INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of the International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE). The library is available here: https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands of papers published under the Auspices of the ISSMGE and maintained by the Innovation and Development Committee of ISSMGE. MAIN SESSION 4 SEANCE PLENIERE 4

DEEP EXCAVATIONS AND TUNNELLING IN SOFT GROUND EXCAVATIONS PROFONDES ET CONSTRUCTION DE TUNNELS EN TERRAINS DE FAIBLE RESISTANCE

Chairman/Président: 0. MORETTO (Argentinal; General Reporter/Rapporteur Général: R. B. PECK (U. S. A.)

Members of the Panel/Membres du Groupe de Discussion: J. ALBERRO (Mexico) M. ENDO (Japan) J. E. JENNINGS (South Africa) T . KUESEL (U. S. A.) W. H. WARD (England)

Chairman 0. MORETTO (Argentina) vided in two parts, in the first one our General Reporter will make a summary of his As all of you know the session of this morn­ well done work over the "State-of-the-Art" in ing is related to deep excavations and tun­ tunneling construction in soft ground. Follow neling in soft ground. This is a debut, ing it, will be corresponding sumaries of because the subject is discussed for the those members of the panel which have been first time in an International Conference, prepared. Contributions over tunnels, in­ as far as I have used the word "début" compa£ cluded the one of the chairman of this ses­ ing it with the slang utilized in the world sion, which I anticipate to you is caracte - of p^optacle, I would say that is a "world rized to be what in Argentina we call a début" in particular related to tunnels in "metido". The first part will end with an soft ground. interchange of ideas between the meraberE of the panel on the points that for disagreement Probably many of you who had the opportunity will need to be claeified. Finished thi6 to read the report of our General Reporter, first part there will be a 10 minutes break, when finishing your reading had reached to sharp, I make special mentions on the word the same conclusion of the speaker; the sis- sharp, because we will be very estrictily on tematization and ordering of the data are so that. well realized, that although becomes in dis­ agreement in many aspects with the final The second part will be develop in a similar evaluation, by the way of the enthusiasm one way to the first one. During this 15 minutes could let carried out, and classified the break, the table will receive questions about work in the following way "And from the both subjects, this means Tunneling and Open chaos becomes the light". Cut Excavation, which the General Relator, together with the Panel Members, will try Following the same road, I do not think I to answer. am exagerating to much by saying that with respect to tunnels, such a report will make Only is previewed the incoming of the audi­ a land mark in the soil mechanics history, ence in the measure of the remaining time it will be a "Before and After" but not only allows it. If there would be any possibility, that, it will be a "Before and After" appear surely we will be forced to make and involun ed just in the right time and here I can noT tary but strictly selection. Whatever be do less than praise without restrictions to done, we exhort all people who want to make the "Mexican Organizing Committee" for hiB any contribution, to do so by writting, fol­ vision in choosing the subject and for his lowing the prescribed procedure by the success in the appointment of the man who Organizing Committee. composed it. I think it would be petulant of my own to With the extraordinary increment in popu­ introduce you the General Reporter, who is lation and in véhiculés that have been under well known, in spite of that I pass the word going the greatest cities of the world the to Dr. Ralph B. Peck, Professor of the Uni­ subway conptruction either in tunneling or versity of Illinois in the United States and in open cuts excavations as well as other elected President of the International Socie type of tunneling lor utilities, is passing ty of Soil Mechanics and Foundation Engineer thorought, an uncommon period of prosperity ing. with special wideness and profit to the sub­ jects which to day atract our attention, from which the one related to excavations forms that we would call the classic section of our knowledge. For that reason nothing General Reporter R. B. peck ( u . s . a .) is better that thi6 up to date of our knowl^ edge corresponding to today’s session. Prof. R. B. Peek’s State-of-the-Art report appears on pp. 225 of the State-of-the-Art For it's development the session will be di- volume.

3II SEANCE PLENIERE 4 Chairman 0. M0RETT0 have all been hand-excavated. On six of the 11 San Francisco tunnel contracts, including all drives lo nger Thank you very much Dr. Peck for a signifi­ than 3,000 feet, mechanical excavators are being us ed. cative summary of the first part or your gen eral report. The machines have perform ed w ell in the dense sands The firBt contribution from the panelist, and the cohesive granular soils, and have successfu lly will be in charge of Mr. T. Kuesel, member negotiated short lengths of soft clay and weathered of the firm Parsons, Brinckerhoff Quade & rock encountered on M ission Street. No running or Douglas of New York. flowing ground was encountered.

M ost of the m achine-excavated tunnels would be Panelist T . KUESEL (U. S. A.) classed as potentially ravelling ground. The machin es generally have closed faces with narrow slots or do ors SYNOPSIS through which the soil is taken in as it is scraped off the face. This perm its supporting the face through Experience on the construction of 14 miles the m achine's thrust jacks. Nonetheless, in the ear ly of soft-ground tunnels and six large subway operations a serious cavity developed above one tun ­ station excavations has been accumulated on nel, which was traced to a failu re to keep the adva nce the BART project. This paper covers the of the machine equal to the volume of soil excavate d. bases of design, construction methods used, and field observations on movements, distor­ tions, and effects on existing structures. Use of Compressed Air

Scope Mandatory com pressed air was specified on six con­ tracts, generally where dewatering might consolidate The San Francisco Bay Area Rapid Transit System com pressible soil layers. Partial dewatering was (B ART) com prises 75 route m iles, including 13 m iles permitted in a number of cases. In Low er M arket of cut-and-cover and tunneled subways in San Fran­ Street, com pressed air was used to reduce the m ove­ cisco, Oakland, and B erkeley. All tunnels are single- ment of soft Bay Clay into the heading. A pressure of track sections, approxim ately 18 feet in diam eter. 12 psi reduced the value of N = T H/S^ from about 6 There are 15 tunnel contracts, including 42 individ ual to below 5. drives with a total length of 14m iles. U sually ther e are two parallel tunnels at the same level, but under M arket Surface Settlem ents Street in San Francisco there are four tunnels (two over two), and there is also a three-tracK section (tw o With isolated, exceptions, all tunnels have been dr iven over one) in Broadway, Oakland. At present writing with a prevalent surface settlem ent not exceeding two (July 1969), 27 tunnels com prising 10 m iles have be en inches, and usually less than one inch. The greater com pleted. part appears to be widespread settlement associated with groundwater drawdown. Settlem ent attributable Soil Conditions to loss of ground is generally less than one inch o ver the tunnels, and for tunnels in the center of the s treet, Sim plified soil profiles along M arket and M ission the settlem ent along the building lines is negligib le. Streets in San Francisco are shown in Fig. 1. The predominant stratum on M arket Street is a dense, fine, slightly cohesive sand. A large wedge of soft, As a general exception, there is alm ost invariably plastic clay intrudes near the shoreline, and scat­ greater settlem ent at the start of a tunnel drive. This tered lenses of peat are encountered around the C iv ic occurred with alm ost all contractors, types of grou nd, Center Station. O nM issionStreet, the interfingered equipment, and methods, owing to a lack of team ex­ granular and cohesive alluvial deposits are m oderately perience with the particular combination of circu m ­ consolidated, but occasional pockets of com pressible stances represented by that tunnel. Local settlemen ts m aterials are encountered, as well as several pin­ of three inches w ere not uncommon at starting areas - nacles of weathered serpentine and sandstone rock. some cases reached 8 inches. Among conditions con­ TunnelB lie as much as 50 feet below the groundwate r tributing to exceptional settlem ent w ere; table. In Oakland and B erkeley, the soils are all a l­ luvial, and fall generally within a classification of 1. Tunneling under street intersections repeatedy m oderately firm , cohesive granular m aterials. disturbed by u tility relocations.

