The text ONLY of the Paper is given below; but it will probably be more useful to read the .pdf file that includes illustrations and can be found in the Library.  The text is given below simply so that Google and other "crawl engines" can find the text.

Lessons from a NATM (SCL) tunnel break-through in difficult geological conditions.

by

A. Paul Takacs Independent Consultant, Bucharest, Romania

Robin B. Clay Dar Al Handasah Consultants, Adana, Turkey

RÉSUMÉ:

La rencontre lors du perçage d'un tunnel, n'est pas seulement un événement très attendu qui marque la fin de la phase la plus critique, mais c'est aussi un défi technique pour les perceurs. Le procédé théorique n'est pas encore bien compris, mais l'expérience a démontré que des problèmes peuvent survenir lorsque les deux branches s'approchent l'une de l'autre, donc pour la rencontre finale des précautions additionnelles doivent être prises. Ce document présente un cas particulier de rencontre ou la géologie du sol était plus faible que prévue. L'ordre des événements et des actions prises y est décrit, et un model suggéré est ajouté pour discussion additionnelle. Même si un système de surveillance des déformations était sur place durant les travaux, ses bénéfices étaient limités à cause de l'imprécision et de la difficulté d'interprétation des résultats.

1. INTRODUCTION

The Project in question consists of five twin-tube three-lane motorway tunnels, each having a crosssection of more than 150 m2 . Generally located in fair to good geology, the tunnels are constructed by the NATM (or Sprayed Concrete Lining) method. As a general construction sequence, the cross-section was divided into top-heading and bench, and an invert was excavated as a third phase for sections where the geology was considered to be weak. A Rock Mass Classification system based on the Austrian Standard (ÖNORM 2203) was used to select the support for the normal driving conditions; however additional support was often requested when the behaviour of the rock support was not within the expected limits. During the construction of this project, nine break-throughs have taken place so far, in various geological conditions. This Paper focuses upon one particular case, where the geology was weaker than anticipated, causing delay to the programme, and also prompting the Contractor to apply a special excavation procedure used previously to overcome two collapses which had occurred in the parallel tube [Clay & Takacs, 1997]. The situation at the break-through position was also complicated by interruptions to the work occasioned by Public Holidays.

2. GEOLOGICAL CONDITIONS

The geological conditions at both faces can be characterised as thinly-bedded and heavily fractured sandstone with intercalations of shale and quartz along the main bedding planes. In both approaching drives, a nearly vertical fault zone some 500-750 mm wide had appeared in each left side-wall, with a strike of about 15° to the centreline of tunnel. These ABSTRACT: A tunnel break-through is not only a long-awaited event which brings to an end the most critical phase of tunnelling, but it is at the same time a technical challenge for the tunnellers. The theoretical process is still not clearly understood, but experience has shown that problems may occur as the advancing faces approach each other, so that for a safe break-through some additional precautions need to be taken. The paper presents one particular case of a break-through where the geology became more weak than expected. A sequence of events and the actions taken is described, and a suggested model is added for further discussion. Although a deformation monitoring system was in place during construction, its benefits were limited due to lack of accuracy and difficulty in interpreting the results. Fault zones were parallel and offset by about a metre; the adjacent rock along the zones was heavily fractured and sheared, with some clay infilling as shown in Figure 1. As the break-through point approached, the geological conditions at the Eastern face were slightly more favourable for the location of the break-through and so this drive was stopped, while the advance of the Western face continued towards this location.

