25th OYO Anniversary Lecture Meeting Tokyo, October 1979 DECISION MAKING AND FIELD MEASUREMENTS IN TUNNELLING K. Kovari, Prof. Dr., Head of Rock Engineering Department, Ch. Amstad, Senior Research Engineer FEDERAL INSTITUTE OF TECHNOLOGY ZURICH (ETHZ) Zurich, Switzerland SUMMARY After a brief review of the factors governing the structural behaviour of underground openings a survey is given of the type of structural problems encountered during the design and execution of tunnels. The use of deformation measurements during construction to assess safety and for a better understanding of the structural behaviour is illustra~ ted. Principles are formulated in order to establish adequate and efficient measuring programs for tunnels. The problems connected with the correct interpretation of the observations are also discussed. In the appendix some recently designed high precision instruments are described.1. INTRODUCTION ‘The design of underground openings like tunnels, subways and chambers in soil or rock was in the past almost purely a matter of experience. In the last two decades, however, new methods of site investigation, systematic measurements in the field and computational methods have been introduced as powerful design aids in order to arrive at safe and econo- mical structures. In fact, the increasing worldwide activity in the con- struction of underground openings and the frequency of large projects even under difficult geotechnical conditions call for a continual im- provement in design principles. The basic cause for the development of displacements in the ground around the opening or for the occurrence of rock and earth pressure phenomena is the disturbance of the stress field in the virgin rock or soil due to the creation of the opening. Each step in the excavation process involves a redistribution of stresses and strains in the ground, thus transforming the primary state of stress and strain into the secondary state. Temporary and permanent support like anchoring and tunnel lining have the task of restoring a new state of equilibrium, firstly for the construction period, and secondly for the service life of the structure. In many cases a new equilibrium state is required under the rigorous condition of limited displacements around the openings, for instance, in subway construction settlements of buildings and traffic surfaces have to be kept to a minimum. 2. THE STRUCTURAL BEHAVIOUR OF UNDERGROUND OPENINGS The tunnel support (lining, anchoring, etc.) and the surrounding rock forma unit (Fig. 1) which is looked upon as the actual structure in tunnelling {1}. In practice, the behaviour of this structure is often characterised by the nature of the rock pressure, i.e. the effective contact stress between the ground and the lining. The magnitude distri- bution and time variation of the rock pressure are important indicators ROCK ANCHORS TUNNEL LINING INVERT ARCH Fig. 1 Tunnel support and rock forming a structural unitROCK INITIAL CONDITIONS | en, od a DIMENSION METHOD OF ‘SUPPORT AND SHAPE EXCAVATION MEASURES Fig. 2 Factors influencing the structural behaviour of a tunnel of the sort of problem arising in tunnelling. The deformations of the tunnel section and the displacements in the rock together with their time dependent characteristics, however, are also good, and in many cases practically the only, indicators for the behaviour of the struc- ture. The protection of the opening against rockfall, keeping the rock pressure under control and limiting the deformations respectively in the most economical way often presents the main problem in tunnelling. In the solution of this problem it must be kept in mind that the behaviour of an underground opening depends essentially on the groups of factors shown in Fig. 2. 2.1 Rock Conditions The scope of the problems which may arise in tunnelling is best illu- strated by the fact that tunnels may have to be driven through comple tely cohesionsless soil, hard rock mass or through any intermediate type between these two extremes. The materials in tunnelling are not chosen as in some other branches of structural engineering; rather, they are encountered. Their mechanical properties are determined by means of geological surveys and soil and rock mechanics investigations. As far as possible this information should be obtained well in advance of con- struction. Generally drill holes or adits give access to the material in the area of the planned underground opening. Often, important infor- mation is gathered from outcrops on the surface, as well as by using the experience gained from previous constructions under similar geo- technical conditions. The rock properties on the scale of specimen size together with the structure of the rock mass determine properties on the scale of the construction. The rock structure is given by strati- fiction, schistosity and jointing. The latter constitute actual or potential surfaces of separation or slip. Therefore, their frequency and orientation in space are generally of great importance. The mate- rial tests in the laboratory comprise soil mechanics investigations,uniaxial and triaxial compression tests and frequently direct shear tests on surfaces of weakness. Load tests in boreholes or even trial sections in tunnels or chambers on a reduced or on full scale can, in certain cases, be applied with advantage as further methods of investigation. Of the many aspects that are important for the geological conditions only two are given special mention here, namely the presence of water and the rock types containing clay or anhydrite. Water inflow in even relatively small quantities into the opening may affect substantially the progress in excavation. The water may reduce the strength of the material by decreasing its cohesion or by the development of pore pressure decreasing the effective normal stresses. when tunnelling in saturated soils special measures, often very expensive, must be taken in order to prevent infiltration and to stabilize the ground, for example, grouting, ground water lowering, utilization of compressed air, hydro-shield or ground freezing techniques. Rocks containing clay or anhydrite give rise to special problems in tunnelling. Such rocks, e.g. marlstones and anhydrite, can swell, i.e. increase considerably their volume due to absorption of water, whereby a substantial amount of heave in the bottom of the tunnel may occur. The tunnel lining (invert arch), in resisting the heave, may be subjected to high swelling pressures. In tunnelling practice, unconstrained heave of up to 70 cm may occur {2} and swelling pressure of up to 3,5 MN/m* have been repor- ted to act on the invert arch (3}. Many of the unexpected difficulties that arise in tunnelling can be traced back to an inadequate knowledge of the material properties. The actual rock conditions are often, in fact, first known as the under~ ground opening is under construction. This is specially true for deep tunnels, for which borehole explorations either for technical or econo- mic reasons are out of question, or else can only be carried out to a very limited extent. Also, one only has to think of the possible varia~ bility of the material with respect to its petrographic composition and its structure (jointing etc.), then it becomes evident that it is especially important to determine the ranges in which the rock mass behaviour may be expected to vary. Here, not only statical but also purely constructional considerations can be important. The greater the degree of mechanisation in the method of construction, the more impor- tant possible extreme cases in the material occurrence become. For instance, when using the shield tunnelling method in soils, if the cutting edge comes up against occasional boulders a big time delay in construction may result, which leads to increased costs. Turning to another example, the economical application of a full-face boring ma- chine is not only limited by poor rock quality (too short