Assessing ground interaction effects and potential damage on existing tunnels before and after new excavation works

Lead Research Organisation: Imperial College London
Department Name: Civil & Environmental Engineering

Abstract

Tunnels and underground railway systems are a popular solution to the increasing need to provide fast efficient transport systems in urban environments. Because in many cases the ground contains dense networks of existing tunnels for transportation and services and also deep foundations, new tunnels have to be constructed deeper to pass beneath these obstacles. New tunnel construction results in stress changes and consequent deformations in the ground and these are transmitted to and affect other existing structures. Services and particularly operational metro tunnels are very sensitive to such movements. Railways operate within narrow 'kinematic envelopes' and if a tunnel deforms so as to restrict this envelope the running of trains is severely restricted, requiring reduced speed limits or shutdown of the network and, in the extreme case, their stability might be compromised. At any of these levels, efficient running of the train network is severely hindered. If it is necessary to halt operation of a train line, even for a short period, there are astronomical daily costs from loss of ticket revenue and from loss of time to the commuting workforce. The societal impact is also enormous with inconvenience to thousands of people, frustration and unrest.Many existing tunnel networks were constructed over a century ago when traditionally linings were made mostly from grey cast iron segments. Little is known about the detailed response of segmental tunnel linings made from this material, e.g. how much deformation they can sustain and when and how they will fail. Grey cast iron is a very brittle material and its response depends on a number of factors such as the force in the bolts, nature of the gaskets/grommets used, and whether the 'pan' is infilled with concrete.Royal Assent for Crossrail, a new transport link across London, was granted this year. Construction is due to start in 2010. The tunnels for Crossrail are larger and deeper than most existing tunnels and they will pass beneath more than forty such tunnels. Many of these are aged and constructed from cast iron segments and transport thousands of commuters hourly.This research project sets out to gain a full understanding of the limits of deformation of existing cast iron tunnels when deformed by underground excavation such as tunnelling. Because of the nature of these linings (connections etc) it is necessary to consider their response at an appropriate scale. A three-path approach is to be adopted with laboratory testing of large-scale rings formed from cast iron segments, field monitoring of existing tunnels during Crossrail tunnel construction and numerical analysis of both laboratory and field conditions. The laboratory study will enable limits of deformation to be established, e.g. yield and failure, under different conditions such as bolt forces, gaskets types. The rings will be instrumented to monitor changes in stresses, bolt force etc as the ring is deformed to simulate typical shapes observed and predicted in situ. This is relevant to understanding conditions both before and after new tunnel construction. Numerical analysis will be used to model the laboratory set-up and refined prior to making predictions of a field situation at Hyde Park where the Crossrail tunnels will pass beneath the Central line tunnels. Comprehensive instrumentation will be installed in the ground and existing tunnels at this location to monitor in detail and understand their response during new tunnelling. The numerical analysis can then be further refined to achieve more accurate predictions for future projects.Bringing together these three research components will allow guidelines to be drawn up on expected and allowable tunnel lining deformations and the effectiveness of mitigation measures for future projects involving tunnelling beneath existing tunnels, such that safe and efficient operation of underground railway networks is not compromised.

Publications

10 25 50
 
Description Numerous findings have been made during the course of the research period, especially in the three main areas of the five themes investigated (see (i) to( v) below). The following passages detail many of these. Sixteen journal publications have been produced so far as listed in the publications section and many of these provide detailed information (as referred to below where appropriate).
The five main aspects to the project are as follows.
(i) Field instrumentation installation and monitoring within and around the existing Central line tunnels and in the ground in Hyde Park to observe responses when the new Crossrail tunnels were constructed.
(ii) Structural testing of specially manufactured half-scale lining segments.
(iii) Numerical analysis of the field conditions and validation with the results from field monitoring.
(iv) Numerical analysis of the cast iron segments and comparison with results from laboratory testing.
(v) Advanced laboratory testing of high quality samples taken during installation of the field instrumentation (ultimately to data from these tests to be fed back into the numerical analysis).