Use of Tunneling M achines 2. Experimentation with new grouting m aterials, and new shield tail sealing system s.

B ART specifications require shields on all tunnels, 3. Vibrations from air com pressor plants. but leave excavation methods to the contractors. Fo r the four tunnel contracts in Oakland and B erkeley,the Soil arching over the first tunnel frequ ently lim ited individual drives w ere all under 1,500 feet long, a nd initial settlem ent to very minor amounts. D riving a construction schedules were generous. These tunnels parallel tunnel generally dislodged the arch-eventu al

312 MAIN SESSION 4

MARKET STREET LEGEND

0*OM fin* Mod

Firm silty and cloyoy toil* ComproMibt« toils - fill,»oft cloy and poot SECTION A-A SCALE IN FEET rvrtT Roc* ...... Ground wotor to Me

“ 'i _ o f i --- H * * ' h SECTION B-B

MISSION STREET

Fig. 1 - Soil Profiles - San Francisco Subways settlements over both tunnels w ere com parable. M arket Street, in soft Bay Clay. Three inches is a pre­ valent value, with som e cases up to 6 or 8 inches. The Figu re 2 shows the configuration of the three-tunne l difficu lty was not so much the weakness of the soil,but section in Oakland, as w ell as surface settlem ent p ro­ rather the presence of nearly 1, 000 buried tim ber piles files m easured over each tunnel as it was driven. T he rem aining from abandoned wharves and foundations. large settlem ent at the start of Tunnels No. 2 and N o.3 These w ere cut off in front of the shield, and air pres­ represents the effects of a tieback anchor in soil that sure was lim ited to avoid blowing chimneys out alon g pulled loose in the shield starting pit. (B yfortu ne or the piles. Some pile stubs remaining above the shie ld foresight, the adjacent buildings had been underpin ned.) invert plow w ere pushed over when the shield was jacked, rem olding the clay beneath the tunnel. Figu re 3 shows two cross sections of the first tunn el contract In M arket Street(the two upper level tunnels Effects on Existing Structures had not been started at the tim e of w riting.)The fi rst section is typical of conditions where relatively l arge There has been no significant differential settleme nt settlements (2 inches or m ore) w ere encountered,and under buildings along the tunnels. (On Low er M arket shows the effects of initial excavation, continued set­ Street m ost of the buildings had been previou sly de ­ tlem ent until grouting is com pleted, and the passag e molished for an urban redevelopment project, and of the adjacent tunnel. (In this tunnel, pea gravel was the principal remaining ones are on piles.) There a re injected into the tail void as the shield was jacke d for­ no significant claim s outstanding for building dam a ge- ward, followed by grouting approxim ately a w eeklate rj a typical complaint is that a door sticks. Many nota­ The second section represents much m ore prevalent ble cases of nonsettlement have been recorded.Am ong conditions, in which total settlem ents w ere a fraction them: of an inch. 1. An eight-story reinforced concrete Bank of G enerally greater settlements w ere observed on Low e r Am erica building', containing automatic m achi-

313 SEANCE PLENIERE 4 nery for processing all of the bank's checks, 3. The footings of a three-story wood fram e rests on spread footings on dense sand. A apartment house lie 15 feet directly above a tunnel was driven directly in front of and 40 tunnel in m oderately firm silty sand. The build­ feet below six 1, 200-tonfootings, with a m ini­ ing waB not underpinned. None of the residents mum horizontal clearance of less than four w ere evacuated, nor w ere any of them awakened feet. A closed-face tunneling machine was the night the tunnel m achine passed beneath. used, with 12-psi air pressure,and ring-by- 4. A thriving hardware store occupies a building ring neat cement grouting. No protective con­ constructed on top of the rem ains of several struction was undertaken for the building, previous buildings demolished by fire. The other than instrumentation and carefu l control structural support for part of the first floor of tunneling. M easurem ents showed thatthe pas­ consists of deeply charred wood beams resting sage of the tunnel caused the building to rise on piles of loose bricks roughly approximating 1 /8 inch. columns. D espite a recommendation to dem olish 2. A telephone switching station was located in an the stru ctu re.it was decided to leave it in place, old brick-w alled steel fram e building that had only strapping the loose bricks together, while survived the 1906 earthquake im perfectly. two tunnels w ere driven directly beneath the Cables had been installed to tie the building basement floor, one with a soil cover of only together across the old cracks. Spread footings seven feet above the crown. This was the last on dense sand w ere located 40 feet directly tunnel driven in Oakland, and benefitted from over the tunnel crown. The tunnel was m achine- the experience of an exceptionally able tunnel­ driven without visib ly enlarging the existing ing crew and foreman. Using an open-face cracks. shield and hand mining and face-support

SETTLEMENT IN INCHES TUNNEL No. I

TUNNEL Or

T UNNEL No. 3 o

so to no

Stiff silty cloy

Vsry stiff, moist coorM.sondy cloy Vsry M n u cloyvy tond Hord, moist, sandy cloy W ry dons«, w*t, silty sond CROSS SECTION Vsry stiff tondy cloy

Fig. 2 - Broadway-Oakland Tunnels - Surface Settlem ent» - m easured after com pletion of each tunnel drive.

314 MAIN SESSION 4

TYPICAL DISTRIBUTION-LARGE SETTLEMENT TYPICAL DISTRIBUTION-SMALL SETTLEMENT

pr ofiles : © 2nd Tunnel Excovoted ® 1st Tunnel Excavated © 2nd Tunnel Grouted (8) 1 st Tunnel Grouted (£) Two Months Loter

Fig. 3 - Transverse Settlement D stribution - M arket Street Tunnels

methods, the tunnels w ere driven through a coming the first flying cistern in San Francisco. cohesive, compact sand, without incident. The upper tunnel was dewatered and the low er tun­ Tunnel Liner Design nel used com pressed air. The charred beams and loose bricks are still there. B ART tunnel liners consist of fabricated steel seg­ mental rings, bolted together to form a uniform Along M arket Street, many of the steel-fram ed build ­ structural tube stiffened with 6" deep ribs about 2 '6" ings adjacent to the tunnel route have had "colum n on centers both ways. Each ring consists of six pick-up" jacking setups installed as a precautionar y welded pan sections 2' 6" wide by about 9' 6" long, m easure. Although only two of the low er level tun­ plus a short tapered key section to facilitate erec tion. nels have been driven, none of the jacking installa ­ tions has been activated, and experience to date in ­ The design is based on the flexib le ring concept, w ith dicates that none w ill be unless there is an accide nt. sections proportioned to carry(at norm al working stresses) the uniform ring com pression correspond­ The m ost prominent underpinning effort involved the ing to full overburden pressure, plus the bending Ferry Building, a historic San Francisco landmark at stress resulting from a shortening of the vertical the foot of M arket Street, which bestrides the tunn els diam eter (and corresponding lengthening of the h ori­ with a forest of tim ber piles. W orking from within the zontal diam eter) of 1/2 inch( ¿D /® *0.25%). The building, steel pipe piles w ere jacked down alongside bending stresses are calculated for a theoretically the tunnel alignment, and capped with prestressed unjointed and fu lly elastic ring, and com prise appr oxi - concrete beams spanning across the tunnels. The m ately 80% of the calculated stress. Consideration of building loads w ere then transfer red to the new fo unda­ the effects of the segm ent joints and of the bolt h ole tions, perm itting cutting off the old tim ber piles as clearances indicates that the ring is actually much they w ere encountered in the tunnel heading. Althou gh m ore flexible, and can absorb several inches of de­ these piles posed a considerable obstruction to tun nel­ form ation elastically. The large ductility of steel also ing, their rem oval caused no damage to the under­ gives the rings a substantial plastic deform ation c a­ pinned buildine. pacity while continuing to support external pressur es.