stand-up time of the rock, insufficient thrust for the advance of the machine) but also in certain circumstances by a very hard, massive rock. The more uncertain the geotechnical predictions or variable the rock conditions, the more adaptable the constructional method has to be.ae 2.2 The Initial State of Stress in the Ground Due to gravitational forces and possible tectonic influences, the rock is already stressed before the underground opening is excavated. Thus, one speaks of an initial or primary state of stress, which, of course, is different from location to location (Fig. 3). There are two ways in which the initial stresses may give rise to difficulties in tunnelling. Firstly, the material in the vicinity of the opening often reacts to the changes in the stress field by failure and creep processes, which may lead either to the closure of the opening or, if it is hindered, to the development of rock pressure. Secondly, in hard rock at great depths the much feared phenomenon of rock burst may occur. This is characte~ rized by the explosive-like separation of plate-shaped pieces of rock often of considerable size, which may endanger the lives of the people working in the tunnel. The mechanism of rock burst has not, as yet, been adequately investigated. All that is known with certainty is that the orientation of the tunnel axis in relation to the directions of the principal stresses of the initial state of stress plays an important role. ‘The stress tensor in the rock cannot be determined theoretically because of the changing topographical conditions, the generally complex struc~ ture of the rock mass and its nonlinear stress-strain relationship, and the tectonic forces which may still be active today. Stress measure- ments in situ are only successful if the rock in the immediate vicinity of the measuring point can be assumed to be elastic, isotropic and homo- geneous {4}. Unfortunately, these conditions are as a rule only ful- filled in those cases in which knowledge of the initial stresses due to excellent rock strength is only of secondary importance. Thus, with re- gard to the magnitude and direction of the principal stresses, we are left with little more than suppositions. For a more or less horizontal surface terrain it is justifiable to assume that the vertical normal stress in the initial state is approximately equal to the overburden stress of the overlying rock or soil. No generally valid statement can be made about the horizontal normal stress component. It can vary from a small fraction to a multiple of the vertical stress. The lower and upper limits for the relationship between the horizontal and the verti- cal normal stresses may be assessed by the failure condition of the Oy Q%3 i = | 821 %22%3 931 932033) Fig. 3. Initial state of stress in the rockFig. 4 Influence of the span D on the stability in jointed rock material in the sense of the active and passive earth pressures. It may be noted that the greater the tendency for the material to creep and the greater the overburden pressure, the closer the initial stresses approach a hydrostatic stress condition. Tunnels located in slopes or beneath the bottom of a deep valley require special attention with re- gard to the initial state of stress. 2.3 Dimensions and Shape of Underground Openings ‘The relationship between the span of the opening and the average joint spacing is decisive, in many cases, for stability considerations (Fig.4). With increasing span D, or 2 respectively, the influence of the jointing becomes more marked an@ the“probability of an unfavourable joint combi- nation, which could give rise to a rock fall, increases. Thus, in the case of a subway through jointed rock the construction of stations gene- rally require special considerations even when the single track tubes might be left completely unsupported.In the particular case of soil with Fig. 5 Large underground chamber with statically favourable shapeQOOO Fig. 6 Possible adaptation of shape to increasing rock pressure no cohesion the vertical pressure on the tunnel lining in the roof in~ creases with increasing span of the tunnel, the ratio of span of tunnel to height of overburden being also an important factor. If this ratio is less than one it is not possible to develop a noticeable arching effect in the soil, not even in heavily~jointed rock. Especially large dimensions in the construction of tunnels or rock chambers are only possible, from the point of view of safety and economy, by imparting a special shape to the profile. A good example illustrating this point is a chamber in the form of a vertical cylinder with a spherical clo- sure (Fig. 5). Statically this shape is very favourable, for horizon- tally we have the effect of a closed ring and a double arching action exists at the roof closure. Cavities of this form and with dimensions of about H = 80m, D = 45 mare at present planned for underground nuclear power plants. The shape of a section is also important in the case of a tunnel. However, as a design parameter, it is, in many in~ stances, not given the attention it deserves. Should rock conditions be encountered in which high rock pressure is expected, the shape of the profile should be selected in such a way that an arching action in the rock and tunnel lining may be developed. In railway tunnels, for instance, this can be achieved by choosing shapes as shown in Fig. 6. 2.4 Method of Construction and Support Measures ‘The method by which the opening is excavated along its longitudinal direction and in its cross-section can have a significant influence on the development of the rock pressure and the displacements in the surrounding rock. In the case of a tunnel the profile can be excavated in a full face operation or by dividing the section in different parts and excavating it in different sequences (heading and bench method, multiple drift method etc.). Difficulties of various kinds can be over- come more easily when working in smaller cross sections. When the rock conditions require it, therefore, the profile must be, excavated in two or more stages (Fig. 7), whereby staging is also employed in the direc- tion of the tunnel axis. The first stage of excavation is in many cases well in advance of the works for enlarging the section to the full pro- file thus providing a useful means of rock exploration. Depending on{ Fig. 7 Examples of multiple stage excavation in the tunnel section (pilot tunnel with boring machine, head and bench, and side drift tunnelling method) whether the problem is to control the rock pressure or to limit the @isplacements in the neighbourhood of the tunnel, various construc- tional procedure may be chosen along the axis of the tunnel. This is illustrated by practical examples, one for a subway construction and the other for a deep tunnel, both driven through a soft rock. For the cross-section one can in both cases proceed according to Fig. 8. For the subway tunnel Fig. 8a, in order to avoid undesirable settlements of buildings in its vicinity, the invert arch should be placed as a) = S77 INVERT TEMPORARY SUPPORT PERMANENT SUPPORT ») INVERT ~05-2 YEARS Fig. 8 Placement of the invert arch a) subway: close to the face b) deep tunnel: at great distance from the faceuickly as possible. The time required to complete a full ring may be only a matter of days or a couple of weeks. Thus, in this way, at a distance of less than one tunnel diameter a closed ring is formed which eat depth tb), where high rock pressures can be developed, consideral is statically extremely efficient. In a tunnel situated at (Fig. deformations may be deliberately permitted using a exible temporary support to keep them under control. In any case, it is impossible to preve! since the pressures that would occur may be of the order of magnitude of the initial stresses (in a depth of 1000 m there would be an over~ burden pressure of about 30 MN/m? in the rock). Thus, with the protec- tion of a flexible, temporary support, one