1) Field instrumentation and monitoring
a) Instrumentation and monitoring at Tottenham Court Road
One of the station tunnels at Tottenham Court Road was to be upgraded as part of the Crossrail works and so some of the instruments and methods to be used in the Central Line tunnel monitoring were tested at this time. Some of the segments were to be removed and so these were instrumented in various ways. It was established that the instruments worked well and there was good correlation between the different measuring systems. The magnitude of change in strain measured using electrical and mechanical strain gauges was found to be similar to that predicted assuming full overburden unloading with a segment modulus of 100 GPa. The work is described in detail by Yu et al. (2014) and in Dr Jessica Yu's PhD thesis (2014).
b) Instrumentation and monitoring within the existing Central line tunnels when new Crossrail tunnels are constructed beneath them Hyde Park
A length of the eastbound Central Line tunnel running between Lancaster Gate and Marble Arch stations was monitored when the Crossrail tunnels were constructed beneath it. The new tunnels had a clearance of 4.2m and were at about 40 degrees skew. Various instruments and surveying techniques were implemented: electrical strain gauges, displacements gauges, eye bolts for tape extensometer measurements, state-of-the-art optical fibre sensors (installed and read by a research team from ETH Zurich, working in collaboration with Imperial College) and a string of electrolevel beams along the tunnel length (results kindly provided by BAM Nuttall, Ferrovial Agroman and Kier Construction Joint Venture (BFK JV).
Main developments and findings
- The deformed shape of the tunnel lining from the tape extensometer measurements corroborated well with the mode of bending in the lining segments captured by the strain gauge measurements.
- It was concluded that the effort involved in installing and reading the strain gauges was not justified given the level of information provided by them. The tape extensometer readings were comparatively quick and the eye-bolts cheap and easy to install and the readings gave a better overall idea of tunnel diameter/chord changes. This is considered to be an important finding in terms of time and economy.
- The magnitude of diametric changes as measured with the tape extensometer were less than ±0.1%, much smaller than the out-of-circularity generally measured in existing tunnels (~0.6%).
- Additionally negligible movements across longitudinal joints between rings were measured with the displacement transducers installed across them. This was corroborated by the measurements with the optical fibre sensors.
- The optical fibre sensors were able to achieve a spatial resolution of about 1 cm with distributed nature of the measurements using swept wavelength interferometry. This is considered a new development in the use of this style of monitoring within tunnels. The fibres were attached using a combined glue/magnet attachment system specifically developed for this situation.
- With the small spatial resolution achieved using the optical fibre sensors it was possible to monitor strains across joints. Results suggest that longitudinal deformation along the Central Line tunnel crown occurred within the segments rather than across the circumferential joints. This is important for understanding the overall response of the existing tunnel. The overall strain pattern observed is of comparable form to that observed in the ground.
- Optical fibre sensors installed circumferentially within the tunnel were able to detect movement across the longitudinal joints as the Central Line tunnel deformed.
- Settlement profiles interpreted from the string of electrolevel beams were of similar form to those measured in the ground in Hyde Park. Comparing the magnitude and form of the two sets of measurements should allow relative bending stiffness values of the ground and the tunnel to be estimated.
The results from this aspect of the work are being written up currently.
c) Instrumentation and monitoring within Hyde Park
Extensive surface and subsurface instrumentation was installed above and around the existing Central Line tunnels and the new Crossrail tunnels. The scope of what could be installed close to the Central Line tunnels was limited as the tunnels run beneath the busy Bayswater Road. Three arrays of surface monitoring points were installed: one along Bayswater Road and two within Hyde Park. Additionally instruments were installed in a total of 38 boreholes to depths ranging from 20 m to 50 m to measure subsurface vertical and horizontal displacements and strains, pore water pressures and total stress changes. Some of the subsurface instrumentation was installed close to the Central Line tunnels while most of them formed a main array in Hyde Park.
Main developments and findings (Wan and Standing, 2014 a and b; Wan et al., 2017a and b and Wan et al., 2019)
- Numerous lessons were learnt during the extensive subsurface instrumentation installation period. The main points have been written up by Wan and Standing (2014). It is anticipated that the detailed consideration of the mix proportions to be adopted for the different grouts used to install instruments will be of considerable interest to the geotechnical community involved with ground monitoring (different mixes were used for backfilling extensometer, inclinometer and piezometer borehole installations).
- The relatively new technique of installing multiple piezometers in a grouted borehole was adopted as part of the monitoring scheme. The piezometers were generally found to work well and independently. Details are written up by Wan and Standing (2014). This paper also gives details about the number of London Clay claystones encountered during the drilling of all the boreholes. Such detailed information is not often reported - especially for a case where the ground conditions have been logged so carefully.
- Pore pressures in the near vicinity of the existing Central Line tunnels were found to be slightly lower than hydrostatic, indicating that some drainage takes place into the existing bolted grey cast iron segmentally lined tunnels.
- The field monitoring allowed differences in the ground response to tunnelling to be observed in a greenfield condition compared with the ground close to and affected by construction of the existing tunnels and the subsequent consolidation that took place. Generally wider settlement troughs were observed where the existing tunnels were located. However, in observing the ground response to the second Crossrail tunnel, it was seen the recently constructed first Crossrail tunnel had a much greater influence on the ground response than the Central Line tunnels (Wan et al., 2017a).
- Observed ground responses were linked with the various earth pressure balance tunnelling machine 'parameters'. Subsurface measurements indicated that the ground responded as a contracting displacement field (Wan et al., 2017b). This was closely linked to parameters such as the face pressure and also is strongly controlled by the depth of the tunnel and overburden. This is considered an important finding as it means that conventional empirical prediction methods can be applied in such cases. An earlier publication concerning the same type of tunnelling in London Clay as part of the Channel Tunnel Rail Link project showed an expanding displacement field which would be much more difficult to predict (Standing and Selemetas: http://dx.doi.org/10.1680/geot.12.P.154).
- Although mostly conventional instrumentation and precise surveying techniques were used, e.g. rod extensometers, in-place inclinometers (MEMS tilt sensors), and vibrating wire piezometers, ETH Zurich also installed optical fibre sensors in two rod extensometer boreholes and two shallow trenches (transverse and longitudinal to the westbound Crossrail tunnel). The borehole installations provide hitherto unseen detailed continuous strain profiles. When integrated these profiles matched the discrete displacements measured at the locations of the rod extensometer anchors. They also allowed cracking in the backfilled grout column to be observed.
- The quality of the data from the optical fibre sensors installed in the near surface trench far exceeded those from the manual micrometer stick measurements. Various type of optical fibre were tested within the installation (detailed in Dominik Hauswirth's ETH Zurich PhD thesis, 2014). The quality of the measurements was so good that it has been possible to propose a new method of predicting vertical as well as horizontal displacements in advance of the tunnelling machine reaching the point of measurement (written up by Hauswirth et al., 2014).
Details of the various instrumentation performances and the ground response are provided in Dr Michael Wan's PhD thesis (2014).
2) Structural testing of specially manufactured half-scale lining segments
Twelve grey cast iron (GCI) half-scale segments were specially cast using ingredients and mix proportions similar to those used a century ago as determined from London Underground archives and exactly replicating geometric details of the original segments. The scale was dictated by the limitation of casting the thinnest part of the segments (the web / skin). At the same time 'coupon' samples were cast for testing. Two suites of tests were undertaken within the structures laboratory at Imperial College. The first involved performing two-segment tests similar to those of Thomas in the 1970s on full-scale segments. The second set of tests used a full ring (comprising six segments bolted together) which was loaded to simulate ground stresses acting on the lining form the ground around them as in situ and then deforming the ring into elliptical shapes of similar form and magnitude observed in existing tunnels. In both cases the segments and joints between them were comprehensively instrumented and monitored to investigate factors such as how stresses developed in the segments and the influence of bolt forces and joint stiffness.