An unusual precaution involved a 30-foot-diam eter Tunnel Lin er D istortions cistern 18 feet deep, one of many placed beneath Sa n Francisco streets after the. 1906 earthquake as an A su rvey of 4, 647 rings in the ten Oakland tunnel em ergency reservoir, and still filled periodically by drives, made at the tim e the rings w ere erected in the fire department. To forestall any possible leak s the shield, shows that in 48% of the rings the dia­ due to ground m ovem ents, the cistern was drained. m eter distortion was within 0.25% of the original However, with no cover other than the street pave­ diam eters, and in an additional 25% of the rings, ment, and only 15 feet between the bottom of the ci s­ within 0. 50%. Five percent of the rings showed up tern and.the tunnel crown, the empty cistern provid ed to 1. 0% distortion, and l% of the rings exceeded 2% insufficient weight to contain the tunnel air press ure. distortion. It was "overpinned" by pumping 70 cubic yards of co n­ crete into it, and thereby avoided the distinction of be- Subsequently, the three Oakland tunnels shown on

315 SEANCE PLENIERE 4

Fig. 2 were checked by measuring vertical, hori­ and the collar plate while the shield jacks w ere still zontal, and diagonal diam eters on every tenth ring. pressurized. When the jacks were released, the tun­ Five rings in Tunnel No. 1 w ere measured shortly nel moved longitudinally 3/16 inch and cracked the after erection, when the shield tail void had been seal, requiring regrouting. packed with pea gravel, but before grouting. In all five rings the horizontal diam eter increased one to D eep Excavations two inches (0.5 to 1.0%).

The cut-and-cover sections include 18 subway statio ns, A total of 123 additional rings w ere measured from of which six are located in the 3-and 4-track sections two weeks to four months after erection and groutin g. and involve excavations up to 70 feet deep, 65 feet In 60% of these rings, the change from erected dia­ wide, and 800 feet long. Five of these deep station s m eter was less than one-half inch(0. 25%) in 30% of are in predominantly cohesive/granular m aterials an d the rings up to one inch (0. 50%), and in 10%, m ore are w ell advanced. The sixth(Em barcadero) is locate d than one inch, up to two inches (1. 0%) maximum. In largely in the soft Hay Clay of Low er M arket Street, the grouted rings of Tunnel No. 1, the horizontal where construction is in an early stage. diam eter generally shortened(in contrast to the un­ grouted rings). The two upper tunnels w ere random­ These stations are up to 60 feet below the groundwa ter ly mixed. The diagonal diam eters frequently showed table. To resist the hydrostatic uplift pres sure, which m ore distortion than the vertical and horizontal diam­ m ay aggregate 100, 000 tons on a station, the walls eters, generally up to 0. 75%, but there was no con ­ and invert slabs are made of concrete, three to seven sistency of direction. feet thick. As shown on Fig. 4, an internal structu ral steel fram ework braces the exterior concrete shell. It m ay be inf erred that the tunnel tends to squat until it is grouted, and that the process of grouting int ro­ Concrete interior floor and roof slabs com plete the duces random, unpredictable distortions. Since the structural system . greatest quantity of grout is frequ ently injected n ear the spring lines, the initial squat m ay be reversed .

No observations w ere made of the effects of passing adjacent tunnels, but in the generally firm ground of the B ART system the effects have not been casually noticeable.

In Low er M arket Street, in soft Bay Clay studded with old tim ber piles, m ore general distortions of two inches (1.0% ) have been recorded with sporadic case s of three to four inches (1.5 to 2.0%). This has re­ quired some recaulking, but has produced no evident structural distress.

M iscellaneous Tunneling Effects

On some 20, 000 rings erected to date, only two buckling failures have occurred. These were en­ countered in a tunnel driven through exceptionally firm sand, and w ere attributed to excessive unbal­ anced grouting pressu re(several tim es the design overburden pressu re) inadvertently introduced into the annular shield tail void.

The two buckled segments w ere cut out, removed,and replaced with new segments. The soil was firm enough to arch across the opening without additiona l tem porary support. The longitudinal precom pression induced in the tunnel lin er by the shield jacks cau sed the adjacent rings to squeeze slightly into the ope ning, so that the new segments did not fit, and considera ble F ig. 4 - Typical Section - M arket Street Subway Sta tion difficu lty was experienced in forcing them into pla ce.

At the foot of M arket Street, twin tunnels are driv en The permanent structural steel fram e is designed to through soft Bay Clay and a special clay fill into a be utilized during construction for support of the ex­ steel collar plate attached to the end section of the cavation. In addition, tem porary struts are require d Trans-Bay Tube, about 400 feet offshore. A tempo­ to lim it the depth of unbraced excavation below the rary grout seal was made between the tunnel rings deepest installed bracing level to 1 5 feet. A fter the

316 MAIN SESSION 4 concrete invert slab has been installed, rem oval of lateral support owing to its sm all modulus of defor ­ the tem porary struts is perm itted. This combination mation. This requires specially fabricated welded. of requirem ents provides maximum support of the H -section piles up to 5 feet deep, at 6-foot center s, bulkhead w alls during construction, minimum quanti­ with sections weighing as much as 600 pounds per foot. ties of tem porary m aterials, and reasonable con­ Individual soldier piles weigh as much as 30 tons e ach. struction working space. Even with these extraordinary sections.it is antici­ pated that lateral inward m ovements of the wall of Types of Excavation Support several inches may develop.

Two basic types of bulkhead w all construction are Consideration was given to underwater excavation used. In the first type, the w all is designed by th e methods and to installation of cross-lot diaphragms in contractor (within design requirem ents established advance of general excavation. These methods would by the engineer) and is not considered as part of the add a substantial cost prem ium , and the selected permanent structure. G enerally, this has resulted method is satisfactory so long as adequate lateral in steel soldier piles, spaced at six-foot centers, support of the toe can be developed. with wood lagging spanning between the piles, and external dewatering system s consisting of eductors or deep-w ell pumps.

The second type of bulkhead w all is designed by the engineer as part of the permanent structural w all. It consists of steel soldier piles with the spaces between them filled with trem ie concrete, and is designated the "SPTC w a ll." The piles are installed in undersized, slu rry-filled augered holes, so that the flanges of the piles are in direct contact with undisturbed soil. The spaces between the piles are then excavated, the re ­ sulting slot being kept filled with bentonite slu rr y. This slu rry is then displaced by trem ie concrete to form the com pleted wall.