3. TIME TABLE

On Monday, 3rd November, 1997, thirty-five days before Break-through Day, i.e., BD:-35, the two approaching faces were just over 80m apart, and progress was a steady one two-metre round per day in each heading, through what was classified as Rock Mass Class 4. At this point, some differences started to become apparent between the Western and Eastern faces. The Engineer suggested that a small pilot drift be driven through, for investigation and to improve the ventilation. This was not accepted by the Contractor, who had already decided to breakthrough on 18 December 1997 (BD: -4). On the evening of 11th December, (BD: -11), a face instability was recorded at the front of Round 171 of the Eastwards drive, 10m before the breakthrough point. Although the geology was notionally too good for that to occur, the fall occurred due to the presence of a transverse fault running across the tunnel that appeared at this time. The following day, 12th December, BD:-10, the proposed break-through procedure was submitted for the Engineer’s approval. For safety, the rock support class was changed from 4 to 5 for the section -6m to +5m relative to the break-through chainage, referred to the advancing West face drive towards the East. The advance from the East was to be stopped at +5m and sealed with shotcrete, and a support core was to be formed from tunnel spoil. The previous four rounds (+5m to +13m) were to be reinforced with an additional layer of wire mesh and shotcrete. For the break-through, a 6m umbrella of grouted forepoling comprising 2”Ø perforated pipe was to be installed. The advance of the West face continued, driving Eastwards. In the afternoon of the day after that, 13th December, BD:-9, in the advancing Western heading, cracks in the shotcrete lining were noticed, and some deformations were recorded by the monitoring system which put everyone on the alert. The results of the monitoring section at -12m, and the geological mapping of that section are presented in Figure 2. After an inspection of the area, it seemed that there was no imminent risk of collapse, but in the interests of safety and to safeguard further advance it was decided to install additional bolting and to strengthen the shotcrete arch, in order to give extra time for the shotcrete already applied to gain strength. On the 15th December, BD: -7, a meeting was convened by the Client. The Contractor maintained that he might yet achieve the break-through three days later on the 18th December, BD: -4, but the Engineer was of the opinion that due to the geological conditions just revealed, the breakthrough would be delayed. On the morning of BD: -6, the advance was restarted, but a new face instability prompted the Subcontractor to apply more face-sealing shotcrete which took several hours, so that when the Designer was informed in the afternoon, the work still had not been completed. The Subcontractor proposed to the Contractor adoption of sequential side-drift excavation, a method that had been applied previously for overcoming collapses. It should be mentioned here that the face instability was only within the excavation profile. The geology did not appear to be deteriorating compared to previous rounds, and the strengthened rounds showed no significant further deformation, i.e. conditions seemed to have stabilised.

Figure 1. Geological longitudinal profile. 

On BD: -4, the Contractor officially submitted his proposal for two side-drifts in each heading, and to re-open the East face. He proposed to drive the lefthand side-drift of the West face, and the right-hand side-drift of the East face, i.e. the opposing sidedrifts on the North side of the tunnel. This did not meet with general acceptance as being the best method for the break-through. The Contractor on site appeared to be under some pressure from his Head Office to advance quicker, and as the remaining pillar between faces was only 11.5m it was his intention to advance on both faces. However, such simultaneous excavation was not within the Specification, so the Contractor was given permission to carry out the side-drift advance, only on the condition that all additional costs were to be at his own expense, and all associated risks also being solely his. This the Contractor accepted, and work re-started. On BD: -2, the advancing side drift of the East face showed signs of instability. The rock pillar between faces was now a little less than 5m and with simultaneous excavation in both directions, it was believed that this might lead to a sudden unexpected break-through, creating a large span which, in the existing geology, might produce a large overbreak. It was therefore agreed that immediate support was to be installed in both headings, and that only the upper part of the remaining pillar was to be blasted first, with shotcrete application immediately after blast. BD. After the break-through blast there was no time for celebration because within about half-anhour the rock started to move, generating an overbreak of some eight cubic metres. The rapid application of shotcrete fortunately stabilised the situation. BD: +3 Religious holiday BD: +9 Official holiday BD: +21 The Excavation and support of the entire top heading was finished. The geological conditions encountered and the results of the monitoring from the break-through location are shown in Figure 3. BD: +35 Bench excavation approached the vicinity of the break-through from the East. After several days of observation and more frequent monitoring it was decided that the Eastern side between +5m and +61m be consolidated with additional bolting, because of larger than expected deformations revealed by the monitoring system. This decision was also influenced by the approaching work-stoppage for a five-day religious holiday. After the stoppage, the deformation curve was seen to have stabilised, despite the constant rate of advance of the bench, as is shown in Figure 4. This leads to a hypothesis that the deformation is caused by a release of energy from the break-through area, as discussed below.

Figure 2. Geological face mapping and monitoring results from chainage km: -12m.

4. DISCUSSION

4.1. Mechanism of the rock behaviour The following is based on the observations recorded during the present and previous break-throughs. The assumptions made herein have yet to be confirmed or refuted by analytical analysis. The statements which are made may help with model creation for such an analysis. The stress field in the undisturbed rock mass is three-dimensional, and the excavation process produces significant changes, which cease only when stability is achieved, i.e. when the deformations cease. The stress re-distribution starts ahead of the driving face and is completed far behind it. This stress re-arrangement is often called the “arching effect”. In general the transversal arching effect is well known and understood. However, with every advance, there is an secondary “longitudinal arching” effect which is, in normal cases, mostly observed only at the advancing face. This longitudinal arching effect becomes predominant when approaching drives close to within a certain distance, see Figure 5. One effect of the longitudinal arching is that the vertical stress increases, inducing deformation, and an effect of this is that the horizontal component of this deformation parallel to the tunnel axis tends to push the ground at the face into the opening created by the tunnel. If the geological properties of the rock mass are sufficiently robust, this effect will not be noticed because the rock may itself carry this stress increment. When the geological parameters are less robust, the stress increment may reach such a level that the rock in the pillar creeps under constant load. At that moment the longitudinal arching loses support from the rock pillar between the driving faces and the span now extends longitudinally between the supported ground back from the face on each side of the pillar. This overall span, being now much increased, produces additional stresses on the supported sections behind the face. The rock pillar is already yielding and yields firther, and the resulting distress causes existing cracks to open and the rock then becomes more unstable. In fact, this is not a critical moment; it merely indicates that the pillar is now redundant and may be removed - of course with the application of the necessary support. The only effect of such support, at least at this stage, is to retain in position some of the distressed material which in general lies outside the desired profile. Further re-adjustment of stress in the surrounding ground then takes place as the newly installed support at the break-through location takes up its share of the load.