a) Coupon testing
The coupon samples cast at the same time as the segments were machined into 'dog-bone' shapes and tested in tension.
Main developments and findings (see Yu et al., 2017)
- Initially several trial tests were performed on dummy samples to optimise the way in which the tests were to be performed e.g. the best way of gripping the ends of the samples and the method of measuring strain/displacement.
- The yield stress and ultimate strength of a number of coupons were established and found to be comparable to those available in the literature. Of particular importance was the yield stress as in testing the segments the intention for most of the tests (especially the initial tests) was to keep the stress level within the GCI within the elastic range. This was to facilitate interpretation of the strain gauge data.
- Multiple tests involving load-unload loops confirmed that the stiffness of the GCI reduced as it was loaded to higher levels and then unloaded. Mapping this behaviour was essential in understanding the overall GCI material characteristics and also for the interpretation of strain gauges when the segments were loaded beyond the elastic range (and subsequently unloaded).
- In addition to the coupon tests, tensile tests were performed on wrought iron bolts recovered from the Waterloo and City Line. The results were compared to the known behaviour for mild steel to confirm that it was appropriate to use mild steel bolts in the laboratory investigations for this research.
b) Two-segment tests (see Tsiampousi et al., 2017)
Prior to testing, the segments were comprehensively instrumented with strain gauges. Optimal positions were selected based on numerical analyses carried out to model the two-segment test set-up (covered under the fourth theme of the research in item 4). Additionally special clamps were designed and made to hold displacement transducers in the vicinity of the bolted joint and instrumented bolts were also specifically purpose made.
Main developments and findings
- An experimental rig was developed to test two segments bolted together in a similar fashion to the tests performed by Thomas in the 1970s (e.g. Thomas, H.S.H. 1977. Measuring structural performance of cast iron tunnel linings in laboratory. Ground Engng., 10, 5, 29-36). The bolted pair of segments, curving downwards, with one end.hinged and the other on a roller, was deformed in small steps using a line load applied parallel to and close to the joint (at the uppermost part of the pair). The level of control and instrumentation on the segments was much greater than those available in the original tests.
- The data from Thomas' tests were reviewed to provide an initial estimate of the joint bending moment capacity stiffness. A series of two-segment tests with varying bolt preloads was then conducted and the results interpreted under the assumption of elasticity. As the conditions of the two segment test were statically determinate, the interpretation of the instrumentation and estimation of the bending moment could be compared against analytical equations. There was good agreement between Thomas' tests with the full-scale segments and those in this research at half-scale.
- The two-segment tests proved valuable in that it gave confidence that the methods used to interpret the instrumentation, especially the strain gauge readings, were appropriate. The same methods of interpretation were used in the full-ring tests.
- A number of other observations were correlated with the results from the numerical analysis described in item (4). For instance, it was found that the load in the middle bolt (the segments were connected together by three bolts) changed negligibly as the load was applied, compared with the two outer bolts.
- The stiffness of the bolted connection was found to reduce as the joint started to open.
c) Full-ring tests (see Yu et al., 2017 and Afshan et al., 2017)
The purpose of performing these tests was to deform the ring to similar shapes observed in situ both prior to and after nearby underground excavation and observe changes in quantities such as stresses within the segments, bolt forces and joint openings. In this way the effect of predicted ground movements from tunnelling could be assessed to gauge the level of distress they might inflict on the existing tunnel lining. In order to achieve this a highly sophistaicated novel testing rig was designed, manufactured and tested over several months. This was a major development (summary details given below). Two series of tests were then performed: a major set of tests with stresses kept within the eastic range and a final pair of tests where the ring was taken to failure.
Main developments and findings (see again Yu et al., 2017 and Afshan et al., 2017)
- The development of the test set-up was a staged process, especially in the development of the control system for applying loads and displacements to the ring. The ring was loaded by 18 actuators positioned at equal distances around a reaction ring that surrounded the test ring. There were therefore three actuators per segment. This allowed much greater ontrol of displacements and loads than could be achieved with soil. The ring was loaded while in a horizontal plane on a structural floor. Each actuator bore onto a spreader pad via a load cell to help distribute the load as a uniform radial stress. The ring was deformed using a combination of computer-activated load and displacement control. Initially a uniform radial stress was applied to the ring via all actuators and then some were held at constant load while others were displaced to produce essentially elliptical forms. Multiple tests were performed to assess the effects of differing initial radial stresses, representing four typical tunnel depths (to about 25 m maximum) and various bolt forces (5, 7.5 and 10 kN). The influence of installing tar-infused hemp grommets within the bolt holes was also investigated.
- The extensive instrumentation installed on the ring allowed both global and local joint responses of the bolted segmental cast iron lining to be studied experimentally and the internal actions relating to a particular deformed shape could be measured. The design and development of the bespoke loading facilities from scratch is a major contribution of this research into the behaviour of bolted segmental linings. This work is written up in detail in Dr Jessica Yu's PhD thesis (2014).
- It was found that at high hoop force stress levels, relevant to deeper tunnels, the bolted segmental ring behaved as a continuous ring while at low levels joint stiffness values started to reduce allowing greater articulation.
- For the small distortions imposed during the first main series of tests, the presence of compressive hoop force was found to render the magnitude of bolt preload insignificant in terms of influencing bolted segmental lining behaviour. This suggests that there may not be much benefit in either tightening or loosening bolts as mitigating measures to excavation-induced ground movements.
- Tests performed using grommets (inserted beneath bolt heads and nuts to make the holes more waterproof) showed that the bolt forces decreased after tightening them to a fixed load. It was necessary to retighten them several times before the forces held. This implies that bolts installed in situ with grommets are likely to be at lower forces than those without.
- At an overburden load corresponding to about 24m below ground level, the onset of joint opening was found to start once the inner radius had increased by about 0.20% by unloading at 90 degrees, contributing to the reduction in the overall stiffness of the ring. At this level of deformation, the extreme fibre stress in the ring was 22% of the ultimate tensile strength, i.e. within the elastic limit of grey cast iron, which in practice is usually taken as 25% of the ultimate tensile strength. At the same level of overburden, the joints started to open once the inner radius reduced by 0.58% from increasing the load at 90 degrees. In this case, the extreme fibre stress in the ring was 94% of the ultimate tensile strength, i.e. the ring segments would deform inelastically before the joint would play a part in the distortion of the ring. This implies that caution needs to be exercised when engineering works are carried out close to existing tunnels which would increase the loading on the tunnel lining, for example piling works.
- Tests performed keeping within the elastic range of the GCI, the experimental measurements indicate that using Morgan's equation (1961) to calculate the maximum bending moment for a tunnel ring assuming elliptical distortion based on a maximum magnitude of radial displacement underpredicts the maximum bending moment in the ring, even before the application of any reduction factor (as suggested by Muir Wood, 1975). The linear elastic solutions appear to be better suited to estimating the bending moment distribution in a tunnel ring from a measured maximum radial displacement. However, this needs to be confirmed for higher strain levels. The results from the two tests taken to higher strain levels (and ultimately failed are still being processed in detail). It should also be noted that in the experimental work (and the FE simulations) the tunnel was assumed to be perfectly circular and any deformations were the result of external loading.