allows radial displacements to 50 cm or more, which in some cases - Any further deformation that might occur prevented using a suitable closed ring-shaped per- ‘This may follow the working face of the heading in a distance of a few kilometers. With regard to the conventional methods @ elementary requirement of carefully controlled he least disturbance of the surrounding rock is ntioned here. The rock should not be unnecessarily loosened by blast~ the deformations completely even when using a stiff lining, of the sides of the opening of up may take a year to develo) can then be safely manent lini of excavation only th blasting which causes ing, as this would result in a considerable loss in strength. In many instances heavy lining is necessary only because of poor blasti work. Such a case is shown in Fig. 9 together with a tunnel in the same rock but with smooth rock surface and with no support at all. The indisputable advantage of blast-£: that they do not effect the in situ rock quality around the opening mechanical excavation methods is In summarising the above it holds true generally that the method of excavation and the type of support system (rigid or flexible) as well as the time and place of its installation hav a profound influence on the haviour of the undergroun 9 Two different blasting techniques applied in the same rock at the construction of adjacent roadway tunnels (Photograp! COEPE, Rio de Janeiro)3, THE PROCESS OF DECISION MAKING IN TUNNELLING In order to obtain a safe and economic structure the engineer has to make decisions on the following item = location, alignment, shape and size of the opening - method of excavation, both in the section and in the longitudinal direction = support measures, temporary and permanent - dewatering, ground improvement etc. Decisions are required prior to, during and in exceptional cases after construction. It must be emphasized that the decisions are not only a matter of purely theoretical consideration but in many cases they are somewhat restricted by contractual aspects. ‘The technical criteria (Fig. 10) for correct decisions may basically originate from the safety of the opening during construction and during its service life or from displacement restrictions, and in some cases from both. The sources of information for the structural decisions are: = geological explorations, field tests - laboratory investigations - statical computations - field measurement - the engineer's own experience Again the flow of information generally extends from the initiation of the project up to its completion. Modern tunnelling is characterized by the systematic use of all sources of information in a balanced manner. A clear understanding of the factors influencing the behaviour of an underground opening under specific conditions can only derive from the engineer's own experience and from his theoretical knowledge. Experience manifests itself in good structural judgement. Together with laboratory investigations, statical computations and field mea- surements it forms the basis for decision making both at the planning and construction stage. To what extent such modern aids should be applied on a given project depends solely on the nature of the prob- lems arising. In the following an attempt will be made to give an up- to-date survey of the possibilities and limitations of field obser- vation techniques. Computational methods as a design aid in tunnelling have been discussed elsewhere {5}. Here we only want to point out that by means of statical computations a prediction of the structural be- haviour of the opening is aimed at in an analytical way. The inter relationship between the various factors, for instance rock proper- ties, shape and dimensions of the opening, initial state of stress etc. may be clearly seen in the calculated results. But although these results are available at the design stage, they are subject to great uncertainties. Measurements, on the other hand, enable its be- haviour to be observed directly, without the actual mechanism, which gives rise to its behaviour, necessarily being completely illuminated. The measurements are usually carried out during the constructional0 CRITERIA SAFETY DISPLACEMENTS ‘SETTLEMENTS CLOSURE, Fig. 10 Criteria for decision making phase and if carefully planned and executed they give a true picture of the behaviour of the structure. From these considerations it is obvious that computations and measurements complement each other and only when combined are they capable of leading to a correct explana~ tion of the structural performance in complex geotechnical situations. 4, FUNDAMENTALS OF FIELD INSTRUMENTATION ‘The practical significance of systematic measurements for a given project depends upon the extent to which the results of the continous observations are able to influence the constructional work. This point is well illustrated by means of two examples. The first concerns the case of shield tunnelling with lining segments. Here the most impor tant constructional decisions, for instance deciding upon the shield diameter based on the anticipated soil movements and lining deforma~ tions, or designing the segments themselves, have to be made well before the start of the construction. ‘The observation of the actual deformations of the tunnel profile, the movements of the surrounding ground or settlements at the ground surface have the function, mainly of checking the structural behaviour with regards to a satisfactory design and proper execution of the works. In this way shortcomings-u- arising in backfilling the space between the rings and the ground or concerning insufficient support of the tunnel face can be detected. Using a tunnelling method with shotcrete and anchoring as a support, which may in cases of favourable ground conditions also be envisaged in subway construction, extensive measurements can really serve as feed back signals for the constructional process. Here on the basis of careful statical computations a concept is worked out for the exca~ vation sequences both in the cross section and along the axis and for the corresponding support measures. If the measurements indicate a substantial deviation from the anticipated behaviour of the structure, then the most important corrective measures in the construction can still be applied. The above comparison of the two methods of construc~ tion restricted itself to the possibilities of influencing the tunnel- ling process by a proper use of measurements and should in no way be regarded as a general evaluation of the two methods. Which of the two methods of construction should be applied in a particular case is decided, of course, by economy and the attainable progress in advan- cing the tunnel. The purpose of measurements. In practical terms the following formula~ tion of the objectives of an observational program are commo! Assessment of rock mass properties Safety control measures Effectiveness of particular support measures Comparison of real performance with theoretical prediction Special problems The above classification of the objectives of measurements are only intended to point out the most essential features, as in many cases several aims are envisaged with the same measurement program. ‘The measured physical quantities: Depending on the particular problems ‘the observations most frequently refer to one or to a group of the following quantities: - strains - relative displacements (no fixpoint) - absolute displacements (with fixpoint) - changes in curvature (in tunnel lining) - stresses in lining and in rock mss - rock or earth pressures on tunnel lining, forces in rock anchors When planning a measuring program some sound principles have to be followed in order to obtain useful results for practical purposes with a minimum of cost expenditure. Principles of field measurements - Correct formulation of the structural problem, the solution of which requires observations = Selection of the most sensitive physical quantities - Assessment of order of magnitude of the measured quantities, conclu- sions with regard to required accuracy ~- Selection of measuring techniques, instruments, location of measu- xing sections, reading programero - Assessment of possible sources of error in the readings well in advance - Application of monitoring with overlapping results for complex situations - Employment of reliable instruments and competent personnel only - Continuous data processing, establishment of tentative emergency levels, correct flow of information. Experience shows that when observing these principles field monitoring really turns out to be an invaluable aid in design and execution of underground works. 