You are on page 1of 227
The Behaviour of Jacked Concrete Pipes during Site Installation by Paul Norris Thesis submitted for the degree of Doctor of Philosophy ‘at the University of Oxford Pembroke College Trinity Term 1992 ‘The Behaviour of Jacked Concrete Pipes during Site Installation aul Norris Pembroke College, University of Oxford, A thesis submitted forthe Degree of Doctor of Philosophy. Trinity Term, 1992 ABSTRAC’ ‘While mach money and effort has heen spent by manufacturers and usew of pipe jacking equipment to develop suitable techniques, this work appears tobe the Fisttostudy'the method at full scale, ina scientific research programme. It has involved monitoring a series of five pipe jacks during construction, In each case heavily instrumented pipe was incorporated into the pipe string to measure pipe joint stresses, pipe and joint compresions and contact stresses between pipe and ground, Taal jacking loads nd movements ofthe pipe string were also measured and all results corcated witha detailed site log, fll wane! ligament surveys, and observed groundconditons, The successof the site monitoring has beenbighly dependent "upon the development of suitable instrumentation anddata acquisition systemin conjunction ‘with appropriate site procedures for working inthe restricted and physically demanding pipe {ack environment without undue disruption to norma site operations. “The buildup of tua jacking force isthe result of highly complex soil-pipe interaction, The loca inteface stresses are essentially frictional most rund conditions, and can be related ‘othe shear strength ofthe ground. The problem isin determining the effective radial stresses ‘which are affected by soil inst stresses, stiffness and stength: groundwater conditions; ate of progress pipeline misalignment and use of lubricants. Relations between pressre distributions at pipe joints and measured tunnel alignments are presented. That small angular deviations between suecesive pipes case severe localisation of stresses on their ends i clearly demonstrated. Careful back analysis shows thatthe linear ress approach of the Concrete Pipe Assocation of Australia ean adequately match the measured stresses and could beusedby pipe manufacturers to provide designdataonallowable jacking forces for pipes on the basis of pipe ize, packer properties, concrete strength and angular alignment, It's also clear from the small pipe barrel stresses that improved packing ‘materials would allow more ofthe pemtial strength of pipes 1 be achieved ince relative angular rather than absolute deviations control transfer mechanisms between Pipes, uncritical adherence 10 specifications based on absolute line and level is counter-productve “When you can measure what you are speaking about, and express iin mumnbers, you kom something about it; ul wien you cannot measure i, when you cannot express iin mambers, ‘your knowledge is ofa meagre and unsatisfactory kind; itmay be the beginning of knowledge, >but you have scarcely, in your thoughts, advanced to the stage of science, whatever the matter may be.” Load Kelvin = Contents, ABSTRACT. Contents i Acknowledgements w Publications ¥ Nomenclature ‘i CHAPTER 1: INTRODUCTION M 1 Pipe jacking with concrete pipes andthe need for esenre). 1.1 12 Phase tv ofthe Oxford research project 13 CHAPTER 2: LITERATURE REVIEW 24 24 Ineodvetion 24 22 Concrete jacking pipes 23 23 Pipe jacking forces 27 2a Pipe end load tanster 22 25 Geotechnical aspees of pipe jacks Ras 26 Insrumentiton and monitoring of Felawork 219 264 Review of tunel, piling and pipe jacking instrumentation 2.19 projects 262 Planning and execution of tunnel fieldwork om active 2.21 CHAPTER 3: INSTRUMENTATION a 31 Introdvction| 31 32 “The instrumentation and measurement progeamme 34 33 Instrument design 35 3a ‘Contact ses ell 35 332 Pore pressure probe 340 333 Pipe joint pressure cell 313 33a “Tube extensometer Bas 335 Pipe joint movement indicator 316 336 Jack ram load cell a7 337 ‘Celesco unit sus 338 {Ground convergence indicator 3s. 339 Data acquisition and power supply 320 3394 Data acquisition 321 3392 Powersupply a2 3a Laboratory performance a4 CHAPTER 4 4 42 424 422 423 428 425 43 43 432 433 44 45 CHAPTER sa 52 521 522 523 524 53 Sa 8s 55a 552 533 CHAPTER 6: 61 62 621 622 623 624 63 631 632 6a 641 642 683 SITE SELECTION AND PROCEDURES Site selection for instrumentation purposes Details of the schemes Scheme I Scheme 2 Scheme 3 Scheme $ Scheme 5 Planning and exccution ofthe fickdwork Instrument instalation Instrument protection Instrument covery Data handing Performance ofthe instruments and site procedures PIPE END LOAD TRANSHER Ineo Direct en of pipe ks Change of pln agra price Elation ult angular election fom i eel Erfaton of gl etn aden of compression tem pint pp nontenieg Pip end gure ul Lind taster jan Predict ppe un blair Dan arrece Fermin ace seas Pecter mal proper Poston inj SOIL-PIPE INTERFACE BEHAVIOUR Irirodvetion| Basie data Jacking records Pipe jacking loads Local interface stresses mobilised during jacking Detailed response of individual pushes (Ground relsted actors Soil-ppe friction coeficients Pipe self weight tition ‘Construction relate factors Misalignment Time factor Lubrication 4a 4 47 47 410 412 416 416 421 423 428 430 431 433 su 3a 32 5.10 su su 519 525 520 529 529 530 6a 61 62 62 6.10 612 630 636 6.36 6.3 6.46 6.46, 650 6.56 (CHAPTER 7: a 72 73 7 75 75a 752 (CHAPTER & &1 82 83 a4 aS 86 87 REFERENCES PIPE STRESSES Inirodvetion| Pipe load paths Pipe barrel strains Pipe barrel stresses Elastic analysis of pipes Example 1: Scheme 5,chainage 111.6m Example 2: Scheme S,chainage 131.2 ‘CONCLUSIONS AND RECOMMENDATIONS, Instrumentation Site work Pipe joint behaviour Pipe barre behaviour Pipe-soll interaction Further work ‘Concluding remark a nm a 16 1 m2 m2 715 sa 81 x1 82 Ba 85. 8 810 Ra Acknowledgements “The nuthor is indebted to Dr G.W-E. Milligan for his much appreciated wuidance and Advice during the couse of this study; it has heen a pleasure working for such a stimulating and receptive engineer, ‘The financial support of SERC, the Pipe Jacking Association, an the five water companies, Northumbrian, North West, Severn Trent, Thames and Yorkstite is gratefully acknowledged, as are the contributions of the various contractors and pie manufacturers ‘ding the various schemes. Inthe development of the instrumentation, considerable assistance and support was provided by Mr Roy Smurthwaite, head of the department workshops nd his technicians Mr Ron Morton, electtonics technician, helped greatly withthe desi, manufacture, site installation and repair ofthe data collection system. The contct stress cells were designed jointly with Mr Clive Dalton of Cambridge Insitu Lid. Thanks are also dae «9 De Andrew Bond (formety of Imperial College) for prowiding details and advice on the pore pressure ‘robe design and the many commercial organisations who have supplied equipment the project and provided excellent after sales support ‘Special thanks go 1 Mr Chris Eggleton, esearch student, who provided invalua assistance on site and during routine calibration ofthe instruments, Miss Nica Andrews for part typing the thesis and Mes Judith Takacs for preparation of the drawings. The whole programme was only possible with the involvement af individuals ton num mention, in the management group, on the various sites, atthe pipe manufactuer’s works and throughout the University (Ona personal note I wish 1 express my gratitude o past and present members ofthe Soil Mechanics research group fr peoviding a stimulating work and social environment and Mott MacDonald for their support and leave of absence to pursue this projet, Finally, L wish to express sincere gratitude 9 my wife Andree, and son Matt providing encouragement, support and the necessary escape from the frstitions associated ‘with research, and o them [dedicate this thesis. ew for “The follawing conference papers and reports hive been published as reso this esearch, Norris, and Milian, G.W.E. (1991), et instrumentation For monitoring the performance of jucked concrete pipes. FMGM 9], Proc. 3. Int Symp. on Field Measurements in Geomechinies, sl, “Milligan, G.W-E and Nortis, . (1991). Conetete jacking pipes, the Onfordresearch project Proc. Ist Int Cont on pipe Jicking and Micstunneling, London, Nomis, P. (1992). Instrument design, manuliclre and calibration for use monitoring the field performance of jcked conerete pipes. Report No, OUEL, 1919872 Department of Enginvering Science, Oxford University Norris, P. and Milligan, G.W.E. (1992) Pipe end load transfer mechanisms during pipe jacking. roc. Int. Conf. on Trenchless Constucton, No-Dig 92, Washington Nowtis, Pind Milligan, G.W.E, (1992) Frictional resistance of jacked conerete pipes ull scale. Proc. Int. Cont. on Trenchless Construction, No-Dig 92, Pars Nomenclature Initial packing material thickness lstc contact width between pipe and soi Cohesion ofthe soi Undvined shea suength ‘Young's modulus of concrete Equivalent joint elasticity eoetiient ‘Young's modulus of packing material ‘Concrete characteristic cube strength Joint compression measured withthe joint movement indicators “Teeeaghi's coefficient of si oad Pipe length ‘Testa radial load on pipe ‘Tova radia invertace sess Effective rail inerice sess “Time required fr 959% response a step load Interface shea stess Dimensionless tunnel silty amber Weight of pipe per meee tun Diametrca contact width at pipe join Radial angle to the point of maximum compression “adhesion” between pipe and clay Angular defection a pipe joints Pipe end squateness angle a socket end Pipe end squireness ange at spigot end ‘Unit weight of the sil Unit weight of water Angle of skin friction between pipe and sil in tt stess terms) ‘Angle of skin friction between pipe and sol in effective stess erm) Packing material compression Pipe deformation Pipe longitadial stain Coefficient of friction between pipe and rock Deviation angle between pipe iavert and pint of contact in ck Pipe anginal stress Hoiontt sui sess Maximum stress at pipe joint Sot stress for uniform hod Rail si stress Required tunnel support pressure Vertical si stress Undrained angle of internal frit of he sol rained angle of internal ection of the soi chap aie CHAPTER 1: INTRODUCTION 1.1 Pipe Jacking with concrete pipes and the need for research Pipe jacking isa technique for forming small diameter tunnels by pushing o jacking Pipes though the ground fom a thst pit toa receiving pit Pipes ae advanced using hydraulic power packs located in the thrst pit a the ground in fot cf the pipeline is rmined. Excavation is noaally carted out within shield sing either proumatic tos ‘tunnel boring machine, wth the spoil being transported along the pipeline to the surface. Steering and adjustments for line and level are mide atthe shield using jacks, in ‘conjunction with frequent sun ying to fied reference poins. The tchaigue is depicted in Figure 11. Pipe jacking is suited to short lengths of tunnel typcily les thin 150m between jacking points and is often used for sections of pipe under embankments, roads and ‘ilways, where open cut methods would be particularly uneconomic because of the need to keep tnffie moving continuously. Likewise there have heen many occasions when the smethod hasbeen used for small sections of long tunnels tht pass below bulings. One ofthe main advantages of using pipe jacking, particularly under existing structure, i hat 1 rigid liner is provided immediately after excavation, in the shape of the pipe, with relatively title over excavation (ypicaly less than 20mm around the circumference) “There is now a tendency for longer and lage contrats to be undertaken wsing pipe jacking as an altemative o open cut method either for diveet economy or to reduce disturbances atthe surface. Wallis (1982) reports an 800mm diameter pipeack, 460m Jong in Londoa clay. Winfield (1986) records 690m of 1950mm diameter pipes being uw jacked through sandstones and siltstone, Perhaps one of the most impressive recent ‘examples isthe jacking of a 14.4m wide by 9.3m high by 35m long precast concrete box. culvert through a clay embankment supporting 2 four tack railway, New Civil Engineer (1989), For the neger diameter pipes and long lengths of thrust, either intermediate jacking, stations, which ate built into the lning ad subsequently removed, or lubrication, usualy bentonite, may be necessary to reduce the jacking forces, Durden (1982). The insertion of polythene sheeting between the ground and the pipetine has aso been wsed successfully for Similar purpose, Toyama (1985) Although the method has many advantages over conventional method of tunneling its wider use i being held hack by a lack of understanding of many favors affecting the installation and performance of such tunnels, and loss of confidence by some specifying authorities resulting from unexpected failures, Shullock (1982). ‘The nee for research in several areas was fist reported by Kirkland (1982) and then in a CIRIA report on the state ofthe att in pipe jacking (Craig °983). Industry, represented by the Pipe Jacking Association (PIA), has actively promoted and supported the ease for research, which resulted inthe Fits stage of the Oxford pipelacking research; laboratory based study of model pipes starting in February 1986 an finishing thre years ee ee ee indngs fom this work emphasised the need for suitable picking material in joins along with carefol contol of pipe alignment, and the superiority of steel banded butt joins to inwall spigot and socket joints forthe transmission of large jacking fees. 1.2 Phase two of the Oxford research project In April 1989, a second phase ofthe research was iitsted to confiem under actual 1 conditions the findings from the laboratory work and investignte aspects of 13 Pooce oss I Termes Fates = ot | _=_—- Pit cs Speci Pps Addon Tes " Desles Aniynis Quaeriy Repre ee Figure 1.2. Research progea Performance thit can only be studied realistically at full scale, This phase involved incorporating an instrumented pipe int pipejacked tunnels during consiracion on five sits. ‘The work was supported hy the Science and Engineering Research Coancil (SERC) in collaboration withthe PIA and five water service companies, Northumbrian, North West, Severn Trent, Thames and Yorkshire. Close coopeation hetween clients, contactors and researchers has been essential 40 the success of the projec. 14 The main uncertainties in pipe jacking arse hecause the alignment of pipes can never be pertzet. The lads between pipes are not transmitted uniformly, and the Inerution between soil and pipe is such that frictional fore resisting the forward ‘movement of the pipe string may be greatly increased. These two effects interaet with each other to increase the jacking loads and cause stress conce ions inthe >ipes. The main purpose ofthe research is therefore w investigate the load transfer between pipes and the contact pressures etween pipes and soil, The measurement are relate to deste Hine and level surveys of the pipetine and to Inca ground conditions. The aim st allow beter prediction and contol of pipe jacking operations inthe future, “The progres ofthe overall programme of esearch is presente in Figure 1.2. ‘The fist year saw completion ofthe design, manufacture and calibration of the instrumentation Minor contractual delays meant that the frst drive was about three months late in siting, With the next year and a half require for the fckdwork on the five pipe jacks. To allow adequate time for analysis and writing up of the work the thee year programme was extended by four months “The eontmet cost for phase wo was £219,500 a8 shown in Tale 1.1 5 SERC ‘atred about 208 of the total, the PIA about 40%, and the remainder was divided equally hetween the five water companies. In addition, exch site operation involved additonal ‘oss of between £10,008 and £20,000, mainly Kor modification ofa stand pine to incorporate the instruments, provision ofa lne inside she instrumented pipe to protect the instruments, some delay and loss of pandcton by the contactor, and cans of retieving the instruments atthe end of the operation, By the end of the contact, each water ‘company had citric these ational costs for one site “The programme of rescafch was overssen by a management group with tw) representatives trom the PJA, one from Oxford, and one from each ofthe water companies. The group met four times a year to discuss progress. The funds were administered via 1s Reseach ast ading apes ‘Onford con incon oserhends (pervon tchicin s¥ppon, ‘epmamphi, nd secre sin). coo Instrumentation an dt loging em ‘TOTAL 219500 Pp Icing Asoc 000 Winter Aarts (Companies) $x 1,000 « sng TOTAL 319500 ‘Table L Costs and Funding of research work. special account with Thames Water, apart from the SERC contibution which was used to support the author directly through the University, Financially, the projet, including the extension, was completed within the original budget This report on the site aed research begins with Chapter 2 which provides a Teackeround tthe pipe joking method of construction, its uses and limitations. jontinues 40 present the curtent methods for determining the magnitude of pipe jcking forces in different ground conditions and the effect of pipeline misalignment om kod transfer a pipe joints. A review of previous site hased tunel, pling nd pipe jacking instrumentation projets is then presented which highlights the necessity of minimising contctual pressures y thorough planning and clear communication at all stages of the work (Chapter 3 sts out the overall concept of the instrumentation and presens details of the design, manufacture and calibration phases. All instruments have been designed 10 eng in ‘operate successfully in the tunnel environment, have minimal effect on the property 4 be measured, be sufficiently accurate, be simple 40 calibrate and to distupt normal site ‘operations a8 Hitle a8 posible. The selection of sites and developmen: of suitable site procedures included as Chapter 4, which also provides an assessment ofthe effectiveness of the site operations. “The ress of the Fework ate covered in Chapters S10 7. Chaater S deals with pipe end tou transfer mechanisms during jacking. Relationships fetween pressure istebatons at pipe joins and measured tunnel alignments are presented and the effectiveness of packing material and the importance of gond conte of line and level assessed Measurements of toa jacking resistance and localised contact pessures between pipes and soil are included as Chapter 6. Detailed reseatch objectives of the soi-pipe {interaction data include determination ofthe areas of contact between the pipe and ground in both granular and cohesive soils, the magnitude of the normal and shear stresses developed at these contact and how these relate wo the overall pipeline resistance. The effects oF stoppages and lbrication on both the local and global measurements ate evaluate. “The resulting stresses ad stains in individual pipes are the subject of Chapter 7, while the conclusions ftom the research and recommendations for future work ate awn together in Chapter 8, w CHAPTER 2: ERATURE REVIEW 21 Introduction Tunnels constructed in the UK may he grouped into three size categories + Small diameter = up t» 3m internal diameter, for sewer, water and cable tunnels + Medium diameter - 3.6m internal diameter, for underground rilways nd associated tunnel + Lage diameter ~ 6m upwards internal diameter, for wad and main line railway tunnels, underground chambers All tunes, except in sound unjoined tock, where the ground is ef supporting, are lined witha primary lining which i designed wo suppor the ground loads and temporary Joads that may occur during installation and for the design life ofthe stucture. Such linings should als exclude or conto the ingress of water ito the tual Pie Jacking isa tecnique for providing a gid primary fining inthe form of a Pipe which eliminates the need for temporary ground support or secondly tnings. I is Principally wed for small diameter tunnels in soft ground. A summary of the conditions in which pipe jacking is « competitive allemative to other forms of primary lining is included a8 Table 2.1 Craig and Muir Wood (1978) “The east ecorded use ofthe Pipe Jacking method was in America about 1910, Richardson and Mayo (1941). The basic principles ofthe technique have heen presented in del by the American Concrete Pipe Association (1960), Richardson (1970), Hough (1974), Drennon (1979) and Clarkson and Thomson (1983). The Britsh Pipe Jacking ‘Association published) noes giving guidance on design and practice (Pipe Jacking Ass jon 1981 and 1986). The increasing use of modern technology and a competitive 24 che (8161) Poo, MY pue Byes) sou SayUH Jo spoqom pareHIng 17 GRY, 22 tendering maicket have led to many innovative methods being introduced in recent years Richardson and Seruby (1981) reported the development ofthe Usi-Tunnel system in which pipes ate jacked forward by inflatable bladders positioned between successive pipes. A new joiting. profile is reported hy Cole (1986) and White etal (1988) presented details of the use of new joint picking materials and microtmnel technology wsing tumrenforced concrete pipes. More diverse apliatons of the technique include a novel tunnel forming method using multiple pipejacks wo form an areh root 4 an underground sation in Milan, New Civil Engineer (1990) and the replacement ofa single brick arch a bridge by a series of precast hox tunnels, Ghosh and Madhusulhan (1991). 