The introduction of the steel piles into the concre te w all promotes control of vertical plumbness, fa cili ­ tates connection of the permanent interior structur al fram ework, provides im proved security of the slu rry - filled trenches against seism ic shocks,and eliminates all reinforcing steel. LATERAL ANTICIPATED PRESSURE WALL The concrete bulkhead w all provides an im perm eable ON WALL DEFORMATION cofferdam which can be excavated without drawing down the groundwater table outside the w all. The dif­ Fig, 5 - Em barcadero Station -C ritical Excavation S tage ferential head between the water levels inside and out­ side requires special provisions for control of see page Instrum entation beneath the w all and hydrostatic uplift pressures on the base. V ertical settlements are monitored by surveys of marks Dainted on existing buildings, sidewalks, and The "SPTC w all" was specified for two of the deep pavements. Some points w ere installed through pipes stations (C ivic Center and Em barcadero), where driven through the pavement into the underlying soil, prVsence of com pressible soils precluded external d e­ to elim inate bridging effect. w atering. It was made optional for the other four stations, but was chosen by the contractor only for Horizontal wall movements are generally measured by one (Pow ell Street), where the bulkhead walls had to means of inclinom eters in vertical casings installe d be installed with less than four feet clearance fro m im m ediately outside the soldier piles. On the C ivic the basement w alls of two busy stores. Center Station, an alternative system of horizontal extensom eters was used. At one end of the Em barcadero Station, the depth of fill and soft Bay Clay approaches 90 feet, while th e In each station, one or m ore zones 70 feet wide was depth of excavation w ill be ¿5 feet. The piles and established for measuring strut loads, through vibr at­ trem ie concrete w alls are toed into firm sand and denBe ing-w ire strain gauges cemented in pairs to opposite clay layers beneath the soft clay to secu relateral sup­ sides of the strut web. port. The m ost critical stage of the work occurs at an early stage of excavation (see Fig. 5) when only th e Preloading upper levels of internal bracing have been installe d, and the unexcavated soft clay provides negligible A ll struts are required to be preloaded by jacking a t

317 SEANCE PLENIERE 4

Horizontal üowo»»o*t - iocfcoo

10

20 -

30-

i 40- i x. m Ô. 50- fil a• 60

70-

80-

Fig. 6 - W all Movements from Slope Indicator Readin gs - 12th Street Station, Oakland installation. The preload is established by the eng i­ Figu re 6 shows two typical patterns of w all movemen t neer after review of the contractor's proposed brac ing recorded by inclinom eter m easurem ents in one of the layout, and has gen erally been about 25% of the des ign Oakland subway stations. earth pressure load. Struts are required to be shielded from direct sunlight, to reduce effects of In general, horizontal movements of the bulkhead wa lla, tem perature changes. have been kept within one inch without difficu lty, and have produced no significant surface settlem ents ad ­ W all Movements jacent to the w alls. No data are yet available on the larger movements expected in the soft clay strata on A ll available measurements represent construction in Low er M arket Street. dense cohesive/granular soils. M ovements are erratic, and apparently influenced by variations in construc tion technique as much as by variations in soil conditio ns. Surface Settlem ents

I^ S T . STATU TUNNELS

Fig. 7 - Surface Settlements - 19th Street Station, Oakland

318 MAIN SESSION 4 Figu re 7 shows contours of surface settlem ent in th e poured, the heat of hydration sw ells the steel stru t vicinity of the 19th Street Station in Oakland, which _ and increases its load sharply. This effect dissipa tes was constructed withtaoldier piles, wood lagging, a nd rapidly, but is replaced by concrete shrinkage, external dewatering. The locations of large settle­ which tends to recom press the beam. Intbn final sta ge ments correlate to locations of shallow com pressible concrete creep produces a gradual relaxation as clay deposits and groundwater drawdown levels. D if­ shrinkage strains are relieved. These effects are c on­ ferential settlem ents across any one building site are siderably m ore pronounced than any variation inearth slight, and have caused no significant distress. pressu res, and make any attempt at constructing a final external pressure diagram on the structure an unrewarding exercise. The com pressible m aterials are absent from the south end of the site, and the fractional settlem ents rec orded Effects on Adjacent Structures there are typical of those that m ay be attributed to loss of ground and w all m ovements. M ost of the subway station entrances are located in the sidewalks directly in front of existing structu res. Strut Loads and Earth Pressures W here these entrances required construction below Design criteria for free-draining soldier piles walls the existing building footings, the footings w ere e x­ specify a total earth pressure based on o H, with tended down to a safe level, generally by means of **o taken as 0.4 to 0. 5, and redistribution of this underpinning piers constructed in sections in hand- total pressure into a trapezoidal shape to allow fo r excavated pits. vertical arching of the relatively stiff soil, and for slight inward m ovem ent of the bottom of the w all. F or W here soldier pile and lagging construction was used im pervious SPTC walls, an allowance for hydrostatic for the deep station excavations, lim ited precautio n­ pressures is added. ary underpinning was generally undertaken.For large , heavy buildings, steel pipe piles were installed in Strut loads have been measured in cohesive/granular sections by jacking them down against the reaction of m aterials. Strain gauge measurements have generally the building footing, and then jacking them to a p r e­ been somewhat low er than the estim ated design loads . determ ined load and wedging them into place. U nder­ H ow ever, the recorded loads at one strut level (not pinning was lim ited to the first row of building co l­ consistently the same level) frequently approach th e umns adjacent to the excavation. Approxim ately 15 design loads. Figure 8 illustrates two typical case s buildings adjacent to the three m ajor soldier pile recorded at M ontgom ery Street Station. stations, and seven adjacent to tunnel sections, we re underpinned by jacked piles, working through the Loads in the permanent struts that are encased in building basements and subsidewalk basement vaults, concrete floor slab construction show considerable without serious disruption of use of the buildings and variation. During excavation, the load increases without structural incident. For lighter buildings, from the preload value as the excavation deepens. "underpinning control piers" were installed in shallow Tem perature variations cause the strut to swell and pits under the front column footings, with hollow ■hrink, changing its load. When the concrete slab is spaces provided to receive jacks to be used to adju s t

if------i.fi------i1;1------1 * J------li------LÌ------L

------*r 1 T T M II II

¡1 h II ■I II ______a. Jl______H------: ‘"-il i » i h li u

Average pressure calculated from largest maximum «train guage reading Averege pressure calculated frcyn smallest maximum strain guage reading

Fig. 8 - Earth Pressures Indicated by Strain Guage M easurements - M ontgomery Street Station

319 SEANCE PLENIERE 4 the footing depth if necessary. water-bearing ground, as happened twice on the Victoria Underground Line North of the Surveys indicate that the process of underpinning h as Thames, progress is delayed very substan­ caused the front of a typical building to settle a frac­ tially - so much so that ordinary hand tion of an inch. The rem ainder of the building has shields have been used more recently on the generally settled slightly as a result of dewaterin g Victoria Line South of the Thames where effects, and the building has suffered little d iffe r­ buried river channels were expected and. en­ countered. ential settlem ent and no significant distress. fortu nately, m ost of the deep excavations on the B A RT I entirely agree that the design and con­ project have been located in the better soils. None the­ struction of tunnels are Inseparable. Too less, the successful construction of such large und er­ often an engineer produces a design for a ground structures in close proxim ity to m ajor build ­ completed tunnel lining and the process of construction Is sorted out by trial and ings without significant damage has been accom plish ed error with an unspecified set of equipment only with detailed planning and carefu l control of con­ In the course of building the tunnel. The struction operations, in the light of the particu la r soil whole endeavour of design and construction conditions at each site. needs to be considered as an Integrated pro­ cess, In the same way that a production line Chairman 0. MORETTO In a factory Is developed.