- As noted above, results from the two tests taken to high strains only preliminary analyses have been performed so far.
3) Numerical analysis of the field conditions and validation with the results from field monitoring (see Avgerinos et al., 2016, 2017 and 2018)
There were three main sets of analyses performed during the course of the study. All analyses were performed using the bespoke Imperial College Finite Element Program (ICFEP). Initially a sophisticated two-surface kinematic hardening constitutive model was calibrated using advanced laboratory soil data from extensive testing of London clay samples. At the timeof this aspect of the study the tests from Hyde Park had not yet started so other well-documented data were used. This constitutive model and the parameters adopted were then used to model the well-documented St James's Park Jubilee Line Extension case study, accounting for different London Clay units. Excellent agreement with the field monitoring data was achieved for both short- and long-term responses.
The same model was used to simulate the new Crossrail tunnel construction at Hyde Park accounting for the previous stress history of the site. Again excellent agreement with the field data (Wan, 2014) was obtained even though the Crossrail tunnels are larger and constructed with a different tunnelling method (EPB vs. open-face shield at St James's Park), indicating the capability of the soil model to predict realistic soil movements from tunnelling in general.
The final stage of analysis was to model the situation at Hyde Park in 3-D. Because of the number of elements and the computing power required it was not practicable to use the sophisticated constitutive model used previously. Therefore a pre-yield non-linear elastic small strain stiffness model coupled with a Mohr-Coulomb plastic model was adopted. Other simplications made were to assume that the Crossrail tunnels were perpendicular to the Central Line (i.e. ignoring any skew) and to model both pairs of tunnels as single tunnels. Parametric analyses were run varying factors such as EPB TBM face pressure and the longitudinal stiffness of the existing tunnel. This exercise resulted in excellent qualitative agreement with the way in which the existing tunnels distorted as the new tunnels approached and passed them. Insight was also gained for the same stages into how bending moments and lining stresses developed. Another important finding from this work was the effect of shear stresses acting on the extrados of the exisiting tunnel(s) on the bending moment distribution within the lining.
Main developments and findings (again see Avgerinos et al., 2016, 2017 and 2018)
- Two very detailed case studies have been formulated (St James's Park and Hyde Park) using a sophisticated two-surface kinematic hardening constitutive model. These cases provide great insight into mechanisms of ground response to tunnelling and illustrate that the analyses can cater for different tunnelling methods.
- Excellent agreement was achieved with the analyses performed for the St James's Park case study for both the short- and long-term ground response to tunnelling. An important component of the analyses is that parameters from intact rather than reconstituted samples were used.
- Implementing realistic anisotropic permeability values for the different London clay units led to closer agreement with the monitoring data.
- The 2D FE simulations of the Central Line using initial earth pressure at rest coefficient of 1.3 to 1.4 showed that after allowing 100 years of consolidation, the maximum distortion in the existing tunnel was 0.2% squat. This implied that although the measured distortion is generally greater than 0.2% in the LU tunnels, it may be due to imperfect construction and self-weight distortion rather than ground loading. A field exercise conducted as part of the research at the LU Acton Depot certainly showed that tunnel rings constructed above ground could display greater than 1% of ovalisation from self-weight and initial out-of-built.
- The influence of the tunnel lining permeability is very great. It affects factors such as the shape of the tunnel after construction (squatting or egging) and also the load distribution with greater hoop stresses being developed in the long term for the impermeable case.
- It was observed that the tunnels deform as a consequence of both the normal and the shear stresses acting on the back of the lining. This was important with respect to the half-scale ring where only normal stresses were applied on the ring extrados.
- The excavation of the Crossrail tunnel below the Central line tunnel axis was found to impose changes on the internal forces in the Central line tunnel lining. The most crucial combination of the circumferential hoop force and bending moment distributions in the Central line tunnel lining occurred when the Crossrail tunnel excavation face is directly below the Central line tunnel axis.
- The face pressure applied to the ground affects the predicted soil movements. An increase of the face pressure (but not exceeding the overburden) decreases the ground surface settlements and the rate at which they develop.
- Using the 3-D numerical analysis it was possible to assess the influence of the longitudinal stiffness of the Central line tunnel. The agreement with the field measurements of the longitudinal strains along the Central line crown (using the fibre optic sensors) became worse as the longitudinal stiffness of the Central line lining was decreased in the numerical analysis. This indicates that the Central line tunnel lining has a significant longitudinal stiffness.
Comprehensive details of the numerical analyses performed are given in Vasilis Avgerinos' PhD thesis (2014).
4) Numerical analysis of the cast iron segments (see Tsiampousi et al., 2017)
A 3-D finite element model using ICFEP was developed by Dr Katerina Tsiampousi in conjunction with the two-segment test laboratory experiment. A linear elastic model was implemented for the iron, while joint opening was allowed in modelling the connections. Results from the test indicated regions where greater element refinement was necessary in the FE mesh, as well as the level of detail required, e.g. the inclusion of the caulking groove.
Main developments and findings
- The main series of tests showed that the stiffness of the bolting system assumed in the FE model was too high. The actual bolting system of bolts, nuts, washers, and the contact between them rendered the stiffness of the system several times lower than the stiffness of the mild steel bolt. Since the stiffness of the bolting system affected how the FE model predicted the deflection of the longitudinal flange, this was a valuable finding for future FE modelling of segmental ring behaviour.
- In general the FE predictions compared well with the laboratory measurements. The FE model proved helpful in providing a guide to the behaviour of the two segment arch beyond the loading conditions tested in the laboratory. An important prediction was that when the joint was subjected to negative bending in the FE model, all three bolts had similar increases in bolt load and the joint opened uniformly along the extrados, i.e. the displacement at the outer edge was predicted to be similar to the displacement at the middle of the segment. This led to a revision of the joint moment capacity estimation for the full-ring test where joints were subjected to both positive and negative bending.
- Additionally, the FE analyses allowed an interrogation of the principal tensile and compressive stresses in the longitudinal flange. Since the analyses were linear elastic the results could only provide a guide to the capacity of the joint. Nonetheless by limiting the tensile stresses in the longitudinal flange as predicted by the FE simulations to below the ultimate tensile capacity of cast iron, it was shown that the initial estimate of joint moment capacity was reasonable. Non-linear analysis would be necessary to determine the true capacity of the joint.
Publications presenting and discussing the results from the numerical analyses of the two-segment tests are currently being written up.
5) Advanced laboratory testing of samples taken during installation of the field instrumentation
High quality rotary samples were obtained from three of the boreholes made for installation of field instrumentation. These were positioned at increasing distances from the Central Line tunnels to investigate the effect of its construction on ground properties. 21 advanced triaxial tests were performed with state-of-the art instrumentation for measuring small strains (three axial and radial locations) and cross anisotropic moduli (using bender elements). Samples were consolidated anisotropically back to their in-situ stresses prior to undrained shearing in compression. Because of the stringent creep periods held at the end of each stage of the tests each tests generally takes 2-3 months. This aspect of the research project has been completed by Dr Ramtin Hosseini Kamal and Dr Khalid Al Haj. An extensive publication presenting and discussing the results from this phase of the research is currently underway.
Following from this work an academic scholar from the University of Sydney, Truong Le, worked within the Geotechnics Section at Imperial College for the first six months of 2016, looking at the effects of creep and strain rate on stiffness. This work is also being currently written up and will be closely linked with the research project.