222 Concrete jacking pipes Precast concrete jacking pipes are generally cast in manufacters’ works and transported to site by toad, Good quality contol and adequate precautions fo the curing of the concrete and regular checking of the moulds can pratuce sceuraely cast pipes of consistent concrete strength. The pipes are manufactured by centrifugal spinning or vertical casting using concrete with a 28 day charictristc cube stength greater than 60 N/mm Spirally wound reinforcement spot welded o fongitdinal tel to form intemal and exter cages ate often used to prevent damage during the temporaty handling and instalation sages Brvsh pipe manufacturers currently assess compliance of a pis’s stength to ‘withstand ground loading imposed on an installed pipe. Class H pipes ae sed fr jacking and a designed to meet Marston (1930) lang conditions. Tey are designed to higher standands and stengts if ground conditions o lent specifications impose more onerous design equiements. Allowance for instalation loads receives seat tention, possibly due 10 the uncertainty in predicting leads, stresses and thee dsthation throughout the jacking cyele, The lack of data required to establish a suitable test assess pipe’ ability to 23 sustain installation end loads is acknowledged in the recent British Standawd, BSS911: Part 120 (1989). "The forward to the standard states “The joint face test is included in the ansence of suitable jacking strength test. Research is presently heing undertaken inorder 1 devise such a test and this may be incorporated into a future revision’ Figure 2.1, Arrangement for joint face test (BSS911: Part 120: 1989), ‘The drafting commitee setled on the joint fice stength test as 2 way of determining the "quality" of the concrete in the jacking face. The testis ilustated in Figure 21 and consists of loading areas at both end of pipe hough a 10am x 20mm 1 20mm see plate. ‘The average uhimate strength of the concrete Face must aot be less than 100 Nim? and no individual result ean fll below 70 Nimm?, ‘The standard points ‘ou that he resulting strengths should not be used to calculate permissible jacking fads For the purposes of instalation i is usu 1 assume that ideal jacking conditions ‘exis in whieh the pipes are subjected w an evenly distributed axial load around the pipe ciroumference, The worst ese is assumed tobe inthe pipe nearest the trast pit which will be subjected to the maximum jacking force from the rams. tn + typical design ‘example for an 180mm intemal diameter pipe, jacking toad of 1000 tomes produces an average direct compression on a fll pipe section of 9Nimm?. British pipe manufacturers specity ther pipes as being capable of sustining a wniformly distributed end stress of ‘between 10 and 15 Ninn? (dependent upon manufacturer and pipe diameter). ‘These Figures appear conservative if toads are applied to the fll end area ofthe pipe, but allow for end tolerances and jacking misalignment in an abitary fshion. tis one of the Principal aims of the Oxford research to enable manufacturers and contractors 40 offer liens theoretically based method for assessing the axial load capacity of jacking pipes. Precast concrete is telaively impermeable material and theretine problems with leakage of water will normally be at pipe joints or at any cracks in the conerete. The functional requirements of «joint on a jacked pipe are adapted from Clarke (1968) + It should he designed to permit angular and axial movements large enough 40 tolerate the maximum displacements likely to accu, without damage or loss of| water tightness + Teshould withstand the Force applied during instalation without damage + Te should remain efficient throughout its working life + Teshoutd be simple 1 make and dismantle inthe limited space of the thrust pit “Traditionally, jacked pie joints i the UK have been the in wal spigot and socket type as shown in igure 2.21, Disguict in the industry about the performance of the in wall 2s Tixreanacoumerer | 42) INWALL SPIGOT AND SOGKET JOINT. ZZ pe ae tel | An eel STEELCOLLAR Sb aR \ : RUBBER SEAL a JOINT PACKING MATERIAL b) STEEL COLLARJOIT Figure 22 Common UK pipe joints joint and is ability to transmit longitudinal loads efficiently bas led to the intoduction of the steel collar joint, Figure 2.2. The main reason for the new joint detail is the belief that jacking loads are beter trnsmited through the cenee of the pipe wal rather than 26 her iene Roser long its edges. The CIRIA technical reports 112 (1983) and 127 (1987) provide further euits of these joins and examples of other less common joining arrangements. ‘The functional requirements of he joints ae in part contradictory, The flexibility ‘offered bythe joint enables the pipeline o accommodate future ground movements without imposing substantial stresses into the pipes, while conversely misaligamen: of pipes relative to cach other during the jacking process can lead to local stress concentrations in both the {ground and joins. These have wationally been minimised by seeking 10 lit the maximum eros in line and level io (piclly 275mm. Compressible joint packing ‘materials are recommended by all pipe manufacturers as an aid to distributing end stresses ‘over lager areas during jacking. However, not all are prepared to specify the type of rateril or advise onthe details ofits use. This aspect has not been coveted in the recent 1BSSO1] Part 120, 23 Pipe jacking forces Accurate preitions ofthe Force required to jack a pipeline are notonlystuctrally important hut also enable the contactor to ease tit Bis equipment and site set up ‘including provision of ineracks and thrust wall arangements) are capadte of producing the required trast complet the dive without damage to pipes and joins rom exoessive stress concentrations. ‘The ttl jacking load depends upon both the Force required to push the shield into the excavation, commonly refered 0 a face resistance, and the fictional resistance along the pipe length, “The amount of resistance encountered a the face depends upon ground cond and the measures required to support the fae. In hand drives it predeminantly related to the edge cutting resistance ofthe jacking shield and overall size & important as larger {imensions proce langer face resistances. Typical estimates vary from 100 to 400 tonnes, fr diameters From 900mm to 2850mm, Auld (1982). In machine drives the face pressure 27 required to suppor the ground must he taken ito account. Currently UK pipe jacking contactors use empirical methods t predict fection forces. These express the required shit in terms of fictional resistance per unit surface sea of pipe. Typical values of ction resistance fr diferent ground conditions ae given in Table 22 Ground Type External oad (KPa) Rock 23 Bolder clay 518 Firm elay 5.20 Wer sand was sit 5.20 Dry loose sand 25-45 il os ‘Table 2.2. Typical values of friction load om pipes (After Craig (1983). 1 will be appreciated that the method is no more than an “experienced guess” hich ean lead to problems with pipe fulures and abandoned boring machines. Injection ‘of 8 lubricant such 4s bentonite slurry into the overbreak etn result in reduced frictional resistance, Research curied out in Japan indicates that a reduction of 365085 for clayey soils and about 209 for sandy soils can be achieved (Ishibashi, 1988). However no temps have been made to standardise the use of lubricants In ordce 19 provide « comprehensive mode! for jacking force predictions it is necessary to review the parameter which inflence the jacking ls, These mast he rationalised in relation wo theoretical methods for evaluating ground pssures on buried ager rt sine “The principal factors include + Primary load from the ground including surcharge, transient live buds and internal and external Mud pressures + Resistance atthe exeavation face + Amount of overeu during excavation + Variation in ground conditions slong the pipeline + Misaignment of pipes + Injection of tbrican int the overbreak void + Use of inermestiate jacking stations ‘Auld (1982), Haslem (1986), O'Reilly and Rogers (1987), Ripley (1989) have proposed various approaches to predicting jacking forces hased on the analysis of eld data from selected pipejacks. Auld presents an analysis fora pipe driven through a cobesioness material for Which the sil is assumed to collapse onto the pipe and exert ral pressure around its circumference. ‘The method is notably based on work by Terzaghi (1943) in. which the analysis of pressure distribution on rectangular buried structures in active flue zones is suitably modified for the eieulr pipe jack situition, ‘The meth takes account of arching in the soil above the pipe redisbuting ambient sitess away rom the pipe. The fictional resistance is then determined by using evefisiens of friction in eonjnetion with the caeulated ground pressure. Unfortunately Avld fails to resolve the horizontal and vertical stresses intra comet rail stress component resulting in a 25% overestimate of jacking resistince. Evaluation of suitbletrieton cveftciens fs abo difficult to define accurately and can be further modified hy the wse of lubricants, Ripley (1989) states that pipejnckers usually adops value of 0.7 times the inermal angle of friction in cohesionless ground, however the value is likely w he stress level dependent, ‘The ability of an unlined tunnel 40 support ise in non-cohesive and cohesive ground hasbeen stati using laboratory and cenit models, Atkinson, Brown and Pows (1975) Davies, Gunn, Mair and Senevirane (1980) and Atkinson and Mair (1981). In non-cohesive ground it was found thatthe support pressure 0, was independent ‘of tunnel depth, Collapse was initiated by a small triangular wedge of material loosening atthe crown ofthe tunnel. For the case with no surcharge loading r umber and is only a function of ¢. (Figute 2.3) is the diameter of the bore. 4 isthe density ofthe sei Te water pressure is presen itis aded t9 a, and (yoy, i used for that part of the Soil below the water tbe o 10 Eng 30" we Figure 23 Relationship between tunnel stability umber T, and érained angle of internal friction @. In cohesive ground, stability is dependent upon the unsupported length of tunnel advance P. For 1wo dimensional tunnels (ie) PID large) and a, = (the short term 210 Stabilly of a tunnel ean be assessed from the relationship between BM and C1D in Figure 24 Figure 24 Criterion for assessment of the circular plane seetion tunnel in clay. of an unsupported Haslem studied the behaviour of pipe jacks in London Clay, with stable bores. The model he adopted was based onthe behaviour ofan elastic eyinder resting i a eplindrcal ‘oid in an elastic continuum, Roark and Young (1976). The shear surtice fr the cohesive ‘material was assumed 10 be within the soil, ther than at the pipe Soil interface. ‘Comparisons made hetween the jacking frees and those mesure in actual pipe jacks ‘were found w provide an underestimate ofthe wt cking force, Adjustments were made ment of pipes hy considering an element of pipe length subject co a local radius of curvature, but this was found to hive negligible effec. (O'Reilly and Rogers (1987) used the ime technique to compare predicted and field measurements in clay and rock (sandstone). Laboratory ests eared out By the authors to assess the contact atea tween pipes and reconstituted cay indicated tha time dependent plastic deformations occur under the pipes resulting in larger contact areas. Replosting the dat published by Haslem and increasing the predicted values in accordance with the ‘experimental data produced a better fit, The analyses for rack were hased on ei + single ‘or multiple point contact relative tothe pipe invert. The model produced a good fit 0 the available data provided the site conditions were relatively uniform. ‘A summary of the models is presented in Figure 2.5. All are based on the assumptions of perfectly staight pipetne and plane stain conditions fr the fining and the round. This is never he case in practi, Some degree of misalignment is inevitable and this will introduce a series of forces normal to the line ofthe drive in-easing the total Jacking foree. The types of force resulting from misalignment, assuming elastic ‘deformations, are shown schematically in Figure 2.6. Stevens (1989) and Ripley (1989) ‘have conducted laboratory based test on jacked pipes main nein deft positions 8 the axial load is increased. Examination of the induced toads in the lateral support systems provides an indication ofthe resistance the ground around a pipe jack must provide if realignment is prevented. The data substantiate the fact that less interaction occurs between pipe and eestaint with smaller deflection angles. The work slbo enabled the resuling stain distribution throughout differently missligned pipes to be recorded and associated failure modes. 9 be observed. 24 Pipe end load transfer ‘The problems associated with misalignment, which generally masifest themselves 4s local spalling of concrete at the joins, hive led to the introduction of tow stitfness packing material between the joints to help distibute the axial foree. Ripley (1989) has tested the most commonly used wood based packing materials obtaining ifoxmation on the effects of angular deviations and packer positioning i the joint. A sumary ofthe resus 212 Litt Reaction Reaction Figure 2.5 Summary of ground loading models. 213 Non Cohesive Auld (1982 o tee net ny Petun 0. P, where @ isthe angle of internal friction ofthe soil 2 is the angle of friction between the pipe and the soil Cohesive with a stable bore Haslem (7986 Feasb where cis, is the “adhesion! hetween the pipe and clay. is the contact width ck 0" oes (1987) F = Wyicos © where W isthe weight of pipe per unit length bis the coefticlent of eto, Pipe | ipe2 Pipe 3 Pipes Pipe 5. a) Misaligned Pipeline “neers oa Tee pec [tr eS en E* sf ts Ib} Farce System to Maintain Equilibrium Figure 2.6 Misalignment forces. suggests that + Thick packing materials are most eneficial, with 18mm hick medium density fibreboard the most suitable from the matriis teste, + Deflection angles greater than 012" combined with high axial loads are 40 be avoided, + Wet materials improve load transfer capabilites by up to hee times + yelic loading conditions need 1 be considered due ta permanent packing material compression, “The significance ofthe packing material prompted Ripley to propose tit pipes shouldbe dived site with « standardised packer already in pace. ‘The henelits of packing material in reducing the effects of axial stress concentrations in vtitfed clay jacking pipes has been established by Boot and Husein (1991), The eonctusons fom this work were similar to those of Ripley. Various atemprs have been made to theoestially evaluate the induced stresses in the concrete as function of packing material stiffness and geometry, Milligan and Ripley (1989), Nagata etal (1990) and the Concrete Pipe Association of Australia (1983). ‘The fist two methods assume linea itain and non-inearstess characteristics actos the joint while the thir simplifies the mod! hy adoping 4 inca stress approich silted in Figure 2.7, All thw analyses il take account of the misalignment shee forces which re generated at pipe joins. 2s enger ae aie “The analysis presented by the Australian CPA is summarised below: ‘Total packer and pipe deformation Baa = A+ AL Deformations related to stresses Se Th yok EE CE Joint Packer Concrete bout r= Sh where t= thickness and f= wall thickness at joint Ee therefore af. the corresponding joint elasticity coetfcient Eis given by £, = Mess _ poning 8 " ene! aE + UE, “The problem reduces to the stress disteution in an annular eross-sectioe where the tensile sresses are disteganded, Marks (1978), From the diagram p= wn'82 = tan XS Rearanging Zum bles, Macks (1978) food Figure 27. Australian joint stress distribution 216 age see Re Misalignment between consecutive pipes can also result fom a ack of pipe end ‘squureness. A test bused on the measurement of joint gaps before and afer one pipe is rotated through 180 degrees is incorporated into BS5911. An alternative method, proposed forthe CEN standard on jacking pipes, involves measuring the genertrices s,s, and the diagonals r, and as shown in Figure 2.8. “The squareness of ends atthe spigot, €, and socket, em, are determined by ese (52 +r? = 52 = rn) em = (5) = Ff = 7 + F240) Where In is the internal bate length, Figure 28 Measurements of lengths. should be noted that pipes complying with nd squateness tolerances of BSS91 I Part 120 could have a joint gap of up to 8mm on @ 900mm internal diameter pipe, ‘This ‘equates to deflection of 0.38 degrees even before pipes are deflected relative to each ‘thes, It is therefore importnt to be aware of the end squareness of pies. 217 25 Geotechnical axpects of pipejacks Deuiled investigations of the subsurtice conditions are requied for all tunnelling projects. The available date must be carefully evaluated to determine construction methods ‘hat will avoid damage to existing surface structures or underground wiles. Factors such 4s high groundwater tables, unstable soil, and the presence of boulders of contaminated ground, if not recognised prior to construction, can make completion of a pipeline much more difficult. Important physical properties which should be determined include the sength, gran size, plastic characteristics and permeability of the ground, and their ‘arabiliy along the pipeline. Typical spacing of exploratory boreholes should be about ‘every 5Om slong the alignment depending upon the degre of complexity ofthe geological ‘conditions. ‘The purpose of these borings is not only 19 determine the conditions at eaeh lcation but also to Uevelop a subsurface profile which highlights houndares between the various soils. The main dificult arising fom changes in ground conditions is when a fault or soft 4 8 unexpectedly encountered at the fave leading t difficulties in maintaining shield alignment. Groundwater also tis an important influence on the behaviour of the ground. Groundwater levels should be determined in the borings and hydraulic conductivity evaluated, [An important factor relate to soil conditions that significantly influences project Feasibility, planning and construction costs isthe potential and magnitude of surface and subsurface setlement. Surface seilement is usually considered the primary source of lamage t roads lites and shallow foundations. In pipe jacking surfice setlement is generally associated with loss of ground dv t soil squeezing or runnin into the heading and overcut on the shield. Empirical methods have been developed for estimating surface settlement duet sft ground tunneling by study of observed setlements on past projects, ‘OReilly and New (1982), O'Reilly (1988). A laboratory based study concerned with ‘evaluating the mechanisms controlling soil deformations around pipejicked tunnels is being 2a cried out at Loughborough Univesity, Chapman (1992), Finally, the resistance that ean be mobilised by the thrust wall is» function of the soil strength. It would seem appropriate to allow the thrust wall to develop the full maximum passive resistance because is a temporary construction Inading situation However the large stains required to develop fll passive resistance can fed significant problems with non-uniform thrust wall movement resulting in eccentric aie loading and pipe damage, Cole (1977), Movement ofthe thrust wall can be contoed by limiting the lowable passive resistance o about one half of the maximum value although the effects of repeated load app cation on the ground need further investigation 26 Instrumentation and monitoring of field work 26.1 Review of tunnel, piling and pipe jacking instrumentation projects ‘The instrumentation and monitoring of tunnel linings and the ground in the vicinity ‘of a tunnel are carted out t9 provide & etter understanding of the mechanisms which affect lining and ground behaviour, during and after tunnel eonstrction, ‘The results of such monitoring help t0 improve the design methods for tunnel lings and enable Indications of the likely surface settlements be established, ad thus th risk of damage to structures tobe asesse Instrumentation programmes for tunnel projects have been carried out for many years, Cording eta (1975) presenta detailed state of the at epor for instrumentation of tunes in soi and rocks. Sohn and Crighton (1989) deserie the geotechnical contol and monitoring being cartied out forthe Channel Tunnel projet. Much ofthis instrumentation i limited 19 monitoring fom within he tunnel using instrumentation stations that remain stationary with respect the tunnel dive. OF more relevance t pipe jacking isthe work that has been carved out om instrumented pies to improve our understanding of pile-structoe interaction, Yong (193) presents an 29 interesting overview of the practical aspects of instrumentation used 10 monitor twelve bored east-insitu piles and two displacement ples, ‘The instrumentation concentrated on determining load distibution characteristics using mechanical extesometes, vibrating wire strain gauges, resistance strain gauges and custom made lad cells. A ccmparatve study of thee performance indicated that the mechanical devices were less prone to flue of damage than the electrical devices, This substantiates earlier work by Lene (1982) who reported from stat a findings thatthe casualty rate for vibrating wire tain gauges is tess than 10% while Reese and Hudson (2968) indicate tht fires in esistance strain sauges can be as high as SOF. The major problems experienced with te stain gauged devices was due to moisture penetration. Coop (1987), Poniah (1989) and Bond and Jardine (1989) report on the peformance of instrumented model piles used to investigate the fundamental behaviour of ‘riven open ended ste piles used in offshore stuctures in clay soil, Measurements of axial loud at diffrent positions along the pile length, radial toa pressure and pore water pressures were reported, The Bond and Jardine pile also included surface contact wransdcers capable of measuring rl total sess an shear stresses specific lcatons cn the pile surface. ‘Such heavily instrumented schemes contrast sharply with the lintel published data from pipe jacking sites inthe UK which ted to be confined 10 tne, level ad jacking pressure wings on mechanised drives only. Overseas the situation is lsh beter with schemes such asthe New York, San Island sewer upgrade reported in Ci Engineering (1988) proving £11 milion allowance inthe contact fr instrumentation, The scheme involved pipe jacking 21Skm of moulded fibreglass pipe through sands und clays. Several pipes were fitted with vibrating wire stn gnuges and surface mounted extensometers to measure longitudinal and cicumfeential stresses. Load cells were pstioned in the jacking system. Subsurface settlement indicators have also been installed aver the length 220 fof the project. Discussions withthe projet engineer, Beloff (1988), confirmed thatthe instruments had performed satisfactorily and indicated relatively uniform eompression of| pipes and the presence of lacked in steses afer jacking pressure tees. No problems With high foal stresses at joins were recorded. Unfortunately data from the project has rot been published. 