Thank you very much for an interesting con­ YIELD AND MOVEMENTS ABOUND TUNNELS IN LONDON tribution related with one of the most impor CLAY tant tunneling work that has been realized at this moment in all the world. Any adverse effects of deep tunnels on over- lying and adjacent structures are caused by the yielding that occurs before the perma­ The next contribution will be in charge of nent water-tight tunnel lining is placed and Dr. W.H. Ward, Head of the Geotechnics Divi­ this 1b the case even under the relatively sion of the Building Research Station of ideal conditions In London. As an Index of Great Britain Mr. Ward. yielding of clay Peck has used the ratio (Pz - P a) eu Panelist W. H. W A R D (England) and Broms and Bennermark have suggested that tl;ls ratio should not exceed about 6 other­ SYNOPSIS wise the clay will flow into the face of the shield. On the other hand Peck suggests The author comments on Peck's State of the that If the ratio is much greater than 5 the Art paper in respect of tunneling in the light clay is likely to Invade the tail-piece of experiences in London on the yielding of clearance on the shield. the ground and the structural performance of tunnel linings of different flexibility. New I feel it is necessary to make clear that information is provided on these two topics. ratios as high as 5 or 6 are really an index of whether present shield tunnelling proce­ dures are feasible at all without a further INTRODUCTION increase in the ambient pressure within the tunnel. This criterion means that It Is My remarks will be limited to Part A absolutely necessary In all tunnels of any (Tunnelling) of Professor Peck's report on depth In normally-consolidated clays to the state of the art in 'Deep Excavation and apply quite substantial pressures continu­ Tunnelling in Soft Ground'. My experiences ously to the face (e.g. by air) to prevent in tunnelling are related mainly to over- considerable influx of the ground. This coneolidated stiff-fissured clays, in parti­ simple fact Is not as well known as it cular the London Clay where the conditions should be. At such large recommended for tunnel construction are relatively ratio values the yield of the ground cannot straightforward, but I have alBo had experi­ be under any real control with present con­ ences in water-bearing sand, soft silts and struction procedures and movement of the various varieties of soft rocks. ground can be quite disturbing so far as overlying and adjacent structures are con­ I agree to a large extent with Peck's philo­ cerned. Moreover the value of the ratio sophy and hiB criticislme of the present which causes a given volume of plastic state of the art of tunnel design and con­ yield of the clay in a tunnelling operation struction. The cost of construction of tunnels In London Is certainly coming down, but this has arisen mainly from Improvements In lining design and the elimination of two ( tunnelling machine similar to the modern processes, namely bolting of the lining and onee was used In the construction of the grouting, rather than the re-lntroductlon(1) Northern Line of the London Underground of tunnelling machines. Present tunnelling . before the end of the last century. I machines are a mixed blessing, while they believe Its uae was discontinued because improve progress when they are operating of the advent of the pneumatic spade they are more liable to produce overbreak which enabled excavation to proceed In and. if they encounter a burled channel of step with the rate of lining construction. 320 MAIN SESSION 4 depends very much on the detailed geometry was then trimmed to its final shape by the of the unsupported or partially supported cutting edge of the shield as It advanced areas of the ground and the duration of again. A few undralned compression tests lack of support before the permanent lining on small borehole samples show that the Is erected. shear strength of the clay is about 7 0 0 0 lb/ft2 at the level of the tunnel. Ae a first Btep vhen considering the con­ struction of a new deep tunnel In clay In a Two sets of observations of the convergence new area I prefer to consider yielding as of the London Clay towards the tunnel were commencing at a pressure-shear strength made by means of sleeved rods anchored at ratio of about 1, which Is the value one one end In the clay and which extended to obtains theoretically from simple elastic nearby underground structures where refer­ considerations of a cylindrical hole In a ence points were established. uniform stress field. If the ratio Is of the order of 1 or perhaps 2 It means that First, a set of lateral convergence measure­ excavation can be carried out fairly freely ments were made at the axis level of the ahead of the lining and that the short-term approaching tunnel at points a, b and c ground movement can be only of a small which were respectively 1.6, 6.5 and 11.5 elastic nature. This Is normally the case feet outside the tunnel excavation, see In deep tunnels at depths In common use In small plan in gig. 1. These measurements London Clay. Although a shield is fre­ were made wlth~reference to an existing quently used in London In these circum­ parallel tunnel at the same level and 25 stances we have no evidence to suggest that feet clear of the tunnel under construction. Its use reduces ground movements to any The total Increase in the horizontal dia­ appreciable extent. Short lengths of meter of the existing tunnel was 0.021* Inch tunnel are often built without a shield and during the construction of the new tunnel, without an obvious Increase In ground move­ but we know that the remote side of this ment. Rather the shield Is a convenient tunnel did not move horizontally from device for trimming the hole to a reason­ measurements made with another sleeved rod ably circular shape and for protecting the extending 20 feet into the clay beyond the miners from occasional falls of blocks of remote side, see Pig. 1. The convergence clay from the roof. Even when a shield is measurements are accurate to a few 0 . 0 0 1 used a length of about 2 feet of ground Is Inch. often exposed behind the tall to construct the ring of permanent lining and sometimes Second, a set of axial convergence measure­ several rings may be left ungrouted for a ments were made at three points A, B and C day. The conditions for tunnelling are at axlB level In front of the face of the obviously good. If we had to tunnel In same tunnel at a location some 160 feet London with , ahead of the first set of measurements, and P /' s u 25 feet before the tunnel entered the tim­ bered end of a tunnel chamber about 30 feet of the order of 6, I would expect ground long and 16.5 feet diameter lined with movements of at least 10 times the present cast-iron segments. The chamber was used values. Such movements could not be tole­ as a reference point, the chamber did not rated in urban London. elongate towards the approaching face and therefore remained quite stable during the In co-operation with London Transport, measurements. A small plan of the situa­ their consultants Mott, Hay 4 Anderson, and tion In front of the approaching tunnel Is their contractors A. Waddington & Son Ltd., shown In Fig. 2. Point A is on the axis we have recently completed a series of ob­ of the tunnel, point B at axis level on the servations on the motion of the London Clay periphery of the excavation and point C at close around a running tunnel of the Vic­ axis level, but 1 foot outside the peri­ toria Line during its construction at phery on the other side. The motion of Brlxton. The tunnel Is 80 feet deep and point C was recorded in a direction at a it was driven with a hand shield 13-5 feet small angle to the tunnel axis as Indicated in diameter and 8.5 feet long with a front in Fig. 2. hood extending a further 1.7 feet. The shield had a bead1 about £ inch thick and The complete converging motions of the 9 inches long behind the cuttjng edge which polntB a, b and c, and A, B and C are extended around the upper 300 degrees of its periphery. The shield was advanced in plotted relative to the position or the steps of about 20 inches, and a rl,ng of shield In Figs. 1 and 2 respectively. (In examining these figures It may help the permanent cast-iron lining of more or less reader to Imagine that the shield and the traditional design was built inside the tunnel are stationary and the ground flows tall of the shield. A gap about 1.5incheB past them.) The Jerky nature of all the wide between the clay and the lining was movements Is real, the faster movements be­ filled with cement grount soon after the ing associated with shoving of the shield. shield advanced. The excavation was made A number of most Interesting deductions can full face and It advanced in steps about be made about the motion o' the clay from 20 Inches ahead of the shield. When the these very simply measurements and some upper h a l f o f the face had been excavated will be mentioned here. it was open-timbered and held temporarily with a few face JackB while the lower h a l f The lateral motion of point 'a' passing o f the face was excavated. The excavation close alongside the shield Is particularly 32I SEANCE PLEN1ERE 4 FEET AHEAD OF HOOD FEET BEHIND TAIL