5. CASE HISTORIES With the help of examples chosen from tunnelling practice the basic considerations given above will be further discussed in the following sections. When dealing with case histories the actual problems arising in the various projects will be briefly formulated, the applied moni- toring technique referred to and the relevance of the obtained results to the constructional problems discussed. The measuring techniques and the associated instruments developed at ISETH are briefly described in the appendix. 5.1 Convergence Measurements ‘The measurement of convergences, i.e. of the changes in distance bet~ ween two points of the excavation or lining surface, is one of the simplest and least expensive operations. In Fig. 11 typical applica- tions are shown, for a cut and cover situation and for a circular tunnel section. Fig. 12 indicates how by a mesh of measuring lengths b) Fig. 11 Typical application of convergence monitoring a) cut and cover tunnelling b) control of diameter change3 7 ee maric SYSTEM OF REFERENCE Fig. 12 Determination of the complete deflection of a tunnel lining by a mesh of individual convergence measurements the complete distortion of the cross section may be determined. The displacement vectors uj and vj of a point are referred to an arbitrari- ly selected kinematical system A-B. Generally three displacement wec- tors should be known or fixed. In Fig. 13 it was assumed that vg * va = ua = 0. Let us now determine the displacements uj, vj and ug of the points B and i in such a system from the measured quantities lag, Mlaj and Alp; based on the given geometry (Fig. 13). It can easily be shown that the following equations hold: Mag = Up Mai = ujreosa + vysina Blgi = Alagrcosa + vysing - uj-cos and also in the inverse expressions up = Alp Alaj*sin8 + Alap‘sina+cos ~ Alg;-sina Sina-cosB + cosa-sin8 Alaj*cos8 - Alap-cosa-cosB + Alp; -cosa Sind-cosB + cosa-sin8-14- Fig. 13 Kinematics of a single triangle for determination of displacement vectors It is advantageous to introduce in a mesh some control lengths "c" as indicated in Fig. 12. In such a manner the reliability of the indivi- dual readings can be checked by the simple relation te = (uj - a) = where e indicates the error actually encountered. In case the value e is negligible the correctness of the readings is confirmed; if not, the readings might have to be repeated. Using a computer program (e.g. program INVAR of ISETH) for data processing acceptable values of e can also be adjusted by the method of the least squares, thus increasing overall accuracy. ‘The DISTOMETER-ISETH, based on the principle of the invar wire, is now in operation world-wide in more than 20 countries and it has proved to be a suitable instrument especially in the adverse conditions of a tunnel construction. A detailed description is given in Appendix 1. In many instances such simple measurements are carried out merely to ascertain if a state of stable equilibrium has already been reached, will be reached, or instabilities are to be expected. 5.1.1 Imigrantes Highway This three-lane 55 km long highway connecting Sao Paulo with Santos in Brazil involved the construction of altogether 16 viaducts,and 11 tunnels with the total length of 3825 m {6}. The tunnels - most of them slope tunnels - were constructed under difficult topographical and geomechanical conditions {7) and excavated simultaneously by diffe- ent contractors, The tunnels present a considerable size of profile with a sectional area of 120 m’. The general sequence of excavation is-4s- o | | i ‘CALOTTE EXCAVATION @CALOTTE PERMANENT TEMPORARY SUPPORT LINING @©SIDE WALL CONCRETE @ INVERT EXCAVIATION AND CONCRETE Fig. 14 Imigrantes Highway, construction sequences of the tunnels {9} shown in Fig. 14. The convergence measurements {8} were expected to give information about: - The type of rock pressure which may develop in different sections of a particular tunnel. It was important to identify the type of possible rock pressure phenomenon. = The stability of the slopes affected by tunnel construction. Different arrangements of the measuring lengths used may be seen in Fig. 15. The observations had to cover all stages of the constructionre Fig. 15 Different arrangements of the measuring lengths in the tunnels of the Imigrantes Highway {9} according to Fig. 14. Special attention was paid to the behaviour of the temporary lining in the calotte which consisted of steel ribs with or without shoterete. Fig. 16 shows an example of readings in the case of a local instability which occured in tunnel TA-4 near to the face. By means of monitoring, this instability could be detected in its very early stage, thus permitting the installation of emergency supports by wooden timbers (Fig. 17). The monitoring of the permanent lining in the calotte during core removal and side wall construction (Fig. 18) was of great interest too. In some cases the effect of these construc- tional measures on the readings could be clearly observed (Fig. 19). TIME {DAYS} EMERGENCY SUPPORT SHORTHENING OF THE CONTROL LENGTH PLACING OF PERMANENT LINING cm Fig. 16 Indication of failure in the tunnel TA~4 by the Distometer readings, Imigrantes Highway {9}-18- 12 LENGTHENING o we rr & 2-12 - ot z7 ive Sey | Zz -16 Aeon Sef 4 | Ee SIDE WALL 2S -20 a as 30 60 90 120 _TiME [Days] Fig. 19 Influence of the core removal and the side wall construction on the control lengths. Tunnel TA-9, Imigrantes Highway (9) When interpreting the readings for the total of 21 monitored measuring sections in seven tunnels it was kept in mind that - The permanent lining forms together with the surrounding rock essen- tially a three dimensional structure. This statement holds especially true when considering an asymmetrical excavation process and step by step side wall construction. Simplification to a two dimensional case is allowed whenever uniform conditions prevail in the vicinity of a measuring section. - The type of rock pressure phenomenon and the supporting effect of the lining may be estimated from the order of magnitude and rate of the deformations. = Due to time limitations a useful back-analysis with the aid of modern computations could not be carried out. Conclusions from the results of field measurements in the Imigrantes Tunnels: It could be shown quantitatively that no exceptional rock pressure phenomenon occurred in the tunnels. The rock was self-suppor- ting as it had been supposed at the time of the first site inspection. only loosening pressure has occurred. It was clear that the permanent Linings of the tunnel were considerably overdesigned. Although the measurements were initiated at a very advanced constructional stage important savings in concrete and steel could still be achieved. The readings further demonstrated that the core removal could take place-19- in one working unit and not in short trenches as it was originally assumed, thus speeding-up the excavation and making it less expensive. As a further two tracks for the Imigrantes - with new tunnels in the same area ~ are planned, the measurements also provide very important data and conclusions for these new major projects. With a view to these two new routes a systematic geological mapping has been carried out in each of the tunnels. The problems