Figure 3. Geological face mapping and monitoring results at the break-through position.

4.2. Interpretation of deformation measurements

As was mentioned above, the deformation process actually starts ahead of the face. The effect is generally reckoned to extend ahead of the face by a distance equal to about twice the tunnel diameter. If the ground is going to deform by a certain given amount, then, if the excavation is very slow, most of this deformation will actually have taken place before the ground is opened and an “initial” measurement can be taken. Subsequent measurements will suggest that the total deformation is small, simply because most of it took place before the “initial” reading. However, if the tunnel advance is rapid, then the readings will suggest quite large deformations relative to the “initial” reading, although the actual deformation from the undisturbed position is the same in both cases. It is therefore over-simplistic to regard the recorded deformations as representing the complete picture. The rate of approach is not taken into account in any way, despite its very real significance. Current practice, “state of the art” though it may be, is still very much in the early stages of development.

5. CONCLUSIONS

• Decisions are taken in the “fever” of events, and in some cases the first impressions on which decisions are made and the correctness of these decisions are often not confirmed by the after effects of the events. The decision-makers are on the Horns of a Dilema because whatever decision they make may subsequently be shown to have been either too cautious, leading to unnecessarily increased cost, or too rash, if things go wrong.

• When the question arises of providing additional support (at additional cost), it is not an analysis of this cost nor even necesseraly what effect is intended that is the centre of attention; what is of the greatest concern is the time available for decision and for application. A perfect solution applied too late is worse than an imperfect one applied in time. In dramatic situations what has to be decided is not just what solution to apply, but what solution can both be applied in time and will stabilise the siuation, as well as the probable effect which that solution will have in the later stages of the construction.

Figure 4. Geological face mapping and results of monitoring at chainage km: +52m.

Figure 5. Simplified sketch of the assumed stress distribution during breakthrough.

• Deformations are a function of time, but not just of the elapsed period from the initiation of monitoring. Deformations are greatly influenced - determined even - by the driving speed. Thus an interpretation of the amplitude of recorded deformations that ignores the rate of advance could lead to erroneous conclusions and decisions.

• Apart from the mechanical properties of the rock, deformations are a direct result of the stress field existing in the rock mass, and this depends on other factors besides the quality of the rock mass and the amount of overburden. Construction sequences and previous mechanical actions such as tectonic action may, in some locations, produce stress concentrations which, when disturbed, induce deformations much greater than the amplitudes recorded in previous “similar geological conditions”. Although deformation measurements do have a part to play, their interpretation is not an easy nor simple part of the decision-making process in tunnelling.

6. ACKNOWLEDGEMENTS

The authors are grateful to Professor Abdullah Memon for valuable suggestions during the preparation of this paper, and to Mr. Ferhat Aydin for his assistance with preparation of the geological mapping.

7. REFERENCES

Clay, R.B. & A.P. Takacs, 1997. Tunnel Collapses - case study. Tunnelling `97, p.481. Institution of Mining and Metallurgy / British Tunnelling Society, Olympia, 2-4 September.

ABOUT THE AUTHORS

A. Paul TAKACS:

During seventeen years with Contractor Tunele Brasov, Romania, Paul Takacs held different managerial positions for the construction of the following tunnels: Stana, Beia, Mestecanis, CirligulMic, Manastirea-Turnu, Grosi, Cerna, Gibei, Plostina and Basarab II. For the last four years, he has been supervising the motorway tunnels discussed in this Paper.

Robin B. CLAY :
A British civil engineer, Chartered for twenty-five years, Robin Clay has worked on many tunnelling projects of all sizes, and in many parts of the world, from the 80 km long Orange-Fish Tunnel in South Africa to a 30 metre long mini-tunnel one metre in diameter, and including the Victoria Line, the Hong Kong Mass Transit Railway and the Jubilee Line Extension, as well as sewerage schemes for Southend-on-Sea and Istanbul. For the last four years, he has been supervising the motorway tunnels discussed in this Paper.

Robin B. Clay 30th March, 1998