In summary all of the work planned for this research project has been successfully completed and in fact a great deal more has been achieved (originally aspects 4 and 5 were not part of the proposal but during the course of the research it was decided to take the opportunity to perform them. The five individual but interlinked themes make the research project very comprehensive and robust, making it possible to gain a very wide appreciation of tunnelling in London and the effects of new tunnel construction on existing tunnels segmentally lined with grey cast iron. Writing up of the findings is still underway, although it is now largely complete.
Exploitation Route The findings contribute to the understanding of: (i) ground response to EPBM tunnelling (from the field monitoring) - useful to those involved in predicting such responses for future projects; (ii) the response of segmental grey cast iron to load and deformation changes, in both the elastic and plastic regions (this will contribute to current design methods); (iii) numerical modelling of ground and structural response to tunnelling (again this will be used by those involved in numerical analysis in general and also those making more complex predictions of ground and structural response to tunnelling); (iv) the behaviour of London Clay (from the advanced laboratory testing) and in particular reasons for the range stiffness values observed (there has been a great deal of uncertainty in respect of this with major implications on analysis and design using numerical methods). All of the themes can and will be used by those in academia performing numerical analyses requiring high quality field and laboratory data (for calibrating and validating the models used).
Recently a discussion was sent to Geotechnique about the paper by Avgerinos et al. (2018) by Professor Andrew Whittle and his co-workers at MIT. This is available ahead of print via: https://doi.org/10.1680/jgeot.18.D.008. It is also worth noting that for most of 2018 the same paper by Avgerinos et al. (2018) was recorded as being the most downloaded paper on the Geotechnique website (ICE Virtual Library)
Sectors Construction,Education,Environment,Transport