related to slope stabilities deserve some attention. The large dimensions of the tunnel sections, the un- favourable alignment of the tunnels in steep slopes with slight over- burden in the portal zones and some critical cuts for the "service road" have raised the problem of slope stability. During the first site inspection it could easily be recognized that the problem of slope stability in the portal zones has in some cases been overestimated but in other cases almost completely ignored. By means of an adequate and early initiated measuring program considerable savings in time, effort and money could have been achieved for this type of problem too. 5.1.2 Tunnels with cut and cover method Recently in many cases cut and cover solutions are given preference over above ground solutions in road and railway connections for the sake of environment protection. The interaction between the back-fill material and the concrete arch is decisive for the load bearing capacity of such a system. For safety control and for a better understanding of the interaction between the thin reinforced concrete vault and the com- pacted soil behind,the springline displacement measurements yield in~ valuable information. As indicated in Fig. 20 a relatively coarse mesh may already provide well interpretable results (Fig. 21) and the levelling of some selected points may confirm accuracy and reveal possible differential settlements of the foundations. One of the prob- lems of special interest in the case of the two lane twin tunnel Maria Zell near Lucerne, Switzerland, was to find out the deformations caused by the partial filling up to the crown (Fig. 21) and by the subsequent earth layer above the crown. In Fig. 22 the displacement vectors refer only to this additional loading case. It is clearly seen that this load Fig. 20 Layout of the measuring lengths for the determination of vault deflections @: levelling point)Fig. 21 Pig. 22 Fig. 23 = 20- Displacements of the vault caused by the back filling up to ‘the crown Displacements caused by the load of the cover only (hatched layer) Total deformations of the vault-2- moves the vault in the spring line area against the compacted soil thus activating a lateral support from the earth. From Fig. 23 which repre- sent the superposition of the displacements from the two proceeding figures one may conclude that the effect of the initial lateral loads is greater than those of the vertical ones. In all cases under discus- sion the result of the control levelling agree satisfactorily with the Distometer readings. The monitoring program together with CBR-tests revealed already at an early stage that the overall soil compaction was not sufficient, giving rise to the need for some soil improvement pro- cedure. From the results of the complete measuring program the authors were able to draw important conclusions for correct computational models of such cut and cover structures. 5.2 Changes in Curvature and Strain of the Tunnel Lining Rock pressure causes the lining of a tunnel to deform. An alternative method to determine the deflection of the lining is given by systema~ tic measurements of both change in curvature and strains along the intrados of the tunnel lining in consecutive points. As an advantage, the internal forces (bending moment and normal force) and the rock pressure also can be calculated {9} provided the material properties of the lining material are properly defined. 5.2.1 Theoretical considerations ‘The following investigations are restricted to deformations and loa~ dings in the plane of the arch. It is also assumed that the deforma~ tions are small enough to permit the use of small-deformation theory as is common in structural analysis. An arch unit with the measuring points A, B and C is shown in Fig. 24. These points are defined by measuring bolts which lie at a distance from the axis of the arch. The chord AB designated the stretch L and the perpendicular distance between C and L defines the "height" F of the arch. The measured quantities are the changes in the lengths F and L. feF- l]=L-f ‘They are chosen positive for a shortening of the lengths F and L. For the computation of the curvature « (which in fact should be called the change in curvature) and the axial strain ¢ from the measured quantities f and 1 the principle of virtual work is used. For small angles ay, it follows that 8 1 Be k=—f, es —+ 2 ee ele £ The state of deformation of the arch is completely defined by « and c along the axis of the arch without the need for any stress-strain re- lationship for the lining material.22 (deformed) (undeformed) Fig. 24 An arch segment in the undeformed (A, B and C) and deformed (A, B and ©) state From a given curvature x and axial strain ¢ the strains ¢, in any arbitrary fibre i at a distance y; from the neutral axis can be found. Introducing realistic stress-strain relationships o = £(€) for each set of values of curvature « and strain ¢ (or the measured values of £ and 1 from which they are obtained) a unique pair of values of M and N can be found N J£(e)da, A = " Syf(e)an BR With regard to plasticity theory it would be necessary to take into account different loading, unloading and reloading behaviour in the material laws. In order to develop the fundamental equations of the problem in apaly- tical form it is assumed that the rock-pressure loading is a continuous function of the arch length s. In Fig. 25a is shown a finite segment of the arch with the distributed rock load having components in radial and tangential directions, p(s) and t(s) respectively. Let us consider an infinitesimal element cut out of the arch (Fig. 25b) with all the for- cies acting on its.23 Fig. 25 Acting internal and external forces a) a finite segment of the lining carrying a distributed rock load b) infinitesimal element for equilibrium considerations Neglecting terms of second order magnitude the equilibrium condition for the element leads to the equations no am Co R as? 1 au aw oo + Rds as ‘Thus the method for pointwise determination of the required load quan- tities pj and ty from discrete values of Mj and Nj is given by sm oL Pos oop mat = am + a 1 1 = oy Mian 7 Mien + 2a - My-1)) For 1< i « (n-1) Rock Pressure Determination in a deep Tunnel The Integrated Measuring Technique found its first application in the Gotthard Road Tunnel Heading North {10}, where steel ribs were used (Fig. 26).- 24 Fig. 26 Gotthard Road Tunnel. Method of construction in Mezosoic section {13} 40. L © Change in F wi \\ + Change in L MEASURED VALUES [;4; mm] ° 250 500) 750 1000) 1250 ARC LENGTH [m] Fig. 27 Distribution of the measured values f and 1 along the arch- 25 - Fig. 28 Distribution of the calculated moment and normal force along the arch * Load cell measurement of normal force N at A and B To give an example of a series of readings we consider Fig. 27, in which the measured values f and 1 are given in terms of the arch length. The scatter and some stray points (point P) are typical. The full line is the smooth curve obtained by spline interpolation. ‘The corresponding distribution of the stress resultants M and N along the axis of the arch is shown in Fig. 28. The derived rock pressure distribution and its change with time is shown in Fig. 29. The maximum values of the radial load was determined for the last but one reading and found to be around 200 kN/m?, The I-section used here has a load capacity of about 500 kN/m? for a uniformly distributed load. ‘The distometer measuring lengths indicated in Fig. 26 were used to check independently the accuracy of the curvometer and deformeter mea~ surements (see Appendix II). Displacements in Soil, Rock and Concrete Such measurements require the drilling of boreholes (or the placing of some casing tubes into the concrete during casting). In the following we shall restrict our attention to the possibilities offered by multiple Borehole Extensometers.