262 I ming and execution of tunnel eldwork on active construction sites Instruments shouldbe designed 1 be easily installed and if possible, fxed or cast, imo the lining before its incorporated into the permanent works this minimising delays in the tunneling programme, and, where possible, eadings shouldbe taken cemotely from the location of the instruments “The instruments should he relatively cheap «0 enable sufficient numbers to be installed, All instruments shouldbe reliable and constructed to withstand the envionment in which they will be installed. This environment in a pipe jacking tunnel is likely to be ‘vet, dity and very confined, They shouldbe sufficiently robust withstand th curiosity of the tunnel tabour fore and covered to withstand shock. The instruments should be lesigned forthe required accuracy and be simple to calibrate. Instrument installed inthe lining should be designed to cause as rmodifiation tothe structural peeformance of the lining. 38 posible, otherwise they will he measuring conditions ditfereat to those of the standard ining The readings obtained from the instruments should be analysed as the work proceeds, enabling ertors in the instruments or the reading techniques tobe found during the course of the work rather than at later date, ‘The output should be designed to suit the results required forthe final objectives and not obscured by a large quantity of data, that may never he used. All ofthese objectives have been incorporate int the design of the instrumentation scheme forthe pipe jacking research reported in Chaster 3 gar ieee Te willbe apparent shat instrumentation programmes for tunnels are normally cated ‘out during the construction phase of the project. The use of instruments & not merely the selection of suitable devices but a comprehensive step-by-step engineering. process beginning with a definition of the objective and ending with interpretation of the dats ch sep fs critical w the successor failure ofthe entre programme ard none more sO thao the planing and implementation of fieldwork on ative construction sites w minimise the contractual pressures that inevitably arise, This particular aspect isthe subject of a forthcoming code of practice which is being drafted by the SERC Geotechoies Steering Group. Draft guidance notes ist the following key element for successful field instrumentation research Planning Phase + Maximise lead in time ~ be involved wih the design + Have clear and realise objectives + Understand roles and motivations of putes involved + Work with consutans for proper provisions in contrat documents + Explain aims of esearch t the Contractor and encourage his intrest + Use expertise of experienced groups such as TRL and BRE Fieldwork Phase + Tey to install instruments when the site fs not busy + Make research staff part of the Resident Engineer's team + Arrange fora member of the eonteator's sta to be personaly responsible + Communicate clealy = beware of wnbrifed subcontractors + Be prepared for delays and disruptions + Process data promptly and feed back wo sie staff 222 CHAPTER 3: INSTRUMENTATION 3M Introduction This chapter presents dels of the instrumentation developed for the research programme. Specific instrument designs ae described and calibration procedures and resuls discussed, The development ofthe instrumentation has been heaw ly influenced by Hannah (1985) and Dunnic fe (1988). A full description of the design, manufacture and calibration phases including design drawings, supplier detis and individual ation data Js presented in Nomis (1992). 3.2. The instrumentation and measurement programme. The field instrumentation consists of three instrument clusters or monitoring stations 1s shown in Figure 3.1. Two stations are incorporated int specially prepared "standart eonetete pipes which cam be inserted at any positon ina ain pipes, while the tind prstoned at the jacking pit. The lead instrumented pipe s only used in deives through cohesive material and is fited witha Ground Convergence tncator whch measures the rate of ground closure above the font end ofthe pipe sting particularly when jacking is halted overnight or at weekends. The mai nsrumente pipe located further back in the Pipe sting and contains the Following istumsnis (2) Four Contact Stress Cel to measure both sail total stress ad shear stress acting on the pipe surface, with their active face flush withthe surface ofthe pipe and provided with uM Figure 3.1. Schemati of instrument arrangement. simile surface finish (b) Pour pore pressure probes adjacent to the contact stess cells, measuring the loa pore pressure and hence allowing determination ofthe effective radial sess; (6) Thre joint movement indicators a each end of the pipe, f0 measure the movements cro the joint gaps in three dimensions; (@) Twelve pressure cell bul nts the packer in the joint at each end of the pipe 40 measure the magnitude and distribution of the loads teansfereed across th joints (6) Six tube extensometer fied othe inner surface of the pipe and eys-spaced around it w measure the compression ofthe pipe under load and (9. A modular data acquisition system and purpose built stable power sippy In the jacking pit the total jacking lod is recorded by Iwo oF four loud cells pesitioned ‘otwcen the jac rims and the thst rng, and the forward movement of He pipe sting hy 4 displacement ransdicer mounted above the tunnel entrance, Resdings ftom the instruments ate backed up by a detailed log of al site activities, a daily srvey of line and 32 Onur level of the full length of the tunnel, and addtional observation, sampling and testing of the ground conditions to supplement site investigation dita. Pipe dimensional checks, concrete strength and stifness testing, and where appropriate surface setlement surveys have also been eartied out. A flow chart summarising the objectives ofthe instrumentation is shown in Figure 3.2. Loci ail and ‘Avcrnge sees oer sage pipe eng Average fe an sbeat Pipe wemtle ound Inne te expressed Ineffective see terms. Ground converge ner tnd pipes insTRUMENT, CONTACT STRESS JOINT PRESSURE ceLis' Ack LOAD ceLLs [AND PRESSURE CELLS Tue EXTENSOMETERS PORE PRESSURE ‘oRouND CONVERGENCE INDICATOR PIPE JOINT MOVEMENT INDICATORS IPE PERFORMANCE ‘ees pp pss wo ven pipe cempresion ‘sie jot me provi eontipe between $Heaet a pt Figure 3.2. Flowchart indicating objectives ofeach instrument type. Fey {An inventory ofthe equipment and cxsis at 1991 prices are presented in Table 3. im Descrton| 1) Hoong 1) Came wasscer 2 Pore Presa Pate (Gctding Dk PCRS) 3. Pipe Jot Pressure Cet 4. Tite Bncrsometes (ictaing LVDT) (acting LYDTS) singe dimension mode 1) ee dimension model octane ap) 2. Celesen Une 1 0-100" 0s" Ground Comergace Innor 9. Das lager equipmeat ntaing orp bil stable pose soppy Sesh eaceares an fee 10. Ps 4) 386 OPUS modet ') 286 NESS model 1, Cain ) este power ra power sem 12 Nae onset 13. ostrment cosine ‘Table 3.1 Equipment inventory ember Relaemet cost at 1951 prices) {excaing VAT) 5 1 M ss 1 0 1 1500 ° 1200 ‘bm: 38 Trem S00 1 30 TOTAL a4 19 pes ep (esa VAT) 16700 1290 2550 ao 0 © ass 33 Instrument design “The main instrument charicteistics required forthe pipjacking rescarch include: + Ensy installation into the pipes prior to incorporation into the permanent works, and subsequent removal, thus minimising delays to the tunnelling programme. + Reliabitity and robustness inthe tunnel environment + Designed for required accuracy and simple to calibrate + Minimal modification o the structural performance of the pipe and the soil-pipe + Readings taken remotely from the location ofthe instruments All of these characteristics have featured in the evolution of the individual instrument designs S31 Contact stress cll “The comtatstess cel isthe most complicated of the research struments. It rmensures both the radia ol ines and the shea stress ating at specific locations om he Pipe surface. ‘The assembly shown in Plates 3 and 32 and Figure 23. AC the beast ofthe instrument is a Cambridge Earth Pressure Cell which i based on the orginal work by Arthur and Roscoe (1961) and Stroud (1971). The cells machined on a CNC miling ‘machine from a single block of aluminium alloy 20144 and is witel up with three independent stain gauge cuits, 10 1 sense the radial tess and the ther 1 Sense he shear sttess. Asiembly of the instrument involves bolting the pressure cll to the main tainess steel housing and covering it with a two part eap which protects it from direct, contict with the soil Any load that hears ont the loading platen i tans to he main housing via the earth pressure cell apart fom a small portion of the load that & lost through the hot honded rubber seal, Ground water is prevented fom enaring the instrument by four seals. The loading platen and the frame are joined by a hot bonding. 3s Plate 3.1, Contact stress cell Figure 3.3. Schematic of cell Plate 3.2. Exploded view of contact stress cell ‘rovess which injects pressurised rubber between the adjoining edges which are held in their comet relative pitons by an aluminium moulding jg. The two prt cap i seed aginst the main housing by a uber “O° ring. ‘The hols which connect the long platen to the Cambridge cll are sealed with malleable copper washers and finally the cable outlet atthe base of the housing is detiled to accept a watertight cable gland. ‘As full details of the design calculation ae not included, the main design requirements are ote + Lond capacities: Radiat = 500 kPa : Shear = 200 kPa + Linear relationship between signal and fod + Minimat compliance, especialy to radial sess + Minimal eros-sensivity + Adequate capacity against buckling of the cell webs under working toads + Fast response to changes in fod + Loading platen size based onthe recommendations of Kalstenius aed Bergau (1956) to minimise the effects of soil arching over the instrument. The assembled contiet stess cells are calibrated in a calibration rig which applies simulaneously shearing and compression forces 10 the active face, Pte 3.3. Lid pplication was in accordance with the procedure summarised in Table 32. A typical set of responses for an instrument under radia stuess and sear sires calatons is presented in Figure 34 Profiling and surface texturing of the ative face of the contact stress cell and adequate fixity of the instrument ia the pipe wall ate of paramount inpotance i the instrument isto record representative field data, series of interface shea box test using Leighton Buzzard 14/25 sand and various constuction materials indicated that a ground aT 7 ® Table 3.2. Calibration procedure for contact stress cells oe i ; z ; 5 = Figure 3. Shear box interface tess. polymer modified mortar finish was the most suitable, Figure 3S, satisiying both frictional fe and ease of incosporation into the instrument by gluing a Smm thick mortar dsc ino a ress in the loading platen prior to profiling on a CNC machine. Direct adhesion ofthe cells into specially cast high tolerance holes using a fast curing stuctural glue was performed on site 33.2 Pore pressure probe “The pore pressute probes ate fcated as close as posible to the contct stress ells ‘A fas acting probe originated by Bond & Jane (1989) is used to establish whether rapid ‘pore pressute buildup and dissipation occurs as the pipes are pushed through the ground ‘The probe i illustrated in an explode view in Figure 36 and Plate 3.4. AL the heart of the pore pressure probe is a Druck POCRAI pressure transducer. 1 is» very sensitive instument, giving a full-scale ouput of * 1SmV/V. ‘The wansducer is soplied with its + removed an glued ito the titanium holder sing Act wa part epoxy resin, cate being taken to ensue tht it does not pote From its recess in the holder 310 Figure 3.6 Schematic of pore pressure probe Plate 3.4 Pore pressure prube assembly The mounting block and thrust ring are made from acetal copolymer, a stable ‘engineering plastic in order to avoid sw metal coosion between dissiniar metas, The mounting block caries sintered stainless ste! porous filter, in its font face, which is fixed in place by heat shrink iting and hence & changeable (On assembling the probe, the holder is clamped into pion by the thrust ring which sews tothe hack ofthe mounting block, An O-ring seals the lee aginst he mounting block. Hecause the pressure sensing diaphragm of the transcucer is set hack from the front edge ofits casing there is a small evity (# 180mm’) between the diaphragm and the bck of the probe's porous filter Bond and Jardine the principal design features as + Response time ty < 1 second (in the Iboratory) + immediate recovery fom cavitation + Base of assembly and saturation in the laboratory + No loss of saturation during transfer from the laboratory tothe ound + Robust design for site ase + Porous fier changeable Saturation ofthe fer pores with glycerine is cavied out using a vacuum chamber to evacuate the ir beforehand. Transducer ealieation and trusportationt site ae carried ‘out ina perspex cylindrical container, Plate 3.4, which can be connected 4 Drock digital pressure indicator vin glycerine/itrogen interfie. ‘Transfer of the saturated probe 40 @ ‘Stainless stel housing which i hued into st hoe trough the pipe wall le until the instrumented pipe is ready for jacking, thus minimising possible problems with esaturtion, 312 33.3 Pipe joint pressure cell “The pressure cell used in the pipejack joint isa commercially available instrument hich comprises a rectangular Mt jack formed from two sheets of stainless steel, welded sanound the edges; the space between the plates being filled with Tellus T46 oil. A closed system is achieved by connecting the cell stain gauged diaphragm pressure trnsdcer, Plate 35. Plate 3.5. Pipe joint pressure cell ‘Unusual features for the purpose ofthe esearch include a Smm diameter handle to assist in extracting the cell from the pipe joint, once the monitoring pei i finished, and a limited number of cells ited with platinum resistance temperature probes o monitor insta temperature changes. Calibration of the istrument was performed in a 50 ton Denison loading machine. ‘The tests were conducted wsing concrete blocks as Hoaling media with 313 the cell and pucker composite sandwiched between them, “The response of the instrument under the ey! © application of a uiform sess was found to be linear and repeatable. However, principal assumption in th: performance of the cell is hat its high ratio of area to thickness approaches the ia! ofan infinitely thin clement minimising the influence of stress dstotion de to aration in mus between te sensor pad and the surtounding material. This assumption wos found ta he invalid For the pipe jack joint Figure 3.7 di Figure 3.7 Stres-strain responses of packer and composite, “The stifer response of the compete inthe instrumented joint leads to redistribution of ‘normal stresses ever the cell and an over prediction of the fee ied valve, A subsidiary set of calibration tests investigating tis effect has been conducted by Barton (1992), It should be noted thatthe effect can he minimised by inserting dmmy cslls to provide a ‘complete ring of cells and this has been done on sehemes 4 and S 334. Tube extensometer Te tube extensometer design is base oma tunnel convergence mecsurement device developed hy Chekan and Babich (1982) and enables eonerete movements over a 1.6m tgiuge length to be recorded sing un LVDT mounted atthe centre f the deviee Plate 3.6. Tube extensometer: The instrument as illustrated in Plate 3.6, consists of two 25mm diameter lengths of silver steel machined to enable one length to slide inside the other for a distince of 100mm. Friction is minimised hy ing. two Fnear motion hearings into the sleeve, Support Fnrackets attached to each end ofthe deve a tate holing tte blocks cast into the Pipe surface and a cen 1 sliding support minimises sagging over the central portion. The LVDT s used inthe tube extensometers were calibrated against a micrometer serew bench prior 40 assembly in the tube extensometers, The tube extensometes were in tun as rated against an optical grating connected w a digital display wnit on a miling machine bed. typical response is illustrated in Figure 38. we! Figure 3.8 Typical tube extensometer calibration response. 3.35 Pipe joint movement indicator ‘The pipe joint movement indicator provides detailed information on the join gap ‘compression, angular deviation and shear movement as jacking proceeds. The instrument, Plate 3.7, consists of two aluminium blocks geometcaly shaped to hold htee LVDT sin ‘orthogonal dieetions. A removable cover plate holds the two parts in tei correct relative Pesitions prior t surface mounting on citer side of the instrumented pipe joint. The installation procedure involves bolting one block tothe instrument pipe while it above round and then completing the assembly inthe tunnel hy gluing the second block wo the concrete surface using the cover plate for support, The LVDTs are calbrate against a -orometer sorew bench, 316 Plate 3.7. ‘Thre direct nal joint movement indicator. 