 
Description There are two main sources from which the findings are disseminated and used. One is through publications. So far sixteen journal papers have been published and a further one is about to be submitted (to Geotechnique). Three PhD theses have been written and defended by the three main researchers working on the project (Jessica Yu, Michael Wan and Vasilis Avgerinos). There are still a small number of papers are underway from these PhD theses, an MSc dissertation and also the aspect of the project covered by post-doctoral researchers on the advanced laboratory testing of the London Clay. There are many findings which we feel will be of great interest to the geotechnical and tunnelling communities. A summary paper with some of the main findings was published in the proceedings of the European Conference on Soil Mechanics and Geotechnical Engineering which was held in UK in Edinburgh, September 2015. This provided an excellent opportunity to highlight the research to the wider geotechnical community. Another summary paper, but more extensive than the Edinburgh paper, has been published by the IoM3 in Hong Kong as part of the conference on Underground Design and Construction. It was not possible to attend this conference as it clashed with the Edinburgh conference but the PI (J Standing) presented the paper at a specially organised lecture in November 2015 which was sold out with more than 200 people attending. As there are numerous major tunnelling projects underway in Hong Kong this was an excellent opportunity to broadcast our findings. The lecture was received with great interest. In 2017 the PI (J Standing) was invited to give a keynote address at the 3rd GEOTECH Hanoi conference which summarised lessons learnt from three EPSRC funded tunnelling related projects, including this most recent one. This was also received with great interest, especially as plans are afoot to construct a metro in Hanoi. Last year a 'Symposium-in-Print' (SiP) was held by Geotechnique. The theme of the SiP was 'Tunnelling in the urban environment' and in fact this was put forward to Geotechnique by the PI with a view to promoting the research done within this grant. Two papers were published in the special issue (Avgerinos et al. and Wan et al.) and both were presented at the SiP (by Prof Potts and Dr Standing respectively). The SiP, held in September (14.09.2017) was a great success and the presentations generated a great deal of discussion and the quality of the field monitoring was particularly commented on. The second main route for disseminating the findings from our research was by feeding them back into the Crossrail team (and members of the research steering group). The field monitoring in particular took place early in the project and so preliminary results were sent through to help with the assessment of both ground movements and also the effect of tunnelling beneath existing grey cast iron tunnels. The broad results from the research have also been incorporated into two books compiled by Crossrail in conjunction with many other parties involved in the Crossrail project. The books are intended to compile the 'lessons learnt' from many aspects of the works. In addition to the PhD theses written several MSc and final-year undergraduate dissertations have been written incorporating aspects of the EPSRC-funded research. Six undergraduate students have also undertaken UROP placements as part of the project. For several years Imperial College has hosted an afternoon seminar in advance of the Rankine Lecture which is hosted at Imperial College. The theme of the seminar in 2015 was 'Tunnelling in London: learning lessons from Crossrail'. Presentations were given in the first half of the afternoon by speakers from industry and in the second from researchers with two of the presentations given by members of the academic team (D Potts and J Standing) covering the five main themes of the project. Both the main lecture theatre and the over-flow room were full to capacity with people standing at the back in both, indicating an audience of more than 300. In March 2016 Emap ran a two-day conference on geotechnical monitoring and instrumentation and J Standing presented some of the results from the field monitoring - there wasbe very good attendance at this event. In April last year (25.04.2018) we held a half-day open seminar to present and discuss the main findings from the project to practising engineers and academics interested in tunnelling. This event was attended by 95 participants (it was free) and attracted much interest and discussion. J Standing gave a lecture at the Global Engineering Congress held at the ICE in October 2018 which was open to delegates and was also a joint British Geotechnical Association and British Tunnelling Society event. There were about 150 to 170 attendees and it was also broadcast live (and the recording should soon be available on the ICE website. Following this JS was invited to give a lunchtime talk at Matt MacDonald based on the same lecture. Dr Michael Wan was invited to give a presentation at the 5th International Course on geotechnical and structural monitoring held in Rome, May 2018. He presented some of the finding from our pore pressure measurements and discussed the use of fully grouted piezometer installations, a technique we used in our filed monitoring at Hyde Park. Work continues on the numerical analysis of the results from structural testing of the grey cast iron segmental ring tested by Dr Jessica Yu and Dr Sheida Afshan. This work is being carried out by Agustin Ruiz Lopez as part of his PhD (supervised by Dr K Tsiampousi and J Standing).
Sector Construction,Education,Environment,Transport
Impact Types Economic