= 26 - Fig. 29 Distribution of the calculated rock pressure for different readings in time 5.3.1 Tunnel _in Swelling Rock The highway tunnel T8 between Biel and Sonceboz in the Swiss Jura passes through a mudstone formation (Effinger layers) for a distance of about 300 m and then for 600 m through limestone beds (Fig. 30). For the section of tunnel situated in mudstone two kinds of problems were faced. Firstly, exploratory drill holes in the vicinity of the portal revealed a very weak rock. It was feared, therefore, that ex- cessively large deformations would occur, as in cases of the presence of genuine rock pressure. Secondly, it was known that the given mud- stone possesses a considerable swelling capacity. The bottom of the tunnel, therefore, was expected to develop heave, or if this were pre- vented by an invert arch, to exhibit high swelling pressures. Obser~ vations made in the past on a great number of major tunnels in Switzer- land {1)had shown that swelling effects were only likely to be of im- portance at the bottom of the tunnel. On the basis of these considera--27- 300 m 200 m Sy RS SERQVnss Fig. 30 Geological profile of the tunnel 78 with the measuring sections M1-M9 tions it was decided!) to drive the tunnel using the top heading method and a selective cutting machine for excavation (Westfalia Linen). In this way a smooth rock surface was obtained and no harmful effects on the rock were caused by excavation. As temporary support systematic anchoring with reinforced shoterete lining was used. A permanent lining was necessary to support the water-proofing. Because of the expected swelling effects it was planned from the beginning to have an invert arch as part of the permanent lining. By means of measurements {12} during the construction period the following problems had to be in- vestigated: a) Determination of the optimum anchoring (i.e. load capacity, anchor length and spacing) as well as the strength of the shotcrete to guarantee the stability of the roof and the walls and to limit the movements. b) Determination of the in-situ swelling capacity of the mudstone along the axis of the tunnel, i.e. collecting the necessary data for designing the invert arch. Tosolve these problems measurements were carried out in three cross- sections MS, M6 and M9 (Fig. 30) with convergence observations and borehole extensometers. In the six other sections only convergence measurements were made. The measuring sections were set-up directly at the respective faces or following a bench excavation. The arrangement of the borehole-extensometers is shown in Fig. 3la. The borehole ex- tensometer (Fig. 31b) is a further development {13} whereby with this type several measuring heads can be installed in the borehole one after L)~ pr. P. Kellerhals (Geologist, Bern) ~ Suiselectra (Consulting Engineers, Basel)- 28 - HEAD FOR MONITORING @ THE RELATIVE DISPLACE MENT BETWEEN@DAND® HEAD FOR MEASUREMENTS 4} @) WITH THE MOBILE ELECTRONIC GAUGE ANCHOR. ROUTING b) Fig. 31 a) Layout of the borehole extensometers b) Details of the Unirod-Extensometer ISETH (here shown with two heads only) 40 30 DISPLACEMENT ff 0 200 400, coo | 80 1000 1200 [DAYS] | a | a) | | 8 8 | ° ° ° 3 fs Poo RY el x a x x ou Fig. 32 Readings of the extensometers in the bottom of the tunnel for different phases of construction (M5)~ 29 - another. Thus parts of the extensometer with the corresponding measu- ring head can be removed in the course of excavation and the remainder will be still in operation for continued readings. As readout unit for the deeper installed measuring heads a portable inductive transducer is used. The readings taken for the floor in section M5 with the multihead ex- tensometer are shown in Fig. 32, The swelling behaviour of the material in stage (I) if construction is with a flat bottom was observed over a period of 150 days. These initial measurements gave an early warning of the danger of swelling. With the successive excavation of the bench the upper head of the extensometer and one connecting rod were removed and the second head was exposed for direct readings. This new stage (II) of excavation caused an accelerated heave of all extensometer points, especially of point 2. This conditions could be observed for 450 days. Following this a large-scale load test was made by using excavated material as load. The corresponding stress on the floor was p = 60 KN/m?. Even with this small loading the floor heave could be considerably slowed down, in fact nearly to standstill. This somewhat reassuring result, however, was not of significance to the whole stretch of the tunnel in the marlstone, as one knows from oedoneter tests that swelling depends, not only on the mineralogical composition but also fon the load history of the material. Greater heights of overburden means higher initial stresses and thus an increased swelling capacity. This was in fact observed in section M9 where the overburden is 300 m giving a vertical normal stress of about 8 MN/m? in the undisturbed state. In this measuring section three extensoneters were installed at the bottom of the tunnel after the bench had been removed (Fig. 33). The results of the measurenents showed, in comparison to those of section M5, a much greater rate of heave of the floor, but also a non- uniform distribution across the section. In Fig. 34 a comparison of the swelling heaves in the various measuring sections is shown. Once more we note, that different in-situ swelling potentials may be traced DISPLACEMENT [inn] ° 200 400 [Days] Fig. 33 Nonuniform swelling heave of the bottom in section M9ao M5 M6 Mg DISPLACEMENT [mri] 9 1 10 [YEARS] 100 Pig. 34 Comparison of the swelling heave with time extrapolation in the various measuring sections back to the combined action of different mineralogical compositions and of different overburden stresses. In a semi-logarithmic represen- tation the time variation of the movements during the period of obser- vation can be approximated very well by straight lines. The extrapola~ tion, therefore, to a further 100 years seems to be reasonable, pro- vided the conditions of the construction remain the same. The authors are aware of the great uncertainties involved in such an extrapolation. However, it is felt, that the displacements obtained in this way, are on the safe side. From these considerations it is concluded that the heave of the flat bottom of the tunnel, without the restraint of the invert arch, could reach up to 20 cm depending on the height of over- burden and mineralogical composition respectively. Three alternative constructional measures were considered: a) the prevention of heave by means of an invert arch construction b) permitting the rock to expand into the tunnel during the entire service period c) permitting a large amount of swelling and to resist further expansion by a strong invert arch construction. ‘The basic relationship between bottom heave and swelling pressure p is shown schematically in Fig. 35. Of great importance are the intercept values p* and 6 on the coordinate axes. In the present case the value 6-31- y+ NCOSTHALNED HERVE 4 Svein presse Fo ao weave 4 HAVE OF worTon Poe TD THE PACENNT OF Ive eC a ACTON HEE OF IWveRT akc a+ ACTUAL PRESSURE On INVERT INERT ARC BOTTOM HEAVE ov SWELLING PRESSURE p Fig. 35 Characteristic lines for the swelling rock and invert arch could be adequately estimated by means of measurements (Fig. 34). The extreme value of swelling pressure p*, however, remains an open question, as from the few laboratory tests carried out it was not possible to ob- tain a reliable data basis. Therefore, it cannot be ruled out that p* could even take on values corresponding to the primary state of stress. This is all the more likely, since the content of minerals causing swelling (montmorillonite, illite) in the marl increases with the ad- vance of the tunnel, At the maximum overburden depth (Fig. 30) the ver~ tical normal pressure could, as was mentioned above, reach a value of up to 8 MN/m2. On account of this uncertainty and the lack of knowledge