33.6 Jack ram load cell The jack ram load cell is a commercially available 200 toane heavy duty compression type, The standard cable and its connection t9 the cell hive been moved to ensure operation under submersed conditions. typical cel i illustted in Plate 3.8 The cells fited with a canrying handle which eases instalation and provides physical protection t the cable connection, A domed loading cap greatly reduces the effects of offset loads and damage tthe thrust ing. ‘The cells atache to the main jack rams wsing 4£ MID threads in its base and a series of coupling plats. A protectivs hood prevents accidental damage to the oad cell. The jack lad cells are calibrated aginst x 50 ton Dension machine, concentric compression being achieved wing a bl beating and spreader Plate 3.8 Jack ram load cell 3.7 Celesco unit The Celesco position displacement unit provides an electrical signl proportional to the linear extension of a stainless stel cable through the use of a precision rotary potentiometer. It is a commercially available instrument and is illustrated in Pate 39. Laboratory calibration is carried out using » series of one metre rules Ind end to end, 33.8 Ground convergence indicator Monitoring the progressive closure of soil onto a jacked pipeline is useful in ‘understanding the mechanisms of load transfer between pipe and soil. filly it was intended to use por in the pipes normally proved for grout and Beaute injection wo measure the rate of closure but problems with continuity of access to the pipeline meant Plate 3.10 round convergence indieat 319 ‘that an automated method of measurement was necessary tv supplement perndie manual readings. The ground convergence indicator his heen designed for his purpose and consists ofa circular main housing which fits into a stel liner bolted 4 he inner surface ‘ofthe pipe. The principe of operation involves monitoring the movement of fin attached toa shaft using a rotary potentiometer. ‘The fn provides a maximum movement of 30mm relative to the tp ofthe housing which is slighly larger than the anticipated wvereut on rmost pipe jacks. Care has been taken to seal the compartment housing the potentiometer from moisture ingress while allowing the wid inthe recess chumer accommodating the fin wo stabilise withthe surtounding ground water. Jamming of the fin is prevented by PTFE wipers. The device i illustated in Plate 3.10, The device was called against 8 300mm sel rule held perpendicular to the top surice ofthe main hewning and parallel to the spring loaded fin 33.9 Data acquisition & power supply Dunnicife (1988), Smith (1988) and Kimber (1989) provide + good fasis for evaluating the advantages and limitations of automatic data acquisition stems. Many of the advantages suggested by the authors such as continuous monitoring of large number of instruments, increased reading sensitivity and aeuricy andthe prise of data storage in a suitable form for divect computer analysis are applicable 1 the pipe jacking avironment with it limited access w the works. OF the limitations, the principal factor, absence ofa knowledgeable observer, citcumvented bythe author keeping detailed log ofall cow afecting the progtess of construction, The ideal dats logge system for the Pipejack site should he modular in form minimising encroachment ypor working spice rugged enough t» operate under harsh site conditions, capable of reading 4 vary of litfeent transducers, capable of operating off site generated electical power and 320 Faeiiting dtect tansmission of dita 40 a PC located above ground 33.9.1 Duta acquisition Figure 3.9. Schematle of data acquisition system, ‘The data acquisition system is shown diagrammatically in Figure 39, A modular ‘das acyuisition system is mounted inside the main instrumented pipe 4 enable readings to be taken remotely fiom the location of the instrumems, For simplicity serial information communication technique has heen employed which allows a 286AT personal computer toe used asthe basis of eontol, The system enables a family of commercially available standard "Datascan" contol and analogue input modules capac of accepting information from diferent types of transducer to be located clase to tne measurement Station. Short lengths of analogue signal cables and communication in digial form between the control modules (7010/7220) and the host computer minimise the isk of signal corruption, The system is capable of 16 bit measurement performance and is readily ‘expanded 1 80 channels which i the maximum required during the research. Each contol module contains & non-volatile memory which resin the setup information, even after system power down, which is particularly desble because the power snd signal cables whieh enter the pipe jack are disconnected each time a pipe i introduced into the drive ‘The system is powered by a 24 Volt de supply. Auto range is standard allowing ranges fof 20mV, 150mV, 1.3V and 10V, “The 7020 provides the facility to connect 16 de voltage signal inputs. The 7021 ‘Provides cipht channels of inputs for strain gauges and platinum resistance thermometers ‘with on hoard encrgsation of LAV. Measurement accuracy wing 16 bit solution forthe 20mV range, the most commonly used range in the esearch, is = 02% rdg + 001% range + 5yV. whichis equivalent to 0.006% ofthe fll range vale, “Measurement speed i dependent upon the number, type and measurement resolution ‘of channels connected 10 each measurement module, the network ditribution of the measurement modules and the communications fink with the PC. For the arrangement shown in Figure 3:9 single measurement on every channel can be mate in 2 seconds, “This is acceptable because the instrument logging strategy du i he pipaacking research requires measurements every § seconds during peshes and 1 minute at all ther times, ‘Configuration ofthe system is caried out using a commercially available software package called Unigen Ic willbe appient from Plate 3.11 that the modules have nt been designed or packaged for rough field conditions and that it has been necessary fe house thet in individual steel boxes to protect them against the tunnel environment, mechanical damage 322 Plate 3.11, Regulator and datascan and the possibilty of total immersion. ach bow is designed to be din mil mounted and internally arranged to accommodate a single input module and a dedicated power supply 3.3.9.2 Power supply ach transducer regutes an energising supply of electric eutrent which cannot he provided hy the datascan units. A purpose built stale” power supply his therefore Been designed and incorporated into the s em. The supply provides a SV or IOV de voltage which is maintained at + 0.0256 of the nominal value For the pipe jacking research the power source is generally the national grid although a 20kVa portable generator has been wsed on one site. ‘The incoming 230V, SOHz ac supply is ist converted to OV with a centre tapped earth, portable ssaton stepown transformer, ‘This isto satisfy the curentsfety regulations for site wse. The HOV ac is fed toa “retitier unit’, housed ina steel box where it is converted via transformers, bridge rectifiers anda smonthing eiruit to low voltage de supply. ‘The low voltage supply is then fed vin sreened cables the steel bones housing the high stability voltage regulators for each transducer and the datascan analogue input modules. ‘This areangement ensures that the magnetic fields associated with the transformers ate physically separated from the sensitive parts of the system (ie the datiscan input modules). The low voltage supplies tre isolated fom earth so that earth cutenis cannot produce a “noise sigma. Cate has also been taken to ensure that vations in the mains supply and spikes do not affect the energising supply tthe transducers. Spikes do not easily transfer through tansformers and even if they_do they are conducted t ground via small value ceramic capacitors in the Wherever practical, lamps have been installed to indicate the presence of supply voltage. This helps greatly in fult Finding, Exch transduce has its own separate supply to enable the system to comtinve operating inthe event of a short citcit in one of the transducer energising ovtpuls, In addition transducer connectors have bwen standardised so that 4 way connectors ate used for signal cables and S way and 6 way ised for 10V and SV transduc cenergisation respectively. Lemo environmentally sealed connectors hive been use to interface the transducers to the steel boxes. Wherever posible, cables are screened and twisted pair to reduce the possibily of electwy magnetic interference fom other equipment. 34. Laboratory performance Reliable, safe calibration procedures are & pre-requisite for felt measurements. ‘This is highlighted by the ineeasing application of modern electronic trunsducers which in most eases ate interpreted as an indirect measurement ofa physical uanity. Calibration procedures are therefore an integral part of the instrument quality assurance and should Provide reliable information on the response single instrument suected 10 varios kinds of insitu excitation, It esent hat instruments are tested and calibrated on regular basis using simple procedures. It ako important that calibrion norma are incorporated into the procedures. Extensive Ihoratory caltation wf the instruments has been carried out prior to the sito eich ite contrat. A fll set of ealbeations typically lakes 135 hours to complete. Typici instrument performances under laboratory conditions| ave presented in Table 3.3, All ofthe instrument types praluce nea linear repeatable responses, Moisture ingress, the main eiuse of ero deft i stain gauged tansdcers, has been eliminated in the designs resting in small ero changes. Lang term changes in the calibration coefficient ofthe instruments are negligible (<25) with the exception af the Pipe joint pressure cell These appear the sensitive kv mine damage sustsined during extction from the joins, Temperature coefficients are not shown in Table 3.3. Field monitering of pipe jack tunnel temperatures indicates that a stable environment (# 1°C) exists when the instrumented pipe is instu. Temperature changes ate therefore not mie varie -aoueuuojiod woes mau sways ayy uo pose axe wo) Ua4a}309 so seus ot a oe eo ne wv z vos co sodtegswatises | (6g 9) soma ta) eet | ae as 24130 esge> tap asus w se 215 oeact Acianay Asawa Asawa ce ASinuines OF je Gionewogey ge GR, fenqye>sea4s ayy 19pum anno a os ON peor ue ams me synod ag a 0 < ss sats bao 326 CHAPTER 4: SITE SELECTION AND PROCEDURES 4.1 Site selection for instrumentation purposes. The research contract was setup withthe intention of instrumenting pipe jacks on five separate schemes. It was envisaged that one scheme would be provided by each of ‘he puricipating water companies, and that number of sites would he offered by each ‘over the contact period, allowing Selection ofthe most suitable sites tobe made, It was further expected that there would be suticient lead time on all schemes forthe research activities to be built ito the contact documents, so thatthe contractor would be fully aware of what be was taking on and the research assistant would have a clay established position onsite (On this basis, review was made ofthe large numberof parameters which might alfet the performance of pipe jacks. These incl: Ground material Pipe size Lbecation Depth of cover + Overburden toaing Length of deve pasion of instrumented pipe Pipe joint profile Packing material Method of excavation Time Method of pipe manufacture 11s apparent trom studying the lst that monitoring of Five pipe jacks cannot cover al the operational and performance aspects of the construction technique. twas therefore necessary 0 judicially assign appropriate levels of importance to the varinbles so that the 4a ‘most significant parameters cou he investigated in relatively canto manner. Decisions were then taken as tn which wuld be kept constant aad which varied between schemes. ‘The result appears in Table 4.1; it eam he Scem hat the major variables were chosen to be the typeof ground, the sizeof pipe and the type of pipe joint. Drives 7 ‘hough cohesive and non-cohesive material were chosen tenable exsting analy {ground loading mols to be eviluted. Firm ely and sundclected fill were considered the mast suitible soil types since they were reputed in the Hiteature a producing the largest jacking fonees per metre run of pipeline, Typi 1 lengths und depth were chosen ‘with relatively uniform ground conditions heing preferred ove the instrumented sections, ‘The use of lubrication and sites with sapidly warying overburden or heavy buildings above the tunnel would e avoided, a8 they woul! complicate the lead dificult iterpetation ofthe instrument readings. Two pipe sizes, 120mm and 18hn internal diameter, were chosen to suit the laboratory work and provide dit fr extrapolation of sults Feom one Size tunnel to anoter in similar ground types. Boh type of UK joint profile were to be investigated 40 esta sh which provides the hest hea! distribution and hence east susceptibility 9 damage though overstessing. The picking material was specified as medium density fbreboard, onthe basis f the laboratory work Ripley (1489): use of the same material throughout was esentl allow proper calibration ofthe pressure ets in the pipe joint. 11 as considered thatthe method of excivation would hav 2 sigoticat effect on the tate of progress an stenghtness of a drive, Hind drives were prefered 40 machine drives hecause it was felt that the tunglling machines were generally only wsed in poor ground conditions and would generate signticant electrical noise in che tunnel. tn stndardising the method of excavation twas hoped that rates of progress would be similar between contracts Finally, it was always intended that the first drive wovld be a pilot scheme, sing 42 Major Vaile be investigated in reserch “Tobe hept as ‘contact selection Notes: MDF - Mediam Density Fibrebard Table 4 ‘Seheme Growed Type Pipe Size Pipe Joint Depth Length Poston of Te Pipe Lubrication Packer Type Exevaton Technique Pipe Manufacturing Process overburden ning Monitoring Period Firm Fim Sand Sand! hy Cy Gravel Grivel ram | 1800 wo | 1200 Bun be Ban watt Wal Sm | SIm | Sm Sim lot 100m 100m <160m Pipe Pipe | Mid Pie Nas NO Noa No No [No | No MDF | MDF MDF MDF. Hand | Hand Hand Hand Any Any Any Any Simple Simple | Simple Simple Shoa | Short Short Short Tem Team Term Tem deal site requirements for research. 43 5 Fim cay 1800 ‘at 12m No MDF Machine any Simple Long Term the instruments but in seduced numbers, identify problems with the instument designs, ste procedures and data eolletion. There would then be time avaiable 19 make any modifications necessary before the four main schemes, which would use the fu In practice the process of selecting schemes was very different Details of the actual instrumented schemes are given in Table 4.2. A location map showing the {geographical distribution of the sites is included a8 Figure 4.1. Figure 4.1 Geographical distribution of sites. “The pilot scheme proceeded more of les as planned bu by then the effec of privatisation ‘of the water industry were having 4 major influence on both the organisation and immediate plans of the various areas. Athough a number of pipe jacking jobs were ‘dentited, very few were even reasonably suitable in terms of locaton, timing and ground conditions forthe research work ‘After long delay, the next owo were taken on at short notice, withthe contactors “already onsite, and the research introduced 28 variation to the contract. Scheme 2 was 4a ' 2 3 ute ‘Avpst | January 1991” March 1991 Lscton Beton, | Gatdead, Hone On, Lamy” Tyaesiie SE London tem SW Wate Norbunbrin Thames Water Boon | Warr Mac CComncior Lethe | DCT Pipe Super Buta ARC Pe LD (am) 13050 co Ground Type | SHIT | Wathen London Cay Gc Mudiowe coer) det Ted net Dove Leth ho 78 oo Position of Tet Fige No Pye NOD Pye Nos Pre Latriction | Na yo No Picker + MDF MDF Mor Excretion Hand Hand od Notes: * Monitoring only for part of dive 4 MDP = Medium Density Fibreboard ‘Table 42 Details of actual schemes monitored, 45 Chote, NW Water corey be. Halcro Dem sity Sand no ana 5 December enn, Gl Cictentam i Spun Conte Lowe Saad & DF Suny TM Tate replacement for one that had heen in an advanced stage of planning, but was then postponed: while scheme 3 cesely followed scheme 2 requiring an early ext fom scheme 2 and negligible time for repair or recalbation of instruments between jobs Jn parallel with schemes 2 and 3, arrangements were being made for scheme 4 on 4 job in Yorkshire, Contract documents incorporating the research work were in preparation, when this Scheme was abo pestponed and would be uw le tobe inched inthe thiee year programme, Fomunately two other sites hecime available, both in ‘ohesionless material, and this allowed a fal set of five schemes in a varety of ground conditions. In hoth eases, the research was introduced a a variation toa contract that fad already heen le, Thus only inthe case of the pilot test was the intnded procedure followed, “The choice of parameters sted has hus in effect heen decided by the avalabiity of sites although an attempt has sill been made to Keep some coherence inthe research programme. The main variable has tuned out 4 be the ground ype, with drives insti tlacial clay, weathered modstone, stiff plastic (London) clay, dense fine sity sand and loose sand and gravel. Pipe intemal diameter has varied between 1200mm and 180mm, and cover depts fom 1.Sm to a8 muchas 21m under an embankment ‘One variable has dlssppeared. Either as result of field experience, ot fim the findings ofthe laboratory tess the supeioiyof but to n-wall joins whenever significant Jacking forces are expected appears i have been accepted, and steel-banded pipes have been specified forall schemes. The three drives through cohesive ground have been completed without the use of lubricant, though with some concern on scheme 3, The non- cohesive drives were lubricated, although in scheme 4 it was posible to monitor part of the drive prior w lubricating. Schemes 1404 were hand excavated, The final scheme was riven using an Eur-Iseki Unclemole tunel boring machine 46 Coe Se St nd ser 42 Details ofthe sehemes 421. Scheme 1 “The pilot test was completed succesfully om 120mm internal diameter, 6m long dive in Botton, ‘The een, Bollon Metropolitan Borough Count aeting as agents For Noth West Water Lid, designed the Bury Road Resewerage Scheme 4 alleviate foul flooding which vas atfecting premises in the Breightmet area of Bolton, Part of the Scheme, the Milnthorpe Ros Retcoton Tank, was brought fava as 2 separate phase when idemitied by North West Water as heing suitable site fr the pil test “The contract was awarded to Laetbore Ltd. Wonk commenced on site 23 July 1990 withthe research element being cated wot between 26 July and 24 August 1990, Pipe jacking was chosen hecause the retention ink pase ditty elow the driveways of 8 number of houses nd the client wishes minimise inconvenience wo the ‘owners and a superstore which could only he accessed fom Milathorpe Row, Figure 4.23, Vertically cast reinforced concrete pipes supplied by C.V. Buchan Lid were used ‘The ead pipe wis fied with a single round convergence snealor andthe main instrumented pipe was positioned Sm behind the shiek, This ened the pipes ta be pushed out ofthe tunnel into the reepton pit andthe main insrunented pipe 0 be jacked tuck ino the tunnel fora distance of two pipe lengths ‘A simpli longitudinal section through the dive shown in Figure 42h, The tank was constructed with a eover depth of approximately Sm. The face hy taken during the construction period, Figure 4.3 illustrates tha the drive was predominertly though tiff ‘lacal clay, Soil stength an chsitiation details were obtained tom insta and laboratory testing of material a four chainayes. A Summaty af the revels sppeas in Table 443 which indicates tha the undraned shear sttength increased fm SOKPA to SO0KPa with increasing tunnel length. Stblily inices hased on these values sugested tthe Face and tunnel bore would sind unsupported fr some time ater exetion ht Se Sa ad Pon 2) Site Plan Figure 4.2 Details of scheme 1 1) Simpitied longitudinal sectlon 48 Sse) Dt aaa > ‘Table 4.3 Insitu and laboratory test data from face samples obtained during Schemel. [] denotes shear strength coresponding to «, 49 42.2 Scheme 2 Instrumentation and monitoring of scheme 2 was earied out between 21 January and 25 February 1991 on 4 110m hand driven 13S0mm internal diame tunnel mainly through weathered mudstone, coal measures and glacial flay which was. part of Northum Water's Tyneside Western Interceptor Sewer Project, Phase 3. The contractor was DCT and pipes wete supplied by ARC. A location pla, longitadinal section indicating the depth of tunnel and differents types obtained from trial horeholes and a more detailed tun 1 face log reconded during construction are included as Figure 44(¢c). Progress on the drive was delayed by one week when it encountered a backfilled mine shaft, approximately 2m square 1m into the Ave, which was negotiated by constucting 2 timber heading and casing a reinforced conerete base over which th pipeline was jacked, The need to run schent 3 hard on the heels of scheme 2, and the unforescen delay resulted in only 40m ofthe drive being completed before leaving site The instrumented pipe was positioned 25m behind the shield; 15m of the drive was monitored in devil Figure 44a. Site Plan 410 2 ©) Face to Figure 44 Details of seheme 2 423 Scheme 3 Scheme 3 involved instrumening # pressurised pipeline which was installed by pipejcking at a reservoir in southeast London 10 improve the mains water supply by drawing more fully on water storage and pumping facies. The orgial 1h Century brick reserve consisted of four eel interconnected bya network of pipes. Wate lowed, via one cell, 0a pumping shaft where it was pumped into the mains system, The new pipeline, some 110m in length, linked a second cll directly tothe pumping station, thereby impeoving operating capacity “he scheme was put forward by Thames Water Utilities and wax awarded to Barhale Construction Lud. The extent of the civil enginering works is lst in the simplified plan and fongitudial section of Figure 45. A temporary worcing shaft, approximately 16 metes deep was sunk on the line of the pipe jack anda short 25m drive cinco the pumping shaft. This backshun, once groute, provide! the thrust resistance forthe longer instrumented dive under the reservoir. Simultaneously with the pipe jacking work, a reception shaft was sunk though the floor ofthe reservoir. Sine the line was to he subject a 25 har pressure test, non-standard 1830mm internal diameter stel banded, ‘concrete pipes with wall hicknes of 225mm were required and suppliedby C.V. Buchan Lid, Work started on the instrumented hand drive on 4 March 1991 and was completed bby 22 March 199, including 24 hou shifwork between 18-22 March. ‘The instrumented Pipe was positioned 35m behind the shield, Cover depth varied fom Im atthe temporary shaft to 2im under the eservoir. The ground materia, sti everconsotidated London clay, was ensistent throughout the drive as indicated in the face log of Figure 4.6. The highly fissured state of the material resulted in preferential fracturing during excavation as indicated by the blocky nature ofthe spoil Plate 4.1. A laboratory test programme based ‘on samples obtained at chainage 22m was earicd out by the Thames Gectechnies Group: 4 summary of the results appears in Table 4. 