 
Description Skempton Scholarship
Amount £100,000 (GBP)
Organisation Imperial College London 
Sector Academic/University
Country United Kingdom
Start 10/2017 
End 04/2021
 
Description Travel funding
Amount £600 (GBP)
Organisation Emap International 
Sector Private
Country United Kingdom
Start 09/2010 
End 09/2010
 
Description Field monitoring using fibre optic sensors with ETH Zurich 
Organisation ETH Zurich
Country Switzerland 
Sector Academic/University 
PI Contribution We collaborated on field monitoring both within the Central Line tunnels and within Hyde Park. We provided the situation for the optical fibre sensing system to be set up, e.g. in boreholes, and in the ground within Hyde Park, by incorporating their installation work with ours. We also provided measurements made by us using traditional survey techniques to validate their state-of-the-art measurements.
Collaborator Contribution They sent researchers to UK for several periods to oversee the installation works and to monitor during different phases of construction. They also provided the hardware (optical fibres, anchors etc) and also most importantly provided the readout units for taking the measurments (bringing them over every time) which are very specialised (and are worth in excess of £100k).
Impact We have written a paper (Hauswirth et al., 2014) describing the surface measurments made in Hyde Park. This has provided much greater resolution of measurement of horizontal ground strain than has ever been measured before and allowed predictions to be made of both horizontal and vertical displacements in advance of the tunnel approaching the line using a new methodology (described in the paper). We are currently working on writing papers referring to the information from the subsurface monitoring (again resultion of strains both the measurmeents themslelves and spatially better than ever achieved before by others) and the monitoring within the Central Line tunnel as the new tunnels were constructed beneath it. These latter measurements have provided great insight into the mechanism by which the tunnel deforms.
Start Year 2011
 