regarding the exact form of the characteristic line for the swelling rock it was decided to seck an alternative solution to the problen, by permitting a sufficiently large heave, thereby controlling the develop- ment of swelling pressures. The influence of permitting deformations on the swelling pressure is readily seen in Fig. 35, if the values Pa and Bg, which correspond to the different values of 6, and 8, are com- pared. ‘The chosen solution is illustrated in Fig. 36. Essentially it was planned to install an invert arch, s0 that the abutments of the tunnel Lining are stabilized with respect to possible rotation or in- ward horizontal displacenents. Further, it was necessary to construct a safe base for the road deck with long-term stability. The invert arch should be capable of transmitting its own weight, that of the fill material and of the traffic loads to the rock. The concrete lining Lies on supports at certain intervals and the intervening spaces are fillea with foam pads, This foan material, made of a kind of plastic (Polystirol) with a voids ratio of 98 4, serves as the exterior form- work for concreting. For the support pads a lightweight concrete with a small cement content was chosen. Its material properties and its Shape (the ratio of height to base length) were selected in such a way that failure occurs at a prescribed low strength associated with extre- mely large deformability (strains of up to 60 t). Fig. 37 denonstrates that despite of the uniaxial loading a considerable amount of residual-32- INVERT FOAM- — MATERIAL ‘CONCRETE LAVER Fig. 36 Invert arch placed on low strength support pads a) section parallel to the axis of the tunnel b) cross-section through foam material ¢) cross-section through support pad UGHT WEIGHT ‘CONCRETE STRAN € [%] Fig. 37 Stress-strain curve of low strength support pad in "uniaxial compression"istrumented with measuri ndix III) to suppleme iether 1 eleven mo imp LF to be £ixeae Sa VERTICAL S S— DISPLACEMENTS SZEEA 10 7 Ir 7 JO FU] Ce ee MEASURING 54 8 MONTHS * MARKS Fig. 39 Sliding Micrometer measurements in tunnel T8 (M9). Development of vertical displacement after completion of the tunnel “olf Fig. 40 Examples of useful application of Sliding Micrometer a) tunnelling b) slurry wall-~435- possible application of this instrument in geotechnical engineering the monitoring of diaphragma walls, dam foundations and piles could be mentioned. 5.3.1 The Instrumentation of a Metro Station Since September 1978 measurements with the Sliding Micrometer-ISETH have been a constituitive part of an extensive monitoring program during the construction of the subway station Republica in Sao Paulo. This is situated in the downtown part of the city and it is the crossing point of the main north-south and east-west lines requiring a very im- portant construction in cut and cover solution!), as the most important structural element the 35 m deep slurry wall was considered with a total length of about 340 m. The aim of the measurements was to provide an accurate check on a new computational method for the design of this wall, its support (braces and anchors) and to yield an early warning Af unexpected behaviour would develop. On both sides of the wall a ca~ sing with measuring marks (Appendix III) were attached to the reinfor- cing cage and concreted into the wall. From the strains along the measuring lines the curvatures may be found and making simplyfying assumptions about concrete properties the ben ding moments may be estimated. Two measuring tubes are placed in each of three wall sections and the observations have been carried out con- tinuously during the stage-wise excavation process. The results*) for four particular stages of excavation are shown in Fig. 41. The arithme- tic mean of the observed strains € on both sides of the wall might be interpreted partly as the shrinkage of the concrete, whereas, the strain differences lead to the values of the wall curvature x. From these cur~ vatures, by means of some simplifying assumptions, the bending moments in the wall may be assessed. It is of interest to note that based on the results of the comprehensive measuring program and the computational procedures decisions on a substantial reduction in the lower level wall- supports could be made. 6. CONCLUDING REMARKS Field measurements are now recognized world-wide as an indispensable aid for correct decision-making in tunnelling. Successful measurements xequire both a thorough understanding of the specific problems arising in tunnelling and a close familiarity with instrument techniques. In this report an attempt was made, on the one hand, to show the signifi- cance of monitoring by the discussion of some case histories and, on the other hand, to give information on new developments in measuring techniques. Vconsulting Engineers: PROMON ENGENHARIA SA, Sao Paulo 2)gxecution of measurements by COEPE, Consultoria Estudos e Projetos de Engenharia LTDA, Rio de Janeiro| A Mod eae © sTRAN € ‘cunvaTune x <20_0 290 ch coNeRESSION Senrention TRAN ccunvrune x s ‘ co compat srtan ¢ cunarute x pom 0-0 90H, to -90 p20 1 boon rn ~. >> | 1» covenesson orn e cumarune Cor 00 0) -s 010 _mo_x0 2m a0 -m0 0 0 40 o_o wi] Teal (1 commnession © extention: Fig. 41 Metro Sao Paulo: Sliding Micrometer measurements in a section of the slurry wall for different stages of excavation - strain distribution along the lines A and B - curvature distribution for assessment of bending moments-37- APPENDIX I : DISTOMETER-ISETH This instrument (Fig. 42) was developed {14} to meet specific require- ments in tunnelling works, although applications in other fields of geotechnical engineering are also quite common. The following important aspects were aimed at during the development and construction of the apparatus: a) The average error m of a measurement should not exceed for a measuring distance of 1 = 5.0 m. b) The need for carrying out such measurements quickly in order to avoid disturbance of the circulation or construction activity in the best possible way. c) Straight forward handling of the apparatus enabling these mea~ surements to be carried out by site supervision personnel. 4) Sturdy construction in order to reduce the risk of damages. 40.02 mm A high accuracy in the measurements is often required in practical si- tuations, especially when the movement of a point. can only be determi- ned by indirect measurements with an unfavourable error propagation, or when the velocity of creep has to be determined in a short time. The best solution under such partly contradictory requirements was found in the well-known base of the tensioned invar steel wire. With respect to the little space available in the majority of the underground works it was clear at once that a solution with pulleys and a weight for tensio~ ning the wire could not be considered. The traction therefore has to be produced with great accuracy by a mechanical device. This is only possible by using a light but highly accurate dynamometer. The elastic deformation of a 1.0 mm diameter wire with a length of Lag = 5 m and a tension of F = 8.0 kg is Al = 2.4 mm. Since the required accuracy of a single measurement is + 0.02 mm, the force has to be reproduced with an accuracy of + 0.5 % to + 0.8 % which corresponds to a tolerance of approx. 