instructive. Ae the shield approaches, Sets of measurements are made frequently by the movement of point 'a' starts abruptly the resident engineers of the settlements and remains almost linear until it comes of the streets where they cross the line of behind the bead, it then suddenly acceler­ London tunnels at points directly overhead, ates and then slows down towards the tall and at points half depth away and full of the Bhleld. This olearly means that depth away on either side. The overhead the clay after leaving the bead converges settlement is typically 0.25 to 0.5 inch and bears on to the tall of the shield. As and zero at full depth away when a single It passes the tail there is the largest tunnel passes at about 7 0 - 8 0 feet depth. sudden movement which subsequently slows down as the grout sets and the lining takes This corresponds to a loss of ground of support. The nature of this movement around 3.0 sq.ft per unit length of tunnel. strongly suggests, as I have mentioned al­ Sets of observations were taken in two ready, that the use of the shield only de­ streets near to the above underground lays the elastic convergence temporarily measurements, but the results vary consider­ and does not reduce Its magnitude. By ex­ ably, the losses of ground being about 1.4 trapolation of the measurements the net radial movement of the clay at the shield and 5.1 sq.ft. These surface observations boundary Is about 0.57 Inches, so the total are, of course, liable to errors if the loss of ground for the construction opera­ natural reference points In the streets are tion per unit length of tunnel is likely to disturbed and this may have been the case be at least 0.57 times the tunnel circum­ here. However the Io s b of ground of 2.1* ference, or about 2. k sq.ft. sq.ft. estimated from the underground ob-

322 MAIN SESSION 4 eervatlons la of the aame order as the val­ of only about 0.05 inch close to the edge ues estimated from the surface observations. of the face compared with the displacement These small surface displacements are not of 0 . 6 8 inch at the axis, see Pig. 2, re­ normally noticed In buildings. veals a strong dome-llke shearing of the face. It Is this action which causes Turning now to the axial convergence meas­ openlng-up of the fissures which can be ob­ urements the very small axial displacement served at the face. The result is also of

323 SEANCE PLENIERE A interest to the problem of disturbance in in the tunnel with theii horizontal Joints In 'undisturbed' sampling. line. The three types of lining had the following dimensions and properties. When part of this tunnel was demolished to make way for a larger one we found with The first lining which was used extensively much Interest one or more sllcken-sided sur­ on the Victoria Line had 6 segments and a faces surrounding the whole tunnel an inch short key to the ring and was made of grey or two behind the layer of grout. The lin­ cast iron. The width of the segment was ing segments often came away from the clay 20 inches, the skin was 1 inch thick and the on these surfaces, which were all striated ribs 3 inches deep, Inch thick. in the direction of the tunnel axiB. We believe these surfaces are generated by The second lining was quite new; it was made quite small reversible strains arising from of ductile iron which has structural proper­ the forward shoving of the shield, see ties similar to mild steel, It had 12 seg­ curves B and C in Pig. 2, and the backward ments and a short key to the ring. The compression of the lining rings which were width of the segment was 2k inches, the skin spaced with thin timber packing. was -J Inch thick and the ribs 2£ Inches deep ■by £ inch thick. THE STRUCTURAL BEHAVIOUR OP THREE LININGS OP DIFFERENT FLEXIBILITY IN THE SAME TUNNEL The third lining was identical to the second except that each circumferential rib was cut In another part of the parallel running tun­ through at the centre of each segment into nels already mentioned at Brixton a full- a bolthole, so that the effective depth of scale experiment was made to compare the the rib was locally only about 0.9 inch. structural performance of 3 types of seg­ This lining is referred to as 1 cut ductile mental metal lining of very different stiff­ iron'. nesses. The results demonstrate very well the good sense and superior safety of using Some of the structural properties of the more flexible, thinner and more ductile metals and of the segments in circumference materials for tunnel linings and fully sup­ bending are given below. port the earlier thoughts of Terzaghl, which Peck has reiterated in his report. The The much better structural qualities of the results of the experiment are of wide Inter­ ductile iron will be noted. For about half est to tunnel builders. the weight of metal the moment of resistance of even the cut ductile iron is greater than At the site of this experiment the running the grey iron segment. Indeed, even when tunnels (13 ft 3 in. O.D.) were parallel to the ductile iron flanges are completely each other, at the same level and only about broken In bending and the segment is well 2 feet apart. Sections of three different bent it still has a resistance moment of linings were built into the tunnel construc­ about 60 ton in. It will also be seen that ted first. Detailed observations were made the flexural rigidity of the cut ductile of the circumferential strains In each seg­ segment is only about 1/5 that of the grey ment in one ring of each type of lining and iron one. of the ring diameters before, during and after the construction of the second tunnel. The net changes in the horizontal and ver­ The site was chosen with the two tunnels ex­ tical diameters of each of the three linings ceptionally close together, so as to provide aB the adjacent tunnel was constructed along­ an unusually severe distortion of the lin­ side are given below. These distortions ings. All the segments were curved rectan­ are much larger than the normal long-term gular pans with ribs for bolting together at deformations the tunnel would undergo if it the edges and all lining rings were assembled had been built as a single tunnel.

Structural properties.of the metal

Grey Iron Ductile Iron Tensile strength ton/in 12 35 Elongation 0.5 8 to 15 Young's Modulus x 106 lb/in2 12 to 1i* 2k Charpy impact ft lb 1 1k Structural properties of the segments in bending

Welght/10 ft tunnel Flexural rigidity Moment of resistance ______( tons )______(x 105 ton.ln21 _____ (tetti In. )______

Grey Iron 9.8 1 .k 65 Ductile Iron 5.3 1 .2 11+0 Cut Ductile Iron 5.3 0.3 7 5

324 MAIN SESSION 4 I am indebted to Mr. T. O'Donnell and llr. Change In diameter R. Carter of Mott, Hay & Anderson for the in­ formation on surface settlements and also to Horizontal Vertical my colleagues Mr. H.S.H. Thomas, Mr. P. Tedd Ins. Ins. and Mr. D. Burford for their help in carry­ ing out the work at Brixton. The note is Grey iron - 0 . 2 5 0 . 3 3 published by permission of the Director of 0.¿í4 - 0 . 3 0 Ductile iron Building Research. Due ductile iron 0 . U 7 -0.31

(-ive is decrease in diameter) Chairman 0. MORETTO

It will be noticed that: Thank you very much Dr. Ward for your interesting contribution and in special to 1. the increase in horizontal diameter is bring some controversial points with respect greater than the decrease in vertical to the lecture of the General Reporter. I am diameter in each case; this Is associ­ sure that this controversy will be the base ated with a local outward bulge of the of an interesting discussion when the exposi^ lining at axis level on the side towards tion of all the members in this first part the adjacent tunnel; will be over.