42 Figure 4.5. Details of scheme 3. 1) Simplified longitudinal section. Plate 4.1 Excavated material Honor Oak \ Z 3 ® _— Han, Sued a Sol nyo | | Saline say L e Lecton 1 2 3 «os ye) | Sst easy ie 59 o pt @ zp | | | | i 1 ‘ acid ca (ABC) CABC) GAC) ABO) Sterns | / Siok 30Pe foray [9] eon aa “ Ga ° ° wer] comets easo, Varied Sree ‘st pp Feetneter (round ti or wats be oad Seat Vane {Ground si or ws bond Table 44 Laboratory test data from face samples obtained at chainage 22m during Scheme 3 (1 denotes shear strength comesponding 104, aus 424 Scheme 4 Instrumentation and monitoring of scheme 4 was carried out hetween 15 July and 9 August 1991 on a 158m, 1470mm internal diameter drive from shaft 2 to shaft 1 in Figure 4.7. The scheme was part of North West Water's Abbey Villige/Pleasington Sewerage scheme, tage 2, ‘The contractor was Barhale Constuction, sho had eared out the work on scheme 3, 1nd Halerow were the supervising engineer. The conetete pipes were supplied hy Spun Concrete Li, {A detailed face log taken during the monitoring period shows that the tunnel was ‘riven through initally dense brown silty sand with bands of fm clay, then laminated sandy silt with occasional bands of si clay and finally dense brown sane with tf sandy boulder clay intrusions. Grading curves for the predominenly non-cohesive materials are included withthe login Figure 4.8, Stabily of the fice was good as iasrated in Plates 44.2 and 43 suggesting that rapid collapse ofthe ground ont the pipes was unlikely and thatthe intended use of lubrication could be defeted ntl later in the drve, as requested by the research group. ‘The main instrumented pipe was inserted 40m behind the shield lubrication wns introduced after 82m and the research equipment was removed after 115m ofthe drive was complet, to allow sufficient prepaations for scheme S. The addtional site costs for this contact were borne by Yorkshire Water 425 Scheme 5 The Hewlett Road Interceptor Sewer forms par of the All Saints Foul and Surface ‘Water Resewerage strategy of Cheltenham Borough Council who are agents for Severn “Trent Water, The 100mm internal diameter sewer runs below the cent line ofa major road and involved tunnelling through sands and gravels below the water tble, Figure 4.9, which made standard tunnelling techniques dificult without the use of compressed air or ') Simplified longitudinal section with borehole data Figure 4.7 Details of scheme 4. Laminated sandy silt Silty sand Figure 4.8. Face log and grading curves from scheme 4 Plate 43. Laminated sandy silt age She ead Pde 2) Site plan. Figure 49. Details of seheme 5. Simplified longitudinal section. 420 Paper Se Se a Per ‘sound water lowering techniques. The relatively shallaw cover depth of he tunnel 47m, and the presence of masonry houses built around the urn of the century, without substantial foundations, meant both techniques could he fraught with potential problems “The contact was awarded to Lilley Consteution based onthe use of an Eurotsek Unelemole, which provides fll face bentonite pressure support and incorporates an «eccentric cone crusher head which reduces cobbles and small boulders (upto 308 of shield ‘outside diameter in sie) suivietly to eale them to be transported away by the pumped bentonite spoil removal system. ‘The contractor elected to dive the ttl 360m from shat 11 using thee intejac at various locations along the pipeline. “The pipejacking research was caried out between 5 November axl 11 December 1991, ‘The instrumented pipe was positioned 8mm behind the machine and was the fis full length pipe ater the 40m radius curve: he site monitoring was competedafter installation ‘of 2600 of pipeline ‘Asa separate exercise nn additonal survey wis carried out to investigate the effects fof the tunnelling operations on the surtounding. ground using. piezometer and surfuce/subsurface setiement stations, Installation of the boreholes and instruments was rruments Lid, and subsequent loging of the instruments was cated cutred out by Soil ‘out by the Oxford pipejacking research group. 43 Planning and execution of the fieldwork The success of site hased research is not only dependent upon good design of instruments and data acquisition systems but 80 thorough planing of the site procedures And incorporation ofthe research activites nto the contact documents, 3c that subsequent contractual pressures are minimised. Pre-contact mestngs with the client and the contractor were exttemely valuable in exphining the aims of the research and Understanding the rks and motivations of the paris involved, A summary of the 421 important research related ems i included as Table 4, ITEM DESCRIPTION unit 1. Supply of special pipe sum 2, Provide, establish and maintain protective rig set liner sum 3. Removal of protective rigid steel liner Sum 44 Installation of compression load cells between theust ring and jack rams ne and mounting of Celesco displacement unit and dita acqustion bos in the thrust pit 5, Removal of pit bottom instruments a 6. Standing by pipe jacking driving operations br 7. Supply and fiting of 100x12mm thick medium density fibreboard nr BS1142: Part 1 and 2 as packer to joint annular ring in joints with, pressure cells 8. Recowery of joint pressure cells trom pipe joint and make goo! packing nr material 9. Recovery by overcoring of pore pressure probes and contact stress ar cells, removal of cores containing instruments and making good hoes 10, Entry into tunnel for inspection and survey he ‘Table 45 Main research related bill ten omer ems of a more general nature including transportation and handing of the cxquipment container, electricity supply, stey, security andthe contractor's responsibility to the instruments were also covered “Te primary bill items ae all activites which cam have x major disruptive effect on to's programme of work, Fitting out ofthe instrumented pipe was an line sctviy which was the responsibilty ofthe Research Assistant Unforeseen delays caused by the te arch were covered by an agreed stnding by rate 422 Cg Sen nd Pie 43.1 Instrument instalation [As far as possible, preparation of the instrumented pipes 100k pie ofine, The exact methods of fixing the instruments varied with the way in which the pipes were formed (veicaly east r cenit spun). The high risk of nstumett damage during the astng process and subsequent delivery to site ruled out nslaton at he pipework {twas therefore necessary to "build in” the necesaty holes, inserts or brackets, Figure 4.20, and deliver the pipes 1 site with sufficient time to complete instalston. Plate 4.4 itustes the process adopted for pipes produced hy the vertically cast method high tolerance circular inserts were used between the inner and outer mould faces 19 produce holes for the glued in-wall instruments and pre-illed tufnel tapered blocks fr tapping and bolting of the surface mounted instruments Pipes produced by the centitugally spun process presented a problem with regard tothe provision of surface mounta inserts becuse the internal surice wes no is fe Bracket assemblies similar to that shown in Plate 4.5 were epoxy growed into pre-dited holes to provide eoretly spaced threaded dowel (On site, the fist priority was to install he jacking pit instrumens. The jack load cells were coupled to the rims, Pate 46 and the Celese displacement unt mounted shove ‘he tunnel entrance and hooked over the steel band of the pipe inthe jacking pit, Pate 47. ‘Once communication was established between the pit botany dat quis tion box and the PC, data could be collected on the initial pipeline installation loads, and work could start con the “special pipe. The instruments were fited while the pipe was om the surface, and ‘only parts of the joint movement indicators had wo be glued in place ater the pipe was in the tunnel. Plate 4.8 illustrates typical instrumented pipe prior wo lowering into the Jacking pit, The total time for assembly and system checking took betwesn two and tree days. It was convenient over part of this period to monitor the effets of ambient temperature actuation on each ofthe insament types or subset tempest compensation. 423 Figure 4.10 Layout of instrument pipe. 404 Plate 4.5. Bracket assembly for a spun pipe Plate 4.8 Instrumented pipe ready for lowering into the jacking pit. 43.2 Instrument protection Figure 4.11, Protective liner details. Protection ofthe instrument and data acquisition system agninst mechanical damage ‘vas provided hy sel liner which fited ise the instrumented pipe and was detailed to void significant stiffening ofits response, Figure 4.11 and Plate 4.9. The liner was ‘typically 30mm smaller in diameter than she pipeline, (although diferent shapes were wsed to optimise the sie ofthe constition) and was fabricated in two 14m lengths, which was Slightly longer than the 2.5m instrumented pipe length. Each overhanging portion was 428 Plate 49° Protective liner in place bolted to a see cradle fastened to the leading and tiling pipes. The liner was supported inthe pie by a timber cradle which was shaped to cover a 150 degree a about the pipe invert and allow the instruments to be coretly positioned, Cross hoks in the timber Provided ducts for routing cables around the pipe circumference. Antiultion ofthe liner was provided by a central steel handed joint which owe the two halves toside 10mm relative to each other, A set of ramps inthe leading and tailing pipes allow muck skips to uavel through the linet. The timber endle was ite while the pipe wis onthe surfaces the liner and ramps had toe ited once the pipe was in the tunnel, causing up W 6 hours delay. This delay and the loss of production 4s result of the constriction were pried hy the contractor inthe appropriate bil tem. 1t hasbeen found in practice thatthe protection it provided was well wort the slight reduction in productivity that it eaesed 49 43.3 Instrument recovery Removal of al the instruments took place at the end of the drive ui before then, a a weekend. The jacking pit instruments were firstly unbolted from their anchorage points, then the liner was removed and the surface mounted extensometer joint movement indicators and data collection boxes removed from the tunnel, ‘The contact tess cells and pore pressure probes glued into the side walls ofthe pipe were recovered by owercoring nd the resulting holes made good as requited Removal of the pressure cell fom the pipe joints required special anchorage brackets to be bolted teach side of the insttumented joint and eylinder jacks inserted to push the joint apart and allow removal of the cell and insettion of replacement packing ‘materia, Plate 4.10. This was only possible provided the instrumented pipe was near one fend of the tunnel or an aetuil or dummy interjack so thatthe friction on only a few pipes 430 Inad to be overcome. This eovld not he arranged in scheme three: hete the cells were recovered by dling and eating out the packing material in the join, extracting the cells fand_ replacing the packer. This was succesful, but was tedious and caused some superficial damage to the cell 44 Data handling Continuous logging ofthe ack Toad ees an the large cluster of instruments in the special pipe generated up to 20M bytes of dita written to print files iy ASCH format, whichis compatile with Lotus 123 spread shets. To ensue that excessive data were not handled, logging was performed on two diferent time bases. Dring pushes ypcally of two minutes duration, data was logged every five seconds, while at all other times intervals ‘of one minute were used. Regul backing up of data on floppy’ disk was cartied out twice aly, with two copies being made and Kept in separate locations at al tes, ‘8.386 AT personal computer was used for processing. A limitation of the 123, sottwate is that the ASCII files containing site data can only be 240 characters wide (16 channels of dats) for importing into a spread sheet. Approximately 70 channels were used ring the research ad therefore the site data files had tobe split into smaller manageable s0ups of instruments using a Fortran program. Aknough the data coud be sealed and converted to engineering units atthe time of capture, it was preferable ts record only the ‘voltage readings fom the various instruments the conversions being carried out in spread sheets at Oxford, minimising the possibility of faducing eros at source onsite. Traces ‘of eaeh instrument response on a daily basis were produced. Figure 4.12 illustrates 9 ‘typical set of responses ftom the bottom pore pressure probe and contct stress cell fom scheme 1, obtained over single working shift, The graphs clearly illastrite the magnitude of stress changes caused by jacking, ‘These formed the starting point For detailed analysis of individual pushes and eross correlation between different instruments which was 431 nore sensresi Figure 4.12 Typical set of interface stress records during a single shit 432 generally caried out ona chainage basis. Back up copies ofe h stage of processing have been created and stored on floppy dik 45 Performance of the instruments and site procedures (Once monitoring was in progress, good eommunieaton between the miners and the researcher ensured that there was very litle disruption the normal working pattetn. ‘The fnsrumentation was even of assistance 1 the contactor on certain sekemes providing precise and continuous records of jacking loads which helped with decisions on whether lor when to lubricate of use an inejack “The careful selection and design of the instruments fed to few in service fires, “Table 46. The pipe Joint pressure cells, jack load cells pipe joint movement indicators, tube extensometer and data sequision system demonstrated ther fitress for purpose uring the pilot test, The contact rss cell intally highlighted problems due t9 ground water ingress but this was overcome by redesigning the primary seals. The pore pressure probes were found to be susceptible to cable damage during extraction, Only the Celesco unit and the ground convergence indicator perirmed poorly in the field, The Celesco unit was very prone to damage and required frequent repair, although it did provide useful data, The ground convergence indicator fad some trouble with the ingress of fine particles past the PTFE seals causing jamming, bu consents of Programming or ground conditions have in any case prevented this fom being wed afer the frst scheme ‘The performance of the data acquisition system was extemely saifictory. At no time was it necessary, during the five schemes, 1 stop the contractor as a result of a general systems failure, which would have necessitated removal ofthe protective finer to ain access 1 the individual boxes 433 genes per 9p age. = [aes sree . vu yew vw vn ‘The suocess of the ste procedures can readily be seen in Figure 13 which itystrates daily productivity in terms of metreage for each of the four hand drives. i Figure 4.13. Effect of research on pipe ja progress. ‘The reductions of 13% and 8% in average production for Schemes 3 and 4 with instruments in place, ace overestimites since the effect of increased haulage distnse mst also be taken into account, Scheme 2 illustrates no change which isa result ofthe difiuly in excavating the face mates, ‘The data fom scheme Sis not included cause pumped bentonite spoil removal system was wsedl and all surveys were performed outside normal working hous. Comparison of scheme 1 isnot possible because the major prt of the drive was completed with the instruments in place. Problems outside the control ofthe researchers occured during the early stages of the fieldwork. Scheme I strted during some ofthe hotles! Laneashie westher for decades nd was followed by tential rain which affected ground conditions tothe depth of the pipe. An unexpected water main exacerbated the situation by temporarity long the Pipeline. Scheme 2 took place during severe winter wether the extremes of temperature were potentially dhmaging 10 insruments exposed on the surfice ard made correct pplication and curing of the le 1 the in-wall instruments very dificuk. An uncharted bocKFilled mine shat further delayed production on tis scheme. By contrast, schemes 3 9 S progressed surprisingly smowsthly, although the rate of progres on the final scheme was affected by the need to carry out remedial work 4 pipe ‘damage caused by localised crashing of pipe joins. CHAPTER 5: PIPE END LOAD TRANSFER 5.1 Introduction ‘The lack of clear guidance on allowable pipe end loads is reflected inthe elause ofthe PIA‘s “guide to pipe jacking design” which has been one o the best safeguards to date: “The permitted tolerances on line and level shall be within 75mm of tre line and 250mm ‘of rue level at any point inthe drive, Adjustments in line ana level shoul be gradual and the pipe manufacturers permitted draw and deflection must never be exceeded at any joint” od dw oF ‘The main concern for contactors is what is the manufgturer’s permit deflection and what oad wil the pine take when it aceus? Cerin the argular defections ‘quoted in BSS011 : Patt 120 of 1° for pipes between 900 - 200m intemal diameter and 12° fo pipes between 1350mm and 18OQnm shouldbe tated circumsfectly as they are minimum angles for waterighiness, ‘Thischapterconcenvates onthe measured tunnetaignments, the pressure dstebutions in the pipe joins, and the relations between them for schemes 1.34 and 5. Data have been sclectd from these sites forthe Following reasons (Scheme I~ shallow tunnel through firm clay, ground contact at the hase of the pipe ‘only, instrumented pipe near the front end where Face loading conditions dominate: Scheme 3 - deep tunnel in highly pase lay, large contact pressures between pipe and ground: (ii) Scheme 4 tunnel in relatively stable bore through generally non-cohesive materi, instrumented pipe inserted approximately a mid length Gv) _ Scheme’ lubricated machine divein andand gravel below the water table, relatively large deviations from line and level and some pipe damage under jacking los sa Detils.of the instrumentation arrays inthe rear joint of the special pipes ae lasted in Figures 5.1 and S.2. The joint packing material in the insteumented joints was standardised 2s medium density fibyeboand, 1 correlate with the laboratory experiments and assist Interpretation of the joint cel readings. 5.2 Directional contrt of pipe jacks 5.21 Changes of pipetine alignment in practice Excavation atthe face of a pipe jack can deviate fom the intended fine and level Constantconestions othe measured deviations induce the pipe sting w takes ig-zag.couse, known as "wrigule", which esults in deflections at pipe joints, ‘Throughout the researc, ‘changes in pipeline alignment hive been monitored by carrying out ine ai evel surveys (osvally daily) fo he fll ength of the tune during the construction period. The resulting Plots are presente in Figures 5.3 w 5.6, For city only three surveys are shown foreach scheme, Schemes 1.3 and 4 would be considered well controlled drives while scheme $ significantly exceeded specification, Pipe damage caused by localised crusting of the joint ‘vas experienced ina numberof pipes at postions A and B of Figure 56. Examination ofthe ehange in positiono pipes reveals that misaligamentof the pipelines id not significantly alter thoughout constuction, This is surprising since the restining scton ean only he provided hy ground reactions which were expected we diferent inthe various types of ground, Furthe information on this aspect of the work can be Found in Chapter 6, In addition ¢ would appear that the directional contol was geacally poorest at the start ofeach drive, ‘This ean have serious repereussions, particularly on long drives because the worst misalignments ten coincide with the most heavily loaded pipes. Care ‘oken during the setting up ofthe thrust pit nd in paticular accurate and regula monitoring ‘during the early phases ofa pipe jack while personne! are familiarsing themselves with the round conditions should reduce this effect. 52 Figure 8.1 Schematic of the instrumented rear joint in schemes 1 and 3. 53 ABBEY VILLAGE Figure 5.2 Schematic of the instrumented rear joint in schemes 4 and 5. 54 cage: petal 5 i * a : é § o>, a —— i a cy CcrANAGE (ny) Figure 53. Tunnel alignment surveys for scheme 1. ss hes Poet To is HONOR OAK = : 5 . a —_| ° 20 «0 © 80 ANA im E & ° 2 * oo 80 E os P85 20 % 60 0 CHANGE 5 pao ‘nana a) Figure $4 Tunnel alignment surveys for scheme 3. 56 Peat Te i ‘ABBEY VILLAGE i 5 g ° 2 10 comnact E 40 +— —| 20 : ° 20 a 60 % 0 E . . or tae : os 00 : | ° 6 Oy 165 Y99 CHINA om) é 4 Figure $8 Tunnel alignment surveys for scheme 4. 