclined service poles or trees and misaligned fences. If a detailed surface topographic survey exists before, resurvey to determine the extent of the trough and its subsequent development. If a detailed surface survey is not available, establish a grid of surface settlement survey points and monitor them regularly.
Deformation monitoring should continue to be carried out within the tunnel, if this can be done safely. This can provide valuable data about changes in the stress field around the tunnel, loads on the tunnel support system and warning of any further collapse. Once the initial observations and surveys have been carried out and a programme of ongoing observation and deformation monitoring (surface and in tunnel) has been established, attention can be given to planning the rest of the site investigation programme.
The designer of the remedial works needs to know the extent and in-situ properties of the collapsed material and any zone of disturbance around it. This will allow him to decide what ground improvement techniques are necessary, if any, and what method he can employ to re-excavate through the collapse zone.
The geotechnical properties that the designer is most interested in are the in-situ strength, density and stiffness properties of the different materials. He also needs to know if there are voids that need to be filled, or allowed for in the lining design. Such voids may require grouting to preserve the integrity of the ground around the tunnel and the inner lining, in the long-term.
The original groundwater regime may have been changed by the collapse, so groundwater information is required to design both the dewatering provisions for re-excavation and to determine water pressures for design of the inner lining. Laboratory testing of intact samples may also be required.
Design principles
Initial undisturbed ground has some cohesive strength that can be utilised in the design calculations for the outer lining. This cohesion also provides temporary crown support, following excavation of the round length, while the outer lining is being constructed. Minimal, if any, pre-support is generally required. This is a basic principle of SES tunnelling, whereby the ground strength is retained and utilised to obtain economic outer lining and final lining designs.
However, there is effectively no cohesion in collapsed ground, with the stiffness (modulus) also greatly reduced. Thus, two design components need to be considered for re-mining collapsed ground using the SES method. First, the ground ahead of the planned re-excavation must be strengthened and stiffened so that the deformations ahead and above the top heading excavation face are kept within tolerable limits as the face advances. Systematic measures termed “pre-support” may be applied from within the tunnel as follows:
The second major component of SES recovery design through collapsed ground is to design the outer lining so that the strengthening benefit achieved with the ground treatment and pre-support can be utilised to provide a new outer lining that is still economical but can take the increased ground loading applied by the collapsed ground. Generally, the top heading requires early ring closure using a temporary invert. Typical ring closure distances behind the top heading face of about three to six round lengths are required.
Outer lining design
Numerical models are typically used to analyse the ground-outer lining mechanical interaction and determine the axial forces, bending moments and shear forces induced in the new outer lining and, if used, in rock dowels. Most tunnel designs are generally analysed and designed using a finite difference numerical analytical model. The numerical models provide guidance for selection of deformation control values to be used during construction.
The treated and pre-supported ground around the tunnel excavation is typically analysed in the numerical models as “reinforced ground” with improved mechanical properties. The numerical models are developed on a trial-and-error basis until an initial support/outer lining design is obtained that generally complies with standard reinforced concrete design based on the selected design codes and reviewed in terms of overall safety and performance.
Inner lining design
Typically, in the cases studied, the inner lining structural analysis indicated that the lining had to be designed for higher vertical and horizontal ground loads compared with the loads that would be applied in un-collapsed ground. Generally in un-collapsed ground the vertical load, expressed as a height of ground above the crown, varied from 0.5 to 1 x the tunnel diameter. In collapsed ground however, the height increased to 1.5 to 2 x the diameter. This results in linings thicker in both the arch and invert, typically by 15-50%, depending mainly on the overburden, friction angle and the modulus of elasticity of the collapsed and treated ground.
Tunnel monitoring
Re-mining large-diameter tunnels using SES methods through collapsed ground involves a considerable risk of re-collapse. A high-quality monitoring programme is required in the top heading stage to identify and control risk. Typically, monitoring stations should be installed at a longitudinal spacing of not more than 8m-10m for the measurement of deformation and stress of the outer lining. The measured values are compared to the “trigger” values predicted by the numerical models of the excavation-support sequencing. Typically three actions are carried out depending on the ratio of the measured values to the predicted values.
Depending on the ratio of the calculated deformations to those measured, certain pre-agreed actions are implemented when the ratio reaches certain values. For example, “alert Level 1” occurs if the ratio reaches 1:3, “alert Level 2” if 2:3, and “alert Level 3” if 1.0. Control values are determined for each excavation stage, thus allowing for deformations to occur for the heading, bench and invert. The typical relationship between the “Control Criteria” and “Safety Control System ” is provided in Table 1.
Long-term monitoring
Major uncertainty regarding the long-term behaviour of inner linings of large-diameter tunnels in collapsed ground can be quantified by measuring ground pressures acting on the linings, strain and movement of the lining at critical locations. Even with extensive pre-treatment and pre-support of the collapsed ground, there is considerable risk of ground loading increasing with time, as the collapsed ground continues to consolidate and creep. This risk can be monitored, by measuring ground pressures acting on the lining and stresses in the lining at well defined locations such as at the tunnel spring line.
Related Files
Fig 1, left, a jet grouted roof canopy supported on vertical jet grouted footings and, right, a jet grouted roof canopy on semi horizontal jet grouted footings