2. there is little difference in the dia­ By indication of the General Reporter and meter changes of the two ductile linings with the agreement of all the members of the despite a four-fold difference in the panel, we are going to break in this session flexural rigidity of their segments; the tradition and the chairman of it, is going to expose some experiences realized in 3. the diameter changes of both ductile his country. linings are significantly greater than for the grey iron, this is due almo6t entirely to the ductile linings having twice the number of Joints in the ring. I would like to make a few remarks about tunnel construction in the oity of Buenos Aires, where the All of these deformations are, of course, first subway line was built between 1910 and. 1912 perfectly acceptable for the uses to which the tunnels are normally put. using then mainly the cut and cover method.

In the segments themselves the smallest cir­ Buenos Aires city is built up along the shore of the cumferential bending moments occurred in the Río de La Plata river on a plain underlained by deep cut ductile iron and the largest in the grey deposits of wind-blown materials that were modified iron. However, the factors of safety a- gainst bending failure are in the reverse by erosion and redeposition aB sedimentation proceeii order; In the grey iron about 2.5, for the ed and at the same time were preconsolidated by ductile about 7.5, and for the cut ductile capillary action due to drying. This formation, that iron about 7 . 3 . Obviously the ductile iron extends for many kilometers ir. land, is locally is still too strong and economies can be altered by the valleys cut by the tributaries of the made, since it is most unusual to drive tun­ above mentioned river that near their mouthes have nels so close in permanent tunnels. left substantial deposits of soft to very soft clay The distortion in the shape of any circular and loose very fine sand, reaching the soft deposit^ tunnel lining in London Clay in a particular in some locations, a depth of up to 40 m. Consequent set of constructional circumstances is de­ ly, in general terms, the service tunnels for the termined primarily by the properties of the town have to be cut in either one of the following ground. Any stiffness of the lining tends 1 to restrict this distortion. It is clear materials from the results of the changes in diameters of the three types of lining at Brixton that 1) A highly preconsolidated loess like formation doubling the number of ring Joints has had a with properties approaching those clasified by the far greaiter effect on the distortion than General Reporter as cohesive granular soils reducing the stiffness of the segment many times. Yet when both the number of Joints is doubled and the segment's stiffness is 2) A soft to very soft clay and or loose very fine reduced many times as with the cut ductile slightly cohesive sand. iron the distortion is still perfectly ac­ ceptable, and at the same- time there is more Tunnel diging in the first type of formation yields than an adequate factor of safety against a behaviour familiar to such class of materials. bending failure. It is quite evident in­ deed that the ultimate development of a Although the soil appears to fit very well to the light, though flexible lining, in which the use of tunnel moles, up to date only clasical mining distortions are controlled almost entirely in drifts by either the eo-called german or auBtrian by the properties of the London Clay has methods have been employed, breaking the soil in nearly been reached In the cut ductile chunks with the use of light air hammers, as is lining. shown in fig. 1. lowering of the water table posee Acknowledgement no special problem as the average permeability of

325 SEANCE PLENIERE 4

5 3 21

FRONT VIEW SECTION A-B

LONGITUDINAL SECTION

22 23 39 38 40 37

Fig. 2 - Digger shield, with continuous mucking and concrete pouring

15 13

12 19 18 33

LONGITUDINAL SETCION ALONG AXIS OF TUNNEL

^8* 3 - Operating method for digger shield with con tinuous mucking and ooncrete pouring.

326 MAIN SESSION 4 the water carrying strata is of the order of stiff cohesive soil deeappeare and the whole ssotion K - 10~4 cm/eeg and it may be aooomplished using is dug in Boft soil. longitudinal foot draine located in the invert, inside the tunnels and leading to sumps set 100 to A digger shield under atmospheric pressure is being 200 m apart. Barring accidents due to local used to advanoe the tunnel. As first planned, the discontinuities, tunnel excavation does not produce shield had a long tail that before olearing aoted as any notioeable effects on the surfaoe and, there­ the external face of a lengthy, travelling conorete fore, no reoords are kept for settlements. mold so that digging and concrete pouring could be accomplished in a continuous operation, as indicated fig. 2. Muoking was aohieved by mixing the exoavated soil with water, to oarry it away with a sorew oonveyor feeding a pump that elevated the material to the surface. Fig.3 outlines the method of operation. Due to its extensive length and to difficulties arising from the mixed soil profile in whioh the tunnel was being dug, the maohine proved to be unmanageable with a tendency to dip resulting from the preaonce of stiff soi] in the invert zone that could not be corrected. After nearly two years of unsuocesful triale with prolonged stops, the machine was thoroughly modified and transformed into a conventional digger shield. By this time, dus to dipping, the tunnel invert had desoended 4 m to the level ehown in fig. 5t In about half the length set in the project. A temporary support made of steel ribs and planok lining is now being ussd. It is strongly expanded against the soil, aB indicted in fi«. 4

The tunnel axis ooinoides with the vertical line running along the middle of a city street with one to trtro stories buildings on both sides. Therefore, Several tunnels have been dug through the soft d a y measursments of ssttlsments are usually limited to and fine Band in the Riaohuelo river valley and the distanoe between property lines though closs under the river bed for water supply or sewer. observation is kept for building oraoklng. The Preeently, a water supply tunnel is under right upper part of Fig. 5 shows a typioal settle­ construction. The section to be du# in soft ground ment distribution for the ssotion of tunnel where starts at the northen border of the Riachuelo a temporary support expanded against the soil Is be valley where the soil profile is made up as shown on ing used. It takes the form of an error curve as the left upper part of fig. 5» In the first part, indioated in the State of the Art Report, measurable the lower quarter of the tunnel eeotion rune on stiff settlements extending on eaoh side to a distanoe of soil rssting on a sand strata. Further ahead, the about 25 m from the axle of the tunnel. The

327 SEANCE PLENIERE 4 distance i - 4.50 n>, gives a ratio i/R =1.9 and a the:.e settlement observations are plotted in terms volume of the trough, calculated with the expression of time, as in the lower part of fig. 5> a line ie V = 2.5 i ¿"max.» equal to 1.7 m^/m, which is obtained which is identical for every point to equivalent to**10 ^ of the theoretical net excavation that of the surface measurements in the same Since z/2R = 3.4, in fig. 9 of the State of the Art section. Report, a point is defined bearly entering the zone pertaining to stiff-clays. In spite of a ll the measures being taken to minimize loss of ground, the surface settlement The lower part of fig. 5 shows how settlement on thB produced by this tunnel job is several tic.es axis of the tunnel progresses with time as the larger than those reported in the State of the Art shield approaches, passes and leaves the transversal Report for tunnels in sim ilar ground. Yet, the section being observed. It points out clearly that ra tio Pz/su doss not aTfiFf*T most of the settlement takes place as the shield passes and its tails clears the section being It may be of interest to state that in the sections obeerved, indicating that is mainly due to loet where the origin al tunnel machine worked regu larly, ground and overexcavation. However, to make sure the settlements were less than h alf those reported of this evidence, observation points were installed in fig . 5 while, on the contrary, the conventional at several levels to find out how their settlement shield with an improperly expanded temporary lin in g compared with those measured at the surface. The yielded twice as much. upper rigth part of Fig. 5 shows the distribution of settlement with depth in the tunnel axis and I pass the word to the General Reporter in indicates that near the tunnel crown settlement is order to conduct the discussion, related to the disagreement released during the lec­ slightly larger than at the surface, pointing out tures of the several members of the panel. that movement of the ground toward the tunnel Dr. Peck. produces a slight Btretching of the soil ir. its vecinity, a distinct indication that loss of ground General Reporter R. B. PECK is prac4kcally the only cause of settlement. When

APPROACHING MOVING AWAY Thank you very much. I have the impression TIME10AYSI TIME ( DAYS) that there are no very great discrepancies -10 - 5 0 25 30 35 in our points of view or in the data. All three panelists have produced some extremely fine data that add a great deal to the State- of-the-Art; precisely the kind of data of which we need, much more in order to find out where we really stand in this subject.