37 eA | : vera OEE Figure $6 Tunnel alignment surveys for scheme § ss ‘An intresting feature of scheme 5 isthe apparent Suetatin in vertical alignment between chainages 50m and 110m, The level of the tunnel rises by approximately 20mm beeween the surveys of LS November and 22 November an thn falls by Dm asndicted bythe survey 4 Dscember 1991, Comparison ofthe extra dismeter af the pipe and he outside diameter ofthe Unclemole indicate a 20mm difference. Site recon prior the 15 November show that lubrication was delayed util 28m ofthe drive was completed and hat 1 rapid at of pipeline advance was being achieved, typically 20m per shit, which slowed to 10m per shift during the period 15-22 November. It would appear tha the initial detay in fuvicaring and the high rate f advance prevented aly effective lubration sone from ting developed around the pipe. Te subsequent reduction inthe rate of advance and systematic lubrication ofthe pipeline resulted in successful stabilisation ef the tunel bore and pipeline buoyancy. As the machine moved away from Shaft St ehinage 160m the ‘main lobrication effort was shifted tothe nex section ofthe dive with whit appeat to be a sow dissipation ofthe lubricating Mud tothe surrounding ground in the init al tunnel section Figure 8.7 Possible deflection angles under typical line and level specification ‘Theimponanceof the directional contolof pipe jacks illustrated Figure S.7. Based ‘on an allowable tolerance on line and level of + $Omm, Hough (1986) shows how large deflections can become. Two 2.5m long pipes can have a maximum deflection angle of 4.6° and sillbe within ine and level tolerance. Deflections of thismagnitude considerably exceed 59 the I" angular deletion ofthe quality assurance control test of Table 6 in BSS911 Part 120 and would result in he pipe joint Being incapable of performing its required functions. Iti therefore clear hat isnot scien o monitor only ine an level high loads ae being plied but assess the thee dimensional erenation of pipes to enable fll interpretation cof misalignment angles. The joint angular deflection can realy be determined fom lin and sed in section $2.2. evel surveys as sum 5.2.2 Evaluation of joi sngular deflection from line and level surveys ‘Assume two pipes have an angular defection ft their Common joint as ilustated in Figure 58. The spatial postions ofthe centres ofeach pipe cros section relative to the rue alignment are represented hy points A and Cat the ends of each pipe and Bat the centre of the join between them, ‘The coordinate of hese pints represent the horizontal and vertical offets measure from the tunel see during the surveys andthe chainageof the individual Pies Figure $8 Coordinate system for determining from line and level surveys. 5.0 Direction cosines of AB are: (ic) Pema 2 Likewise diretion cosines of BC are: By definition the misalignment angle Bi given by osph = Hkmmn los of angular detection desived rom the tine and evel surveyson chemes 1, Sand are included in Figues5.3005.. Values of between zervand 0.3 were ypcaly measored lon the thre dives. The apparent seater in he values sa result the sensitityof the mode to errors in the messuted difeences hetween successive pipes: these can result from ‘manufacturing tolerances on pipe diameter, shear displacements atin or more importantly hecause ofthe inherent ditficulty of measuring small changes inline and vel off «tunnel laser beam. The angles for scheme 5 could mo be determined because it was necessary 19 reduce the tunnel survey time by reconling the pestion of every second pipe 5.23 Evaluation of angular deflection and centre of compression from Joint gap monitoring ‘The previous assessment of pipe deflection is generally only appicabe to unloaded or partially relaxed pipelines si ost surveys are cari out during break imes. To obtain fan assessment of angular deflection when pipes are being jacked it is necessary to automatically monitor the changes in joint gap. This performed using he tree equispaced cup: Poe blame joint movement indicators positioned at each of the instrumented pipe joints, gues 5.1 and 53.2. A deflected joint may be described in terms ofthe notation and sign convention of Figure 59, Figure 8.9 Notation for deflected instrumented o Coordinates of points A, Band Care 0 NR [aR | ma [2 . respectively al. |e ck 2 2 4, by Ie flows tha: aR vie “GR Vector Al 2 BC 2 aR ° ae 2 2 (hy hh (h=m) 52 ngs pete tote Hvecwrp », ‘Then for po be normal to AB: Ph) [VSR 7 | 20 Py 3K po] [=m Simi for BC and CA, “Therefore for pt be normal w the plane: VIR -aR Ohh) [Py “ vir 0 thot! [py w aR 2 in-h)) — [p From (i) De. (he=h) ER and tom (andi) py Use tn=2h) Pe 3 “The direction cosines f pare: Poo Ps Ps where La pre pre pe L ae pe VRE hae Pelton ie ‘at a close approximation is piven by Where @ is the rial angle 10 the poiat of maximum ym tine = 2 Compression Tih. 24)) (Changes in pipe atgnment, athe rear instrumented joint are preseted for schemes 1, 3 Sand part of Sin Figures 8.3 1 56, Valucs are recom tthe strand en of each push, Itisapparent hat the variation in for scheme 1 i greater than fr schemes 3 and 4; al tree schemes satisfying line and level specification, This was anticipated Hecause the instrumented join in scheme 1 was clove to the shield and therefore nt only influenced by ‘ariations in tine and tevel but eorective action by the mints. The positon of maximum angular msaligament.0.3°,inscheme 3shows excellent agreement with he position of pid ‘change inline and level. tals cleasly demonstrates the advantage of gradal corrections to deviations which result in ypical misalignment angles of 01° forthe remainder of the drive ‘The alignment or scheme 4 was exceptionally vod in terms of line and level, and probably reflectsthe optimum performance thatcanbeachievedon a handive. However the changes indirection are to abeupteausing misignment angles lage a 0.29. which demonstrates that uncritical adherence 1 specifications based on absolute line and level cin be ‘counterproductive, The section of tunnel subject pipe damage, between chainages 90m snd 135m in Scheme 5, illustrated valves of has large as 082 Automatic monitoring of joint gaps at both ends ofthe pipe provide the opportunity tocheok the epetablity of misalignment angles at any given cainage. The relationship fr scheme 3is typical of all the hand drives, Figure 510, luting od agreement fr peak values and overall profile but some divergence at smaller angles. For scheme 5 the instrumented pipe was positioned two half pipes behind the fist interjac. The pots of fi Figure 5.6, show thatthe ea joints uaffectedby the clse proximity of the atexjack adi slignment induced deviations dictate joint behaviour, By contrast the foe joint asteates saa Cher Pc tad tae Te much smaller deflection angles. The interiack may provide additional degrees of freedom Which combined with the increased Mexbilty of the half pipes allows. sig reorientation of the join, —ee eee se cerA er) é aes oe Bs 4 cunase on) Figure §.10 Comparison of the angular deviation in the front and rear instrumented Joints at the same position throughout the Honor Oak drive. ‘The joint gap monitoring ann enabed changes in 4 he established etween the tnloaded tunnel state and when full jacking loud was applied, Responses from schemes 1, 23,4 and 5 ave included in Figures 5.11 and 5.12. Smal reduction, typically 0.02 *, were iwstrated in the unlbricated hand drives suggesting the pipelines try tstighten under oad but are restrained by ground reactions. It would appear that normal tunnel surveys, which se usually undertaken during breaks of work when the pipe string is unkoed, are suficient to establish pipe misalignments for site contol purposes. In the fully lubvicated drive the Pie buoyancy allows lager reductions in joint angle, typically 0.08". Close examination of th end stess distributions in Scheme S, section 5.3, shows pipe damage oscurred at angles hetween 0.5 “and 06 "when combined withthe maximum jacking loads. JOLTON cera io) | fi oo +— 2 HONOR OAK Potala ee ty Figure $.11 Variation in Beta due t» application of jacking loa; schemes 1 and 3. 516 Pend oat Tor marian Pare to 7090m ma j 2 ' as @ 02}—+- - os SE aon | Tee oo | ——_—_I1 ——s rie CHELTENHAM j, restonee 90m : 5 G i EH Af 7072 74 76 78 80 82 84 86 88 90 (Mhoisends) TINE (see) Figure $.12 Variation in Beta due to application of jacking load; schemes 4 and 5. 5.24 Pipe end squareness audit ‘Ackitional angular deviations ean oecur at joints dve ta lack of pipe end squarenes, and in theory these nee to be offset against the permissible misalignmentangles, To evaluate the routinely obtained limits for this eect # numberof pipes were measured atthe various pipe manufacturers works, covering the range of sizes used during the research, The results are dete in Table $.1 oe Seoift satu Ratatat | Seer Notes 1. A.& B refer to two planes at 90° to each other 2. = No reading taken Table 5. End squareness audit. cpa pened Ter ‘The values demonstrate that pipes made by Brith manufacturers ane well within tolerance for end sguarenes, It willbe noted that an infinite combination ofangles can vecur when two pipes ae joined and tht the exact values were not determined athe instrumented joints. Although the worst combinations of angles typically produce values around (1% i ‘was observed in peactce tha the angular deviations comeate well with he alignment profiles suggesting that the end squarenesselfect is suficiently small fori tobe discounted inthe Field tests 18.3 Load transfer at joints “The field work allows joint dfletion angles 10 be coreated aginst joint packer compression and load distribution, Results at selection of angular misalignments ace presented in Figures 5.13 9 5.18. ‘The plots have been grouped according t the observed stability ofthe tunnel bores. Schemes I and 4 were principally stable scheme 3 was subject to large horizontal ground stresses and scheme illustrated pipetine buoyancy. To explain the behaviour of the various examples i is necessary to consider the charcteristies ofthe ‘packing material used inthe joints. Figure S.16 presents the results of series of uniaxial compression teston dry Medium Density Fibreboatd (MDF) which hasbeer used throughout the eld work. Thechoice of picking material was hase onthe recommendations of Miligan and Ripley (1989) Samples of Imm thick MDF andthe MDF-pressue cell composite have ben subjected to twenty eyes of applied stress at 1,20, 30, $0, 30, 10 Nin" intesiies. “The plots show the responses on cycles 1, §, 10 and 20, Examination of the results reveal several pints of intrest the packer permanently deformed as soon ast loaded and unloaded once: afer the fist eyole the materia illustrates litle change in is stessstrain characteristics permanent deformation unless the sess magnitude is increase; subsequent cyeling at reduced stress levels exhibits similar response to tata the macimum intensity, Ie therefore important, although generally impracticable, when madelling the stress lstrbution in joint to consider the previous stres istry ee Peat Tie foapsey wiecage (aseey miuiace Figure $13 Relationship between measured joint angle and pressure distribution; schemes Land 4 5.20 [ ley HoNaR oaK 3 oy a (a1 HONOR 8K isa Figure 3.14. Relationship between measured joint angle and pressure distributio scheme 3. 521 LA Bo 8 g 8 8 & Figure 5.16 Uniaxial stitness tess on joint packer material, 523 “The plots of Figures 5.134, d, 5.144 and 5.150, c,d ate taken at raximum angular deviations and ilustat close agreement between the measured stress distibutons and the postions of maimnam compression derived from the joint movement nicstrs, ‘This because the previous ses history ofthe packer has neglgile eect when the packer i undergoing maximum compression. Localisation of stress cleatly demonstrated even with smaderatedefetion sngles of 0.1". Atsmullermisaignment typically 0.025", Figure .148, the relationship between angular measurement and stress distribution tends to break down ‘because the permanent compaction ofthe packer material reolts in reisibution of loa rom the analysis of joint zaps the maximum compression is expected 0 cur at 6" below ‘he ett springing position, Inspection of the previo tees hitryof the jit, Figure 5.17, and consideration of the stresses imposed by the thustring (not showin the Figure), indieate ‘atthe lowers pressure cells have registered large stresses prior tochainage 10.62m resulting in the major part ofthe loud being transfered though the tp left hand quadrant. It will be ‘ote tat the boom pressure pad in scheme 3 indiates leper compressive stres than ‘would be expected fom the general distributions, This may be the result of spoil being ped inthe boxiomof the joint I general stess levels in schemes | 104 were smaller than 20N mm? and id not precipitate pipe damage. Onscheme 5 the masimumangular deviation, Figure 5.152, oecuredunder pata load conditions (ie. closing the intrjack). Although the localisation of stress was severe the sures levels were small, Atlarge jacking loads ie interac engaged) eased crushing ‘of numberof joins was observed at chainages 110m and 130mm, The damsge wis sustained in sandanl joints ited with 80mm wide by 18mm thick plywood packer, Sireses inthe ‘nsrumented joins. containing 1mm wide by [2mm thick MDF to sasfytherequiremens ofthe research, peaked a $2 10 S8N/mm and did ot inflict damage. Comparison ofthe theoretical maximum jacking loads forthe diferent joint configurations using the design approach of section $.4 with B= 47 suggests thatthe MDF joint should carry 135% more Jad than the plywood whichis probably the reason the instrumented pipe survived Cpr: Pele Figure 8.17 Variation in the reat joint presure distribution obtained from seheme 3. An important pint wo note from Figure 5.15 i the possible fitigue effect in pipes close 1 inverjacks due tthe increased number of large Intensity stress cycles. This effec i dependent upon the contnictn’s interpretation of inerack usage ‘4 Predicting pipe joint behaviour Results connected with joint behaviour seem consistent with expectations and mexsured small deviations fom Tine and level along the pipe sting, "The elfets are essentially ‘geometrical and related mainly 4 the accuricy of the drive, and ate osherwise largely independent of the nature ofthe ground, Exact correlation of joint angles and contact stesses iscomplicated bythe nonclinear and tess orytelated behaviour ofthe Fucking material as discussed in section 5.3. However the site instruments enable total joint loads 10 be 5.25 cp Ppt tne accurately determined and packer material propesties which take account of the previous stress history ofthe join whe used. On this ass comparisons between measured and hack analysed sreses using the Austatian linear stess model, described in Figure 2.7, are presented in Figures $.1910 5.20, The appropriate packer stiffness abelled E, in he figures and the permanent pucker compression is obtained hy pajctng the stifles ine hick to eww stress. Curves linking maxeyo, and manoy,/2R have been proce! from the tables of Marks (1978), Figure $18, 1 packerone a packer oer aus ever/ar/ am Figure S18, ‘The comparisons show thatthe Australian approachcan adequately match the measutel stresses within the limits ofthe uncertainty ofthe data, and could be used t define allowable jacking loads on pipes for specified misalignment angles on the bass of the strength ofthe ‘concrete and te sifiness ofthe packer defined ever a Sultble unload-eload stress range BACK ANALYSSS °° CuroaiTe qe = a j i- i; # | 7 = oooh mane @ BACK ANALYSIS. | | _*T-aeapeane yy | oumna | i ae 4 ee ®) — Figure5.19 Comparison of measured joint stress distribution wo predicted using a linear stress model based on the Australian Concrete Pipe Association methed. (a) scheme 1 ‘ear joint subject toa f, value of 0.44", and (b) scheme 3 rear joint subject toa f value eras" 327 BACK ANALYSIS. | (Nima) 8 | : ee (a) | © om Beta 047 DES FigureS.20 Comparison of measured jointstress distribution to predicted using a linear stress model based on the Australian Conerete Pipe Association methd. (a) scheme 4 rear joint subject toa f. value of 0.28", and (b) seheme 8 rear joint suhyect to a P value of 0.87" and pipe damage. 5.28 5 Design approach In adopting the Australi CPA mel as a bass for evaluating allowable pipe end loads i is frst mvessary to define appropriate limits Rr the various parameters 5.8.1 Permissible joint face strength The field results have demonstrated that the steel handed joint detail ean sustain average steesses over a cell area of 200100hmm equivalent w the characteristic eae strength of the concrete, The localisation of stress is generally severe withthe peak stress pd reducing. ‘This suggests that the joint face strength limits of BS $911:Part 120 may be applicable. For design purposes two values of matximem permissible joint stress are eonsera- OE. which isthe generally acepted maximum beating sess in enforced concrete under well eonfined ‘canditinns and the lower iit of TONimm? from the joint face strength ts 155.2 Packer material properties Medium Density Fibeboard was used inthe instrumented joins ofall Five schemes. Tests ‘on other common construction materials were cried out by Ripley (1989). Exterior Grade Plywood hasbeen selected for comparative purposes because itis currently the mest popular material with pipe jacking contractors, sittness Figure 5.21 compares init sess and joint compression froma single cellom scheme 3 with the laboratory compression tests. The fiek! dats show that Behaviour tends that of procompressed material. For design purposes « linearisation of the reloading stiffness for ‘pre-compressed material, such as that labelled ED in Figure $.21 would be the safest. The value shouldbe determined from packer compression tess using a maximum stess intensity ‘of 0.8 Using this appoch vilues of SSUN mm and TOON mm ae obained for MDF and plywond respectively ys Petia afr Figure $21 Insitu stiffness evaluation using data from scheme 3 and suggested design stiness. Permanent packer deformation ‘The permanent packer deformation at these high sess levels approximates 10 0.3 times the ‘original thickness for MDF and 0.45 times the original thickness for plyweod, 55.3 Positioning in joint Ideally the packer should he made as wide as posible to maximise the avilable load are, However local bursting effects and large shear forces in joints (values up to 25 tonnes are indieatedin Chapter 7) can cause join spalling. To minimise the rsk of supeeiciat palling, itis suggested that the packer is place centrally in he join with a minimem edge distince esuival 1 10 the longitudinal reinforcement cover or 20mm which ever isthe smaller, For the pipe size range inthe esearch, typical wal thickness varied between 11Smm and 225mm With the joiat wall thickness some ISmm smaller The range is clearly large although variations in thickness for the product ranges of a single manufacturer are much more comparable. Two packer widths ate considered in this parametric study, 10mm 0 suit the research and 8mm which was frequently used by the participating contractors, 530 Acypical set of calculations andthe resulting design charts are produced a: follows: From the Australian linear tess apprcch 10 a max, EB “Total force ever hatched atea fo nea distbution of stress ‘Width of element = ah Length of element Ro VER dh ‘Area of elemen Stress on element =u, tH} (RH) Force om element = WRF a “WY t ‘Toul force on area ABCEGD/ 20(h-H WER te Jetty Simianly force on atea DFEGD = [Sea sat epee peta tnt ree Therefore fore on ate ABCEFDA= yan ay (ff 26vR Ram - [nv an = 2H [erw +n [rPas Bem PHI 18H [we eo) sR (PP When F, Bem [a WH [Wow ( ore} Consider the ease of «1200mm internal diameter pipe with a 12mm MDF packer 100mm ‘wide positioned 20mm from the internal edge subject a range of angular deviacons 443 = 0.7 x 12 = 8.4mm (composite joint assumed to he similar to 12mm MDF) ED = SS0N/mm? 8 2 R : 4" Finite (eg) (om) (rom) (mm) (om) ey ou 613 70 620 107 7830) 02 306 70 620 414 4706, 03 208 70 620 S16 3302 oa. 153 720 620 S67 2485 os 123 20 620 397 1870) 075 2 20 o20 038 1039 10 ol 720 620 659 669 CCateulatons can he carried out fora ange of pipe sizes, joint tress limits and packer material types and dimensions, resulting indesign chars similar to those in Figures 522 « S.24 532 MAXIMUM JOINT LOAD (a) MAXIMUM JOINT LOAD 0) oo Pte Tar wes + SN BETA (eg © 1470 mm 19, oni | PACKING MATERIAL | 00mm wide sot Figure 5.22. Permissible pipe end Wnading at various angular misaligrments; 100mm wide Medium Density Fibreboard chp Poe tadtaat Take IPE Sie « s0Nmm2 PACRING MATERIAL om MOF stm MOF FD = SN MAXIMUM JOINT LOKD MAXIMUM JOINT LOAD 6) BETA op) Figure 8.23 Permisible pipe end ading at various angular Medium Density Fibreboard. ignments: 80mm wide capes i. a” 3 00 | 7] \ Wires ape Eatnd mer 7 s0NInne | PACKING MATERIAL TPE ae yamm PLYWOOD BETA Dea) = 700Nmm2 ble pipe end loading at various angular misaligements; 100mm wide Exterior Grade Plywood. 