50 grams. A special gadget is used for the precision tensioning of the wire, This is, fortunately, only required over a narrow range close to the get value of F = 8 kg. The tensioning device moves the end of the wire towards the micrometer with the help of a precision thread (Pig. 42). A longitudinal bearing and a spring are used for the unre- strected precision setting of the instrument. The micrometer scale has divisions corresponding to 0.01 mm and a range of 100 mm. A weight is used for the control of the dynamometer which is calibrated at the be- ginning of each series of measurements. A gauge made of invar rods is used for the control of the instrument length. With this procedure it is possible to compensate any temperature effect. The wire is held by conical clamps constructed in such a way that no bending of the wire ends or an eventual rupture by fatigue may occur. All connections are simple bolts. Universal joints make it possible, to keep a static equilibrium of the system (instrument + tensioned wire) completely free of constraints. There are setting bolts shown in Fig. 42 which define the fixed points A and B. They can be inserted either permanently into drillholes in rock or concrete or temporarily into little base plates welded on to steel members, where they can be removed each time after measurements. In such a way all permanent protrusions apt to be damaged are avoided.or each basis length Lag a separate wire with corresponding length is ared. After use, the wire is rolled up over a spool for protect y of the instru- against any possible damaging or bending. The accurac 10-6 x L (L being the length of the invar wire} ment is better thai sont JOINT DISPLACEMENT GAGE JOINT t { INVAR WIRE 4: k i oa fe (Cpe ao i DYNAMOMETER f |S serine ou TENSIONING DEVICE sernnc eour Fig. ustration of the principles of Photo with schematic TOMETER-ISETEgrated Measuring Technique APPENDIX IT: Instruments of the Int CURVOMETER-ISETH: This device me ture of a straight at a given point {9}. Actually the change in the perpendicular distance between the intermediate point C of an arch unit and the base line AB (Fig. 24) is observed. The device sures indirectly the change in curva~ curved bea consists of a simply supported rigid frame (Fig. 43) of triangular form at whose middle point a mechanical dial gauge is fitted. The sensitivi- 001 mm an ty of the instrument is the accuracy of a reading obtai- nable in situ is about + 0.002 mm Pig. 43 The CURVOMETER equipped with a mechanical @ial gauge and the measuring bolts, ficl applicatio:- 40 - DEFORMETER-ISETH: This is a strain measuring device with a fixed base length to measure relative displacement AL = L - L' (Fig. 24). It consists of two parallel rods, which transmit the relative movements of bolts A and B to the mechanical dial gauge (Fig. 44). The on is less than + 0.005 mm. the measuril scatter of the readings in field applica’ Using linear displacement transducers (LVD?) together with a digital x both instruments the field readout unit and printing facilities works can considerably be speed ul Both instruments are provided with a suitable calibration device to check accuracy and proper functioning at any time. Fig. 44 The DEFORMETER equipped with a mechanical dial gauge and the measuring bolts, field application-4- APPENDIX IIT : SLIDING MICROMETER-ISETH This is a portable high precision instrument to measure differential displacements between two points in a borehole with a base length of 1.0 m {15}. Obviously one also could speak of a portable direct strain meter as well. Principle of measurement: A casing made of hard PVC, is installed in a borehole with a minimum diameter of $ = 75 mm, Measuring marks in the form of rings with special stop fittings are fixed at di- stances of 1.0 m apart in this tube (Fig. 45). The purpose of the mea~ suring marks is to accept temporarily the two heads of the instrument when readings are being made. If the marks move relative to each other than the change in distance between the marks is registered as the dif- ference of two readings. The measurement procedure is such that the in- strument is moved stepwise from one position to another and thereby set in the measuring marks. In each position i readings are taken. 1t may be seen from the schematic view (Fig. 46) that the principle of setting of the instrument into measuring position is based on the cone-sphere idea in order to achieve high accuracy. As either the cones nor the spheres are complete the instrument can be brought alternatively into sliding and into measuring position by a rotation of 45° (Fig. 46) .When setting the instrument into position of measurement the heads have to be pulled into the measuring marks while stretching a spring on the pro- tective sleeve. In this way the heads will be pressed to the measuring marks providing a stable condition for the readings. The exact orienta~ tion of the device with respect to rotation around its axis is facili- tated by a guide beam coupled with the forehead. The instrument is wa~ tertight up to an external water pressure of 15 bars so that casing does not have to be dry. A portable calibration frame constructed of invar steel serves to check the proper functioning and long-term stability of the instrument (Fig. 47). The device is designed to be self-compensa- ting with respect to temperature effects. However, to differentiate bet- ween induced-temperature strains and strains caused by stress or creep in concrete, rock or soil, one instrument head is fitted with a tempe- rature sensor. A possible "pile action" due to the stiffening effect of casing and grout in the borehole is prevented by a telescopic connection from the PVC tubes to the steel coupling elements. The accuracy of the setting under field condition is better than + 0.002 mm. MEASURING MARKS, BOREHOLE 1,0 m|1,0m| INSTRUMENT IN MEASURING POSITIONS Fig. 45 Principle of Sliding Micrometer-ISETH: Instrument positions defined by measuring marks at consecutive points along the boreholeINSTALLING MEASURING HEAD PROD MARK (CONE! Le or ee PROTECTIVE LOT ~ SLEEVE Fig. 46 Schematic view of the ng Micrometer~ISETH a) instrument in sliding position b) instrument in measuring position Fig. 47 The Sliding Micromeea REFERENCES {1} {2} {3} {4} is} {6} {7} {8} {9} {10} (1 ee oe {14} {15} Kovari, K.: Basic Considerations on the Design of Underground Openings, Periodica 3/1979 of the Int. Assoc. for Bridge and structural Engineering Golta, A.: Schwellvorgénge im Planum schweizerischer Bahntunnel, Rock Mechanics, Suppl. 5, 1976 Huder, J., Amberg, G.: Quellung in Mergel, Opalinuston und Anhy- drit, Schweizerische Bauzeitung, H. 43, 1970 Grob, H., Kovari, K. and Anstad, Ch.: Sources of Error in the Determination of in-Situ Stresses, Tectonophysics, Int. Journal of Geotechnics and the Geology and Physics of the Interior of the Earth, 29, 1 - 4 (1975) Kovari, K.: The Elasto-Plastic Analysis in the Design Practice of Underground Openings, Chapter 12 in Finite Elements in Geo- mechanics (ed. G. Gudehus), John Wiley, London Proceedings of Imigrantes Seminar, Revista Construgao Pesada No. 1481, June 1976, Sao Paulo Fiasco, J.: Problemas Diversos na Implantagao de Tuneis (Imigran- tes Seminar), Revista Construgao Pesada, No. 65, Juni 1976 Carvalho, 0.S., Kovari, K.: Displacement Measurements as a Means for Safe and Economical Tunnel Design, Proceedings of the Int. Symposium on Field Measurements in Rock Nechanics, Zurich, 1977, Balkema Publishers, Rotterdam Kovari, K., Amstad, Ch., Fritz, P.: Integrated Measuring Technique for Rock Pressure Determination, Proc. of the Int. Symposium on Field Measurements in Rock Mechanics, Zurich, 1977, Balkema Publishers, Rotterdam Bourquin, M.: Le plus long tunnel routier du monde en construc- tion sous le Gotthard, Chantiers No. 10, Saint Gotthard 1972 Golta, A.: Schwellvorgnge im Planum schweizerischer Bahntunnel, Rock Mechanics, Supplement 5, 1976 Kovari, K., Amstad, Ch. Field Instrumentation in Tunnelling as a Practical Design Aid, Proceedings of the 4th Congress of the Int. Soc. for Rock Mechanics, 1979, Balkema Publishers, Rotterdam Amstad, Ch., Koppel, J.: A Multihead Borehole Rod-Extensometer Design, Proc. of the Int. Symposium on Field Measurements in Rock Mechanics, Zurich, 1977, Balkema Publishers, Rotterdam Kovari, K., Amstad, Ch., Grob, H.: Displacement Measurements of High Accuracy in Underground Openings, Proc. of the 3rd Congress of the Int. Soc. for Rock Mechanics, Denver, 1974 Kovari, K., Amstad, Ch., Képpel, J.: New Developments in the Instrumentation of Underground Openings, Proc. of the 4th Rapid Excavation and Tunnelling Conference, Atlanta, USA, 1979