Description Tunnelling research with Crossrail 
Organisation Crossrail
Country United Kingdom 
Sector Private 
PI Contribution Providing state-of-the-art high quality research findings from the EPSRC-funded investigation into the effect of tunnelling on existing tunnels
Collaborator Contribution Crossrail employed two of the researchers (Jessica Yu and Michael Wan) for the duration of the project and in fact have extended their contracts till the end of 2015. They also paid the College fees for them to register for PhD degrees (~£140k). In addition to this Crossrail also made a major contribution to the cost of the installation of field instrumentation at Hyde Park (est'd ~£600k). They have contributed a number of other in-kind items, especially relating to staff time and providing information on the tunnelling construction works. Morgan Sindall funded the casting of the twelve grey cast iron segments and provided the steel reaction ring for the half-scale full ring tests. They also contributed in-kind with staff time. London Underground Limited provided archive material to enable accurate replicas of existing one-hundred year old segments to be manufactured and other information that helped with the progress of the project.
Impact The collaboration greatly contributed to many of the key findings listed separately.
Start Year 2010
 
Description Tunnelling research with Crossrail 
Organisation Morgan Sindall Group
Country United Kingdom 
Sector Private 
PI Contribution Providing state-of-the-art high quality research findings from the EPSRC-funded investigation into the effect of tunnelling on existing tunnels
Collaborator Contribution Crossrail employed two of the researchers (Jessica Yu and Michael Wan) for the duration of the project and in fact have extended their contracts till the end of 2015. They also paid the College fees for them to register for PhD degrees (~£140k). In addition to this Crossrail also made a major contribution to the cost of the installation of field instrumentation at Hyde Park (est'd ~£600k). They have contributed a number of other in-kind items, especially relating to staff time and providing information on the tunnelling construction works. Morgan Sindall funded the casting of the twelve grey cast iron segments and provided the steel reaction ring for the half-scale full ring tests. They also contributed in-kind with staff time. London Underground Limited provided archive material to enable accurate replicas of existing one-hundred year old segments to be manufactured and other information that helped with the progress of the project.
Impact The collaboration greatly contributed to many of the key findings listed separately.
Start Year 2010
 
Description Tunnelling research with Crossrail 
Organisation Transport for London
Department London Underground
Country United Kingdom 
Sector Public 
PI Contribution Providing state-of-the-art high quality research findings from the EPSRC-funded investigation into the effect of tunnelling on existing tunnels
Collaborator Contribution Crossrail employed two of the researchers (Jessica Yu and Michael Wan) for the duration of the project and in fact have extended their contracts till the end of 2015. They also paid the College fees for them to register for PhD degrees (~£140k). In addition to this Crossrail also made a major contribution to the cost of the installation of field instrumentation at Hyde Park (est'd ~£600k). They have contributed a number of other in-kind items, especially relating to staff time and providing information on the tunnelling construction works. Morgan Sindall funded the casting of the twelve grey cast iron segments and provided the steel reaction ring for the half-scale full ring tests. They also contributed in-kind with staff time. London Underground Limited provided archive material to enable accurate replicas of existing one-hundred year old segments to be manufactured and other information that helped with the progress of the project.
Impact The collaboration greatly contributed to many of the key findings listed separately.
Start Year 2010
 
Description 5th International Course on Geotechnical and Structural Monitoring, Rome, Italy 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Dr Michael Wan gave an invited presentation at the international course on Geotechnical and Structural Monitoring held in Rome, Italy May 2018. His presentation concerned pore pressure measurements and the use of fully grouted piezometers as employed at Hyde Park for the field monitoring on the ground response to the tunnelling for Crossrail.
Year(s) Of Engagement Activity 2018
 
Description Evening lecture (22.10.2018) at the Institution of Civil Engineers as part of the Global Engineering Congress held at the ICE but also avertised by and attended by members of the British Geotechnical Association (BGA) and the British Tunnelling Society (BTS). 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Jamie Standing was invited by the Chairman of the BGA to give a special lecture at the Global Engineering Congress. It was attended by members of the congress and also members of the BGA and BTS (the event was chaired jointly by the Chairmen of the BGA and BTS. There was good discussion afterwards and a number of emails came through subsequently.
Year(s) Of Engagement Activity 2011,2018
 
Description Half-day symposium (Dept Civil and Environmental Engineering, Imperial College London) 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact We held a half-day symposium at Imperial College in April (25.04.2018) to disseminate the primary findings from our research. A series of presentations by those researchers involved with the overall study was given. The symposium was aimed at those involved in tunnelling works ranging from practising engineers to academics. It was advertised through the British Tunnelling Society and the British Geotechnical Association. It was a very successful event with 95 participants and active discussion following the presentations and at the end of the afternoon.
Year(s) Of Engagement Activity 2018
 
Description Lunchtime presentation based on the GEC lecture to engineers at Mott MacDonald 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Professional Practitioners
Results and Impact Following Jamie Standing's lecture at the Global Engineering Congress at the ICE in October he was invited to give a condensed version of the lecture to the geotechnical and tunnelling engineers working at the consulting company Mott MacDonald at their main office in Croydon. London. There were 40 to 50 people in attendance and lively discussion followed. The occasion also allowed for discussion with senior management afterwards about potential research collaboration.
Year(s) Of Engagement Activity 2018