535 he 5 Pre atta Par Examination of the design charts provides some indication of the influences of the varios parameters usd inthe linear stress model. The distribution of ses he end of a Pipe isthe same irespectve of pipe diameter and packer width ‘The observed increases in sllowable jacking lad withdiferentpipe diameters and packer wihsresulfrom the increase in available end area, An 18% increase in allowable load is obtained if the pipe diameter is increased from 1200mm to. 180mm intemal diameter, the percentage increase being independent of Increasing the packer width from 80mm to 10h has less influence at larger values of . Comparisons at 0.1" and 0.5* produce increases in permissible jacking loads of 21% and 48 respectively when MDF is used. Packer thickness hey parameter. An increase fom 2mm to 18mm produces 238% increase in allowable loa at 0.1" and 89% a0 Inereases in conerete strength aso produce significant increases in load carrying capacity at larger B values, 185 a 0.1° and 4% at OS" {Use of plywood with ll oter parameters similar produces a 31% reduction in allowable 4oadat 0.1" misalignment and halves the capacity at 0.5" compared othe response with MDF. “The most sgnianr parameters for increasing allowable pipe end Ina t moderate risaignmentanglesare therefore packertype (ie, bothstiffnessandpermarent compaction, packer thickness and concrete strength, A hypothetical cases presented inFigure 525. The ‘material stness is half that of MDF, with an original thickness of 5mm and 805 recovery under maximum stess eyeing, The maximum joint fae stress hs been limited 10 OS, Sincethe esting sss disribution wl probably have departed significant rmmthe highly concentrated joint face strength estconfiguration. Significant improvemens in oad earring ‘apacityoverthat givenby 18mm MDF, which thebestof anumberof wood based products, isillarated. Clearly the packer should be considered a part of the pipe an fined to tat the pipe works, There is obviously scope forthe development of Bete, possibly polymer based, materials 5.36 capers petted tant \ oeormeia, =m \ omnia Eco 4, =s0Ninme 3 | PACKING NATERIAL i — jo of > BETA Don Figure $25 Permissible pipe end loading at various angular misligements; 100mm wide idealised packing material. Existing data to verify the suitability of the proposed design approach i scarce, The ar Field data from scheme Sis the only suituhe rect comparison. The design vl fr is 207 tonnes whereas a maximum measured value of some 380 tonnes was ultimately sustained hy the pipe albeit with local compressive failure, The appronch appears to have an adequate ierent factor of safety a appoimately 18 ane Se neta hae CHAPTER 6: SOIL-PIPE INTERFACE BEHAVIOUR 6.1 Introduction ‘The provision of sufcient jacking capacity i largely based on previous experience from scheme in similar ground conditions. In genera, contactors desge jacking systems ‘with n excess capacity of 5-305 greater than the expected load tallow for misalignments ‘orunexpected ground conditions. To improve the design situation, it is necessary to review the parameters which influence jacking loud and rtionatse them in relation to theoretical ‘methods and laboratory test data, fr preiting ground pressures on pipes and appropriate friction coeticens. Comparison of theoretical values with actual jacking loads is essential to validate any design approach he jacking fod depends upon the force required to push the shield int the excavation, commonly referred 1a fae resistance, and the fection resistance along the length ofthe Pipeline. The amount of resistance encountered atthe face depends upon round conditions andthe measures eure to suppor the ace. tis predominantly elated tthe edge ertng resistance ofthe jacking shiek! und overall size is important as larger dimensions produce larger fae resistances. Currently friction resistance can only beestimated within arg limits ‘Table 22. The variability within basic soil types throws the credibly or “design back ‘no the individu comratoes experience “Thischaptr initially considers the tot jacking resistance For each ofthe instrumented schemes. A detailed examination of the sil pipe interaction process flows with two sts of parameters being investigated: ground related including soi type, ere coment and rading: and construction related including pipeline misalignment, stoppage timeand the we of lubricants on 6.2 Basie data 621 Jacking records Progress and joking fore recon fo the five schemes ae ilsteatedin Figures 6.140 6.5. Te length drive pleted along the horizontal axis does ot nce he length of shiek lor TBM which has heen assumed to provide x constant face resistance 38 measured off the vertical ans. The average fietonal sistance (acing ote divides hy pipeline length) has been added othe wraps. 'A prominent feature of the jacking recrd for scheme 1, Figure 6.1, isthe change in average jacking resistance at chainage 35m fom uniform rte of 7.2KNim to 29.8kN/m “The change in slope comespond to a significant change in weather fom Bn dry conditions to torent rain throughout he remainder he contact, whichappears wo sve affected the round eonditions tothe depth ofthe pipe jack, Genera progress of én per shit was Achieved. The jacking force increase from an ntl face resistance of 120kN ta maximum ‘of T3KOKN, The effect of overnight and weekend stoppages resulted in increased restart forces picaly 200KN greater than atthe end ofthe previous push. The weekend stoppage ofthe 18/19 August was preceded by tunne! Moding duet fractured war main withthe restart forces some 4O0KN larger and the post jacking resistance remaining significantly react ‘The jacking recond for scheme 2, Figure 6.2 suggest thatthe main jcking resistance ‘ceursat the tunel tice, Extending the average jacking resistance tne aca ests in an intercept of 9SOKN. This high face resistance i the result of strong to sl gly weathered ‘mudstone being encountered in the pipe jack invert. The erate nature ofthe jacking load prohably reflects the variation in the extent of trimming of the exeavation, This parameter could not be adequately monitored during this scheme, Progress was relatively slow at Sm per shifts direct resultof the ficult of excavating the mudstone, The low nite of advance and in putiolar the eleven day stoppage while negotiating the disused mine shatt do not 62 Chaput Sanpete Baer appear to have cauned larger restr jacking foees. The average increas in ekg resistance ‘was estimated as RkN/m forthe mustonesetion with larger value of SAAN m through the boulder ety. ‘The tern of jcking lad build up for scheme 3, Figure 6.shows relatively uniform increase of S4kNim and a fae resistance of 300KN, The blocky nature of the ground upon «excavation made cont of shield trimming dificult, The peaky time dependent response is panicularly pronounced with each posh illustating an increase in resistnce upon restart followed by 4 rpid decay 1 the average jacking resistance line. The magnitude of the scillatons increase with increasing drive length, ‘The contractor changed from single to double shit working 10 reduce the length of stoppages after @ 68% incease in jacking resistance after he weekend stoppage of 16/17 March, The continuous working phase resulled {in ite change in he rato of peak to average jacking resistance suggesting that substantial par of the time ctr” eect occurs within sort period of time, This effects considered in greater detail in section 6.42 “The jacking records for scheme 4 are presented in Figure 6. This scheme provided the Fist opportunity tv monitor the effects of changes in the extent of shield wimming and the use of fabrication, Both ofthese fctrs will be considered in greater detail ater inthe chapter. Prioe lubricating, the jacking fore increased ata uniform rate of23.1KNim. The well defined peaks at various positions along the drive ae the result of increases in face resistnce caused by changes inthe extent shield trimming. ally the miner was wimming approximately 20mm wit he shield t minimise overbreak, Upon requesta numberof pushes were carried out with reduced Uistances. The resulting eye reduction in cking resistance prompted the miner 0 revert generally excnvating tthe outside iamter of the shiek ‘except when changes in alignment were necessiry, and shield resistance wasrequiredwreset the steering jacks. The ipl sions of surface settlement were negligible given the line of the jack through a greenfield site. Styppages ha itl effect om the restart jacking kas over the non-lubricated section 63 ubvicaion was inrluced at drive length of 82m, For the 1 80 days of injection, with a tunnel advance of Sm per shift, the bentonite lubricant hd litle eft onthe jacking resistnce, However between chaimiges 90-102m the jacking resistance dropped by 15%. Coe inspection of the tunel face lo indicates that boulder clay was preset inthe invert Atchinage 102m running sand entered the btom of the face withthe boulder clay horizon ‘moving into the top hal ofthe face. The running sand hus probably increased the Face resistance. The peaky response i key tobe the result ofthe miner burying the shied hood Jo the boulder clay to prevent the shied! fom tipping, The resulting inerease in jacking resistance approches the same value that woul! Rave heen obtained by exteding the average rate of increase prior to lricating to the end ehainage. This suggests that the lbriation| has omy been patil effective, Scheme S was.a mechanised dive the jacking records are included as Figure 6.5. The scheme is characterised hy a high fice resistance of ISOOKN and a low werage frictional resistnce of 17KNim. The lange rate of increase over the intial 25m result ofthe ground being un-lubrcited because of eonceen over blowout 80 lose 1 the jacking pit. Use of ‘bentonite lubrication significantly reduces the rte of increase of the total jacking haads. The Ly appears 0 be x function ofthe rate of pipe installation with the nwer resistances, piniculrly between chainages. 30m and 60m corresponding w an increase in daily productivity fom 17.5m w 278m, Periods of inaetvity and overnight stoppages result in increases in estat jacking forces lange a8 256, ‘The eects of pipstne misalignment ate clearly scen from the plot with an overall nctease of 6h caused by the reverse bend between chainages 110m and 130m, Activation ofthe First inerjack prsitioned Sm behind the chine has produced a handed response the tcking record. The lower hound i the closing upof the rear ofthe pipeline int the interjack and the upper values are the combined resistance ofthe tron and rear sections. An intresting feature isthe much igher resistance between machine and intrjack, ‘This evuld be the result of less effective nbriction as the tunnel length increases and pressure losses along the injection sysiern become more significant = oe 2 | SEEe os é ae 4 | Po cetera | re —— = BOLTON z Ww oe ta 3) —— 8° oe ‘ a os _| a oe oo + a | LENGTH OF DRIVE (m) Figure 6.1 Jacking records for scheme 1. 65 24 20 18 16 2 10 DAYS ELAPSED 264 NEWCASTLE 24 22 20+— 19. 16. 12 JACKING FORCE (iN) (Thousands) oa 02 0.0 +—-— ee LENGTH OF DRIVE (m) Figure 6.2 Jacking records for scheme 2. 66 hee SslPp aer Rbeie 3 mo ae DAYS ELAPSED i i 404 JACKING FORCE (kN) (Thousands) © 20 LENGTH OF DRIVE (m) Figure 63 Jacking records for seheme 3. 67 Cher 26 za 2 20 16 “4 10 ire jet ° DAYS ELAPSED. 20 ° as Pai LENGTH OF DRIVE (m) Sap ef Basie a | © 20 120 ABBEY VILLAGE 30} 25 20 (Thousands) REE eS JACKING FORCE kN) Lear Figure 64. Jacking records for seheme 4 OF DRIVE (rm) 68 Sti ei eer CHELTENHAM Figure 65 Jacking records for scheme 8. grt Sipe ere er 6.2.2 Pipe jacking loads Jacking records are generally the only iaormation on gsound telat kading thatthe ‘contractor has a his disposal. ‘They allow values of nominal face resistnce and average Pipeline friction 49 be determined in diferent ground conditions. The values from the instrumented schemes are included as Table 6.1. Fae resistance can be seen to account For 4 significant proportion ofthe total load. In cohesive round the value sates from 6-9% although there is probably considerable variability in the value for scheme 3 because ofthe highly issued state ofthe exeavated materia. Variability in face reistnce is inherent in allhand drives and is clearly illustrated in scheme 4 were the value varies from 396 when the shield isomly in contaet withthe base ofthe bre to 25% when rimming approximately 20mm allaround the excavation. The hrge valve of 27% forthe machine drive isa funetion of the Slurry pressure used 10 support the fice, while the wile of 439% from the rock drive i function ofthe mudstoe inthe invert f Mensur ce | Morand aserage | mig (43) sctene ww fe | am far | ar To Te loa 0 ee es Wee [oe | z 0 we [x as [2s 3 x00 ox [see [re | san eee wo fa [7 (ss | owe = Value based on 40m monitored length, Total dive length 100m " Unubricated Wer Sand ‘Table 6.1 Average face resistance and pipeline friction. Chet Stipe bine Comparison of the average tition values with typical freton lint, Crag (1983), show thatthe measured valuesare below oratthe lower limit the generallyacceped anges. ‘This isa function ofthe competent nature ofthe cohesive ground in Schemes 103, Figure 66, and the well controlled alignment ofthe drives. For the manly aon-ooesive drives of schemes 4and S small support pressures, 0, (se section 2.3, Table 6.2 and assume gs 40" andthe watertbleis Im above thew othe pie in seheme S) of 10kpaaind 14kPa respectively were required to prevent collapse ofthe material onto the pipes. Inscheme 4 his was provided hy capillary suction within the sity material leading 1 a relatively stable fore for up five aysafterexeavation. {nthe lose snd and gravel of scheme S the support was provided by lubricating the line with pressurised hentonite slr at SOKPa. UNSTABLE Figure 66 Stability of bores. out Chaput Sarge bine “Tosimprove the very approximate average resistances traditionally adapted in designing Pipe jacks tis necessary t gin an understanding ofthe contact pressures hetween the pipes and the soil during jacking. The contact stress cells nd poe pressure probesin the instument Pipe wall allow such measurements 1 be made, 6223 Local interface stresses mobi ed during jacking Seheme 1 ‘The contact stress cells throughout scheme 1 illstte zero cit, Figures 6.7 and 68, {due to moisture ingress. The instrumented pipe was positioned three pipes behind the shield resulting inthe excavation being generally one day old when the pipe passedhnough. Contact ‘between the pipe and soil was ony recorded on the bxtom ofthe pipe thnayghout the drive ‘which is consistent with the overbreak in if soils at low cower depth remaining open and the pipeslidngalongthe hase of the open bore. The total radial stes varied tween -1204Pa and 7S0KPa. These values were for extremely short periods wih the majority of pushes Iying between O250KPa, The larger values are probably the result of local cables pressing onto the cell. The negative values imply suction during the shearing process and generally correspond the locations of negative pote pressures recorded during the wet site conditions, Figure 6. Shear stresses varied hetween -20KPa and 480K, The negive stress 6 an isolated value ands possibly du to recoil atthe end of a push. The majority of shea stresses were betwen kPa and 150K The pore pressures during the dry weather conditions are generally smaller than the ‘changes in total sess which suggests thatthe ground atthe imterface may be unsaturated ‘even though the original material was saturated andlor the high permeability ofthe concrete allows a partially deine condition 4 exist during the jacking process. By contest the behaviour during the wet conditions indicates closer parity between tra stess and pore pressure suggesting undrained haviour; the material having regained is saturated state and the local drainage pth via the pipe interface being greatly diminished by the excess water. 612 Figure 6.7 Variation in total radial stress during scheme I. Se eR 4 oGbEGE aital Cher Sati nice i remESaEIEH 7S hala sadegall aged PES Figure 6.9 Variation in pore water pressure during scheme 1. Seheme2 The variations in lea incerice esses for scheme 2 ate included ss Figures 6.1019 6.12, The anticipated large stand up time of the overbreak ed toa reduced instrument set being concentrated inthe pipe inven. The insiramented pipe was inserted 22 Stn being the ‘hic, passing through the excavation 16 days afer the shel. appears rom the pots ha the pipe was subjected 0 longitudinal rotation withthe font contact tres cell tractng Sificantly larger stresses, prior co it being damaged Total radial stresses approached [75kPa with coresponding friction sresses of 160KPa, However measurements ofthe exc postoning of the cls inthe cored holes showed tha the front cell was Imm proud of the pipe surface and inclined towards the direction of movement while the rear ell was recessed by the same amount and inclined away from the diretion of movement. The Sif nature of the ground has resulted in significant over and under-registation of load. The pore water 6s ctr Sat ni eae pressure during pushes indicate suctions of rypically SOKPa in the font cel. Large positive pore pressures of short duration were measured in the rear cell, the maximum valve approaching 700KPa at chainage 6.1m which was the positon were the lend contatstess cell was damaged. The lage pressure caused the gli line sound the probe to fil. Deep cass the contact sess cell suggest that & high spot may have res tein ‘concentrated loading. thd Figure 6.10 Variation in total radial stress during scheme 2. po Figure 6.11 Variation in interface shear stress during seheme 2 =F a =o] it] }— aol | E 2 ' Figure 6.12 Variation n pore water pressure during scheme 2 6.16 oper Sarge tne Scheme 3 “The local imerface stresses in the high plasty lay of scheme 3 ae presented in Figures 6.13t06.15. Theinstrumented pipe passed through he intial exevaton 9 days afer ‘he shield, High tral ground stresses exacerbated by horizontal deviations of he pipeline Figure 54, were sufficient wo break the glue lines onthe contact stress ces a tunnel ans “Therighteell was displaced under a ail oa ess of 490KPa nda shearstees of H1OKPa, approximately Im ino the drive, The lft cell moved at chainage 4.5m when subjected toa sail otal seesso 6S04Pa andasheastressof M45KPa, Iwill be noted th he cells became vealed in the oles at depths of 2mm and 1Omm respectively from the pipe outer surface and continued to record rail sresses up to 1SOKPa and complementary seat stresses upto SUKPa, The respons af he hotlom contact tess ell was faulty an this would have been Picked up during calibration had scheme 3 no followed hard on the heels of scheme 2 Unfortunately the cll was subsequently crushed in service and could notte recalbrated or inspected to establish the reason fr its sbnormal response, ‘The top cell was displaced by Simm which probably occurred at chainage 10m. ‘The maximum vertical ta rai stress ‘was 4S0KPa with a peak shear sess of 145kPa, The high failure rate of ue lines on this ‘ite was disapointing but certainly saved the enh pressure cells fom Beng crushed. The method of fixity was maid for schemes 4 and § 10 prevent futher premature movement. The pore pressure data when considered together with the toll tres data provides xtemely limited information on the effective stress behaviour, The apparent dormant response ofthe top probe suggests that the cable was damage during installation while he ‘com and lett ais probes were displaced during the drive. The oly elu data were obtained from the right axis instrument; the lrg pressures typically 400-800 imply that the pine could be sujcc to squering fom the sides. Its apparent rote plots that zo pore pressures were recorded during the fist 3m ofthe drive. Visual observations suggest thatthe sutface layer of the clay around the rel entrance may have dried oot Chet Sr per iw Figure 6.13 Variation in total radial stress during scheme 3. 68 sensTesiey nth pet ie h : hull ll Che Stipe Basie A CSE REESE 6.20 hp Sa pe me ee Scheme $ Figures 6.16 0 6.18 las the data recorded during scheme 4. The major stesses cecurted on the tase ofthe pipe although limited collapse ofthe sly sand around the pipe resulted in misalignment induced stesss being recorded atthe crown atl axis postions. “The plotscleatly show dat peak values of wal dil and interface shearstreses were obtained cover short fengths of the drive with Tower overall average for he tts ngth. The peak use values reached 300KPa and 160KPa respectively. The stesses at pipe ais and crown ‘were generally smaller peakingat )SOKPA and 8OKPa, Clserexaminatonof pots indicates that the positions of peak stresses along the base agree well wih the postions of maximum angular misalignment. Variation in ground conditions may al ale the interface stresses tong the tunel length. The detailed unre! fice log of Figure 4.8 indicates that the ground type atthe base ofthe bore was predominenly sity sand, The presence ofa boulder ctiy bund between chainagesSm-61m appears to hive cused isolated peak rl stresses inthe top and let hand contaet stress eels. The alignment profiles of Figure S5 indicate thatthe Pipe was psig througha vertical tough wth the position maximum devaionat chainage 60m. ‘The stifer response of the boulder clay compared to the lxse sand hs allowed the seneration of lange eal interface stresses. Te aia ubrication over hissecton of tunnel appearstwhavereducedthe ratio ofshearstesstoradiatstress, Furter deaf misalignment and lubrication effets are eavere in sections 6.4.1 and 6.4.3, Pore pressures ang the base ‘ofthe exeavation were peaky reaching maximum of 190KPa, Pore pressures in the disturbed material elsewhere around the pipe were small, The injestion of Iubricant it the woid does notappear to he hd a masked efecto the ail and shear stess profiles Fr the majority of the drive, The absence of lange pore pressures along the hase of the bore after lubricant injection suggests hat he probe may have wore down, 621 non ses ROLSTESI, OM SS a oT stress during scheme 4. sen eRe ote Gt Stipe ese hg rE me EES Figure 6.18 Variation in pore water pressure during scheme 4. 6.28 ener SP ne Ber Scheme $ The interface stresses along the length of tunel containing the severe pipeline ‘isaignmentsandpipe damage, chainage Mimo 132m of Figure 5.6, areprsentedin Figures 6.19 19621, Seeses were mobilised when he whole pine sting was jacked from the trast pitand also when the intrjack, which was positioned one pipe in front of the instrumented pipe. wasactivated, An immediate conclusion that can be drawn fromthe plots is that stresses sppear to be independent ofthe location of thus, This is because the tat load passing through the pipe ziven the close proximity of the interjack, must be sinilar and tunnel alignment is leary dictating the interface ses profiles, Maximum tora rail stresses of 300-350KPa were measured atthe tunnel axs with the average stress Mucuatng about the ‘SOkPavalue, Pore pressures were generally very similar suggesting that presursed bentonite slurry had formed a stabilised zone of wil around the pipe preventing the soilfrm developing Jnege effective contact stresses. The elfectvenes ofthe layer of bentonite el between the pipe and sols demonstrate in Figure 6.20 with shear steses typically below SkPu except atihe postions of maximum misalignment changes 110mand 1302 where tresses peaked 1 LSKPa, These values are supported by the average pipeline shear stress value of 3.6kPa ‘Table 6.1, obtained from the jacking record. Surprisingly there is no elee indication of a reversal of shear stress when tvatng te ierek, Hisnotknown whether this isa function ‘ofthe movement of pressurised uid fom the high pressure zones. The suggested buoyancy ‘ofthe pipeline based on tunnel surveys, Section 5.2.1, 8 substantiated by te slightly larger total ail and shear stress vlues atthe fop ofthe pipe co Soiree to 8) Main Rams ) Interjack activated Figure 6.19 Variation in total radial stress (ch, 913m to 131.8) om scheme S 626 punt Sipe ri eer i a z 7 ™“ 8 eee [. is i ! 2) Main rams by Interjack activated Figure 6.20 Variatio 627 Interface shear stress (ch. 913m to 131.8m) om scheme 5. Chet Stipe aie ts PNW Wh | Figure 621 Variation in pore water pressure (ch. 91.3m to 131.8m) on scheme 5. ‘A summary ofthe peak interface stesses hs been reproduced in Ta I: 6.2. The major ‘contct stresses inthe nonlubricate drives were mobilised at the bottom ofthe pipes ps from scheme 3 though henvily overeonsolidated London clay during which lateral stesses up to 6S0KPA were sutticien to damage the instruments on the tunnel acs. In the fully lubricated dive scheme Sthepressrise lubricant his causedthepipelineto oat resulting insimilar values athe top and bottom ofthe pipes. The vious plots clearly show that peak ‘values of aia. fretional and pore water pressures were obtained over shortlengths,Im-2m, of the drives with a lower everal average wale for the total length, Te fe skin friction values are nearly wo orders of magnitude larger thaa the average fiction values in Table 2.2. The generation of significant positive pocepresswres dating jacking ako demonstrate. “To provide some fee! forthe magnitude of radial stresses the values are compared with fll 628 hae Salpeter ine ‘overburden loading, short term Terzaghi loading and elastic pipe self weight streses bused ‘on the formula of Roark and Young (1976) foe the contact width between a solid elastic cylinder resting in a cylindrical cavity: b= Lo(Purka*ce? k= DDs(D-D.) omtact force per unit length 1D, = internal diameter ofthe cavity = extemal diameter of the pipe C= (len, Eyton VE, E, = elastic modulo the sol FE =lastic modules of he concrete pipe 1 Poisson's rains as for E Scheme li prea fee cae al te ten feta one am [on [oreo | rn xn [ion _[ oo 2 [2 [u [om [a [or [or [aso [ams [im 3 a_[uww| os [4 [oa [oa [asa [2am [aa 4 wr [iw [a for for [or [rao [ire 5 w |e [wm |» [oa [oa [ax [ras [rao rapetics wed to determine rial sles pedis shvwn ae Steme | Aston | Ful Pe | Ridin) | Shear(aPe) |] Pore Preware Cra” | owetaren | Sat or) eden | Gy | weit ‘Terasht| ha) “Top axe bcm | Top ani boro | Top ani eins ary) tea tm) | an 2. % si 1 amet 3 [550 650 ist 80 2380 as be fim sm sm [oa wo [wo 0 ww ‘Table 6.2. Maximum local interface stresses 629 Cher Sate Haine Comparison ofthe theoretical and measured values demonstates tht induced earth lowding ring jacking can he larger than the most severe design case of fll overburden, which may have implications for pipe design, Adopting fl overburden foe predicting icing resistance would however seriously overestimate the equied jacking capacity. The more generally cceptedTerzaghi loading which s the bess of most European design appmnaches results in ‘much smllershor term ln intensites which approximate more cisely toe average radial Joading, but significantly underestimate the peak radial bading cowed by pipeline misalignment. In adopting the Terzaghi approach and using short term material properties many pipe jacks through cohesive ground, except in very soft clays, invariably indicate negative radial loading, ue self supporting, whichis consistent withthe stbility checks in section 6.22, In such cases pipe self weight contact esses may bean appropriate estimate ‘of ial loading. Contact his heen assumed to ccur on the bottom ofthe pipes in schemes 1 10 4 although iti suspected thatthe pipes in scheme 3 were suhjeced to horizontal syueezing, In scheme S contact was assumed 1 oecur atthe crown duet pipe buoyancy. “The lca interface plots vary above and below the el weight valves suggesing that he local ‘variation inthe excavated surface andthe changes in pipe end lad positions due to pipeline inisalignment result in uli or down trust ofthe pipes. t wil be seen in ection 6.3.2 that ‘the pipe self weight model in well contlled deve can provide aoe sppximation so total jacking resistance 6.24 Detaled response of individual pushes ‘The compressed nature ofthe interface stress plots makes it ficult evaluate the ‘round behaviour duringa singe push. Typical pushes have been selected om the ist four schemes and ate presente in Figures 6.22 106.30, A numberof cbservations can he made 8) There is excellent agreement between the responses of radial, shear and pore pressure profiles during pushes providing increased confidence inthe validity ofthe readings, 630 by) Steeses vary rapidly over short distnces of es thin Sum and are probaly function of foe variations in excavation profile. The dives tough cohesive ground exhibit apd peak and tough (typically zero tess) responses while a much smoother Muctation is ppatentin the non-coesive atti ©) Sess path plots indicate frictional elationships forthe nonscobesive ives of schemes and 4 Surprisingly the total rail sess pots fr the London clay t pipes aso suggest A ition material response (Ihe effective stress plots could not he constructed). The stress paths for scheme Tae however more emiie which may be fonetion ofthe High stone content ‘of the boulder clay. During the dry weather fictional response & demonstrated with an Apparent change an undraned constant volume Behaviour afer the seriad of tunnel inundation, Figure 6.22 Detailed response of pushes hetween chainages 239m and 415m on seheme I chert [some Stipe ie eee \ Figure 6.23 Detailed response of pushes between chuinages 13.09m and 14.22m on scheme I. les baoggeeae Figure 6.24 Detailed response of pushes between chalnages 44.39m and 46.65m on scheme f Figure 6.28 Detailed response of pushes between chainages 4.49m and 5,02m on scheme 3. Ctr Slept bine : al (fms Figure 6.27 Detaled response of pushes between chainages 7.75m and 9.111 on scheme 3 ~ 7 i. a Rehan at Figure 6.28 Detaled response of pushes between chainages 16.8Sm and 17.86m on om Chr Stipe ane Figure 6.29 Detaled response of pushes between chainages 44.97m and 45.56m on i Figure 6.30 Detailed response of pushes between chainages 62.13m and 64.36m on scheme sd. con Seip asia 6.3 Ground related factors 63.4 Soilepipe frition eveticents “Toestblish whether ther sa fundamental difference in interface behaviour between cohesive and mon-cohesive sis plas of shear tess aginst wal ad effestive ral tess -65ue site value -7he TE9SS 9=95-02=93Nimm’ < 217He site value 245ye TE9S6 9=95412.2-01=21.6Nimm’ <> S0Sue site value SSOje TEOT? @=954 122401 1.8 Nimm’ <= S00je site valve S254 4 -14ye site value -17He ‘Once again there i good agreement hetween the calculated and measured values which ‘demonstrates he sutaity of simple elastic assumption or analytical purpses athe high aquatty ofthe Fed dats. 1138 note thatthe agreement less gon when strains are smal andtersile, This likely tobe a function ofthe readings beng close w the Fit of sensitivity ‘ofthe instrument which was found toe around 124 under ahorstory conditions. Futhee work is needed to optimise the structural design of pipes, but its clear hit better packing materials couldallow more of the pscatiat strength pipes toe achieved when being jacked with small misalignment anges 716 hes lian Ramin CHAPTER 8: CONCLUSIONS AND RECOMMENDATIONS 8.1 Instrumentation One of the maa objectives ofthe research wa o develop suitable isitumentation and appropriate procedures for monitoring actual pipe jacks, without undue distupionw normal siteoperations. Instruments had he selected designed to operat inthe gressive tune ‘environment, hive minimal effect on the measured property and be stfcie tly accurate and simple to ealibate, Whete possible, advantage was ten of he reduced development costs ‘of using commercially available instruments. The pipe joint pressure cells, jack kad cells and Celesco displacement unit il into this category and performed well. The emaiing instruments were specifically designed and manufactured forthe pipe jacking research and -orwhich performed within specification, wih the exception he ground conver genes ind hada poor field performance record, Allo the equipment wis designed reas inconporation tnd retrieval from the permanent worksandsubsequent re-use. The only consectionsPetween {he instrumentation inthe tunnel and the surface hase computer which cenzlled the data acquisition were a power cable anda signal cable which were disconnected ad reconnected atthe same time asthe contractor's lighting cable All intrument readings were recorded on ‘time basis for coreation with a detailed lg of site activites. 82 Site work Dring the planning stige it wis expected tht a sufficient number of potential sites ‘would become available from which those approximating o previously determined optimum conditions could be selected. In practice, very few sites became available, an hen often a ‘very short notice, making incorporation of he research work int the contact more difficult aa tr ins andes than expected, Similar al site work was at the merey of delays and problems outside the contol of the research team. Finishing the work within the available imescale clearly illustrates the excellent cooperation provided by the various clients and contractors on each ‘ofthe schemes, The main variable turned out oe the ground type with dives through if -locil clay, weathered mudstone, sf plastic (London) lay, dense ine sity sand and loose sand and gravel, although it is reretble that none involved an excavation in sft cay. Ltwication was used during the Inter stages of scheme 4 and thoughout scheme 5. One expected variable, the ype of pipe joint, appears io have disappeared. thers esl of field experince or from the findings ofthe laboratory tests of Ripley (1989), the superiority ‘of but toin-wall joints whenever significant jacking forces are expected appears ohave been accepted and stel-banded pipes were specified forall schemes: 83 Pipe joint behaviour The transfrofload through pipe jninsisdominatdy ffets of angubrmisalignments between successive pipes. These missignments eccur due to variations fom exact ine and leveonce they have occured hey appear to change litle either de to applestion of jacking load or he passage of subsequent pipes, Thus once critical misalignments established, all Pipe joints passing that chainage willbe affected by it. Contol of tine end level is often ‘oorestat the start ofa drives tisis ao the location at which maximum pip loads wil occur later in the deve. giving eset severe pipe loading coitions. ‘Typical misalignment angles on nominally straight drives, within scifi tne and Tevel tolerances, were found to be mainly between 0.1° and 03, but could reach 0." Signtcant sess concentrations were observed in the joints (even when a packer of compressible materi such as fibreboard was used) and these correlate well withthe joint misalignments. Angular deviations of 0.1" confined loading to approximstely hal he end area f the pie, and 0.3 fo about a quater, the exact amount depending onpipe size, packer properties andthe ror load passing through the joint, Bearing resesin excess of S8N/mn ceceurred on one site witha jin deflection angle of 0.47" and large jackne fores. Accurate contol of angular deviations during constuction is therefore essential fe the succes completion of highly loaded pipe strings. ‘The establishment of actual agutar deviations achieved in practice allows reasonable specifications in terms of allowable angular ‘misalignments io be considered, instead of o in aditon to curet mits on fine aleve Ie would appear that the Det Swiss code of practice, Craig( 192), has git more oes correct with 20.24 lisiton angular deflection, when high axial lads ar expected. I should be borne in mind, that additional angles ean occur due to luck of pipe end squarenes, and in theory these could add together inthe worst case, ‘The field observaions and an end squareness audit cartied out at three of the UK pipe works has established that this eect could be ignored. The field data have also edt imple method for making decisions about corrective action base on minimising angular deflections, Milligan (1992) a presen this is being retained as confidential information by the industrial partners in the esearch In order to void damage to joints due to overstressing when misalignment has taken ploce itis important 10 estimate the sress concentrations that occur wih differen joint dletlecion angles. Asa first approximation the Concrete Pipe Association of Australia (1983) recommend that sess concentration of about 3 times the uniform jin stress should be expected, Thisisalsoahebasisofthe UK pipe manures practise peiying itn evenly steibutesend stress value. The site data however suggest that local stion of transfer suesses cam he much more severe withthe ae of thumb vale closer toa set 6 Exact contain of join angles joint stresses is complicated by the non-linear and suess-hisory elated behaviourofthe packing material. Howevercarefulback-analysis shows that the linear stress approach ofthe Australian CPA can adequately mach the measured streses and could be used to define allowable jacking Toads on pipes on the basis of pipe sie, pack properties, conrete srength and angular misalignment, Adopting such model asthe asistoradesignapproach equi safe init tobe defined forthe principal parameters, namely packer slfness and. thickness after permanent compaction, and allowable joint Dearing sess. The field data show thatthe stesc-strain behaviour of the packing material ‘tends towards that measuredin unoad-zeloul cycles i laboratory tests andthe coresponding high stfness gives more severe tess concentrations ud should therefore fe used in design ‘Consideration ofthe joim geometry ad the loealsation of force over sill areas suggests ‘that local bearing stresses between #04 ofthe characteristic concrete cue strength and the lower limit ofthe joint face strength test of BS $911:Part 120 can be tolerated. “The packing materia playsan important roein istibutingstessconcenrations caused by misaignment, It is therefore imperative that guidance on prefered material types, propeties and geometry are included in BS 5911:Par 120, Standardisation ofits use could be achieved by considering the packer as part ofthe pipe and fixing ita the pipe works. 844 Pipe barrel behaviour Asthe pipe sing “wriggls" through he ground the angular orientation ofthe postion ‘of maximum compression moves around the joint, Greater movements experienced inthe pipes immediately behind the shield which is consistent with expectations. Ifthe cenres of pressure are a the same angular position at bo ends of pipe the lod is tansmied essentially along one edge ofthe pips i they ae out of phase by 180 degrees. the pipe will ‘eloded across. dlagonal, The ite measurements have sos that pipe ne mialigzment generally takes the approximate form of large radius curves with the resting lad paths ‘within 245° of the single edge loading case. The extreme diagonal loadingcondition would only occur under exceptionally oor contol of the shield. ‘The overall betaviour of a pipe ‘under various field conditions show that sypical maximum stresses afe snout one thi of ‘he peak ointstess, Inabsolue term the valves are within the usually assumed near elastic range forconcree. Even under he exteme conditions of lading in scheme, n which lea compressive flue ina numberof jinsswasexperienced, he pipebareleompressve stresses ‘peaked at 20N/mm and tensie stresses did no exceed the tensile sirength ofthe concrete to produce eircumferemial cracks a4 Further work is needed t0 optimise the srvcoral design of pipes, bt it s clear that improved packing materials and/or better end reinforcement details could allow more of the potenti stength of pipes tobe achieved. Current reinforcement arrangements are generally based onempirca rules, with he main consideration being the provision of wficent hooped reinforcementto withsandlong term groundloads usually basedon trench lading conditions. “Te field daa on the oxher hand have shown that instalation lous ae considerable an he round loading regime during instalation is significant diferent to assumed long term conditions. Fortunately the evolved pie wal thicknesses are sufficiently age to ensure that longitudinal reinforcement is only needed for shrinkage and handling. requirements. Economies in pipe wall hickness wll however equite proper consideration being given 10 inwallaton loads. ‘The Fieldwork has provided a large hank of high qui data which can be used for this purpose 85 Pipe-sil interaction Datahavebeen collected onthe ineacton between pipes grounding resistance athe pipes neti in diffrent around conditions; nd constuction rato factors such 4stime and rate effets, pipeine missignment, andthe effectiveness of labret injection, allofwhichafect the tual jackingresstance. tiscler from the interface stessmeasurements thatthe very approximate average ressances ta jacking through differet ground conitions ‘radionlly doped in desig 1 Pie jock representa gross simplification ofan extremely complicated case of si-srutute interaction, “The interaction mechanisms ate highly dependent upon the short ter stability ofthe excavation. In general pipejuek shields re usually of slghly larger diameter than the ouside ofthe pipes. This “overbreak”consiertby reduces the conic between pipes and ground andhence the resistance 10 jacking, For if wilt at law cover depths the overbreak remains ‘pen athe pipes slide long the base of an open bore; eadings are thea only bused fom ‘he instruments in the bottom ofthe pipe. This happened through most of schemes I and 2 ss Scheme 3 was ata greater depth and in highly plastic and overconsolidved London clay ere some very high contact stresses were registred sufficient to eause damage to most of the sees cells quite carly inthe dive. It appears tar the pipes were "squeezed by high horizontal stresses, exacerbated by horizontal deviations ofthe pipeline, For the period for which he ces were operation the nterace resistance corelated very well with published ata From insrumer ‘riven piles and laboratory inerfce ring Shear tests. In scheme 4, some collapse ofthe silty sand onto the 1p ofthe pipe occured, while stesses onthe side walls varied fom zero to quite large wales dependent upon the rection and magnitude of pipe misalignments. The misalignments are therefore highly significant it their effects on total jacking force as well a6 local join sesses. Peak values at mi ength of pipe ‘typically occur over short tm 10 2m Tengths and correspond t0 postens of maximum

You might also like