I shall address my comments to those of each of the panelists in turn. I do not disagree with Mr. Ward about the proper ratio of net pressure to shear strength. He said rather clearly that he prefers a ratio of 1 or 2 rather than 5 or 6. I certainly do also. I am sure well all would prefer a ratio of 1 or 2, if we could get it. But, if we had to stay within a ratio of 1 or 2, those of us who do not work in such amenable materials as the London clay might not be able to build any tunnels. The point is that with higher ratios there would be. more settlement of the streets. In London these settlements might at first glance seem unacceptable. But that is not quite the whole story. The cost of

SETTLEMENT VARIATION reducing the settlements would be calculable; ON AVERTICALLINE IN the settlements could be reduced to some ex­ THE AXIS OF THE TUNNEL tent at least by the use of a fairly high air pressure. This might for some reasons be a very undesirable thing to do, but it would be a way to reduce the ratio of net pressure to shear strength. The cost of reducing the settlements by this means would have to be balanced against the cost of repairing the utilities or streets, or of underpinning the adjacent buildings, or of doing whatever SOIL PROFILE might be necessary to cope with the damage that might be associated with the movements. Fig. 5. Soil Profile and Settlement of Tunnel So I think another way of saying what Mr. in Soft Soil. Ward has discussed is that it would be far

328 MAIN SESSION 4 more costly to tunnel if the ratio were of using a shield which is not rotating, it is the order of 5 or 6 than it is with the radio just pushed. It has a grill in front and it that happily prevails. This does not mean is being pushed againt the soft soil. Now, that tunnels could not be or would not be does the general reporter think that if driven in London if the clays there were instead of using that type of shield a rota^ softer than they are. It means that more ing shield were used, the loss of ground money would have to be spent either on the would be larger? tunneling procedure or on the protective measures. General Reporter R. B. PECK

Panelist W. H. W AR D I can not really answer that question because I believe most of these things would have Well, the point really at issue, I think, is to be tried out on a full scale to see what that we tried to reduce this cost, of course, really happens. It is very easy to make because there are many other cities contem­ hypotheses about what this or that improve­ plating underground systems where the ratio ment or supposed improvement might accom­ is even greater than 5 or 6, and the diffi­ plish, but until we try it, and also perform culty is here that people do not want to measurements of the sort that Mr. Ward was work in compressed air, it is almost a lethal describing, until we see what really happens process theBe days, and if one is working in the ground ahead of the tunnel as well as even at factors of 5 or 6 it seems to me that behind it, we may be wrong in our notions as there is no precise a method of control and to what would be an improvement. The only this is what we are all looking for and we way to settle this is to make appropiate, have not really got it yet. It should be po£ detailed, field observations. sible to tell, at the shield, more precisely, how much ground you are losing at every small I think there is assuredly a great room for motion of the shield, and this, as far as I improvement in shield tunneling by finding can discover, can not be done sufficiently some means of eliminating or reducing the precisely at the moment. It should be possi­ effect of the annular space behind the tail ble to go through the ground and control the piece. This is a source of lost ground that, loss of ground by allowing the soft clay to as far as I can see, nobody has really succe£ squeeze in at the same rate as the shield is fully coped with yet. There are many proce­ advanced; you can increase the ambient pres­ dures for doing a better job than has been sure by thrusting harder against the clay by done in the past, but I would say this par­ means of the shield. You do not need to put ticular problem has not been Bolved. I think air pressure into this, at least not theor­ that probably more will be gained by looking etically. And I think it should be possible at this part of the shield operations than to get much better control over the loss of at the details of what happens at the face, ground than exists at present. but this may be another one of those hunches that will be proved wrong by field observa­ tion. General Reporter R. B. PECK

I think this is quite correct and it merely Chairman 0. MORETTO highlights that we need to know much more precisely than we do just what are the seats Does any one wish to comment on this question? of the loss of ground and just what construc tion operations could or must be improved in order to make the procedures more feasible, Panelist W. H. W AR D before we change to some totally different procedure. It is not quite so easy as it may In 'one of the soft clay shields I saw in seem, however, to hold the clay with the face, and then permit just the right amount shield lining of movement into the tail piece to balance V the heave that goes with the held face. We anchor tried that a long time ago in Chicago and wound up with a heave to begin with and a I T 1 fixed point settlement afterwards. Even though in a few rod sleevex instances we did come back to about the same place where we started, the massaging effect on the utilities over the tunnel was drastic. finished tunnel

Chairman 0. MORETTO

I want to ask the General Reporter to what extent the difference in various types of V bulkhead J ______L shields may influence the I o sb of ground. Contractors in general do not want to use air pressure because the air pressure is too expensive and I have seen in tours here in Mexico City -probably you have too,- that Device for measuring the movement of for the soft clays of Mexico City they are the soil ahead the shield.

329 SEANCE PLEN ERE 4 Mexico City, the total release between the Chairman O. MORETTO noee of the shield and the first lining is 2i ” so they are losing 2i" times the diameter Well, I do not know if I have made a mistake of the Bhield Just from the design of the in my calculations but the ratio was 6 and shield itself; this is apart from what is the settlements were really much higher than lost at the face. I would just like to make those that one would expect from the State- one suggestion that it is quite technically of-the-Art. The tunnel was cut in through possible, and that iB to measure from the the lower third on hard Boil and the two Bhield how much ground is coming at you by a upper thirds on very soft soil. I do not simple mechanical means: know whether the non-uniform profile had any thing to do with this exceptional Bettlemenx Chairman 0. M O R ET TO but I am sure, absolutely, that every effort was made to decrease that settlement, with I b there any other comment? the collaboration of people who have had very great experience with tunnels, including Ame­ rican people who have worked in many soft Panelist T. KUESEL ground tunnels in the United StateB. There is only one solution which would probably In all the San Francisco tunnels we had in have worked: the use of aire pressure. But, soft clay at the foot of Market Street, the not knowing whether he would get any improve "N" value in free air, which was something ment in the real settlement, the contractor over six, was one of the reasons we specified finally decided to live with the settlement mandatory compressed air which brought the and fix the buildings as long as they con­ value down to Bornething below 5. However tinued to crack. Fortunately nobody came up after we chopped out several hundred piles I with any legal suits so the work is going on am not sure how much contribution the air that way. pressure waB making to controlling the loss of ground because there was inevitable loss of ground when we were working in front of the shield on the piles. The problem her«, General Reporter R. B. PECK is that you can device means of dealing with the prevalent conditions, but you alway6 have The legal profession is obviously not very to be on the look-out for the exceptional l£ aggressive in Buenos Aires. cal face conditions which will give you a much greater magnitude of settlement and dif­ ficulties.

Chairman O. M O R ET TO Chairman O. MORETTO

Does the General Reporter have any comments? Well, with this we close the first part of our session and we will have a 10 minutes recession. General Reporter R. B. PECK

ThiB iB certainly a correct observation. I feel that the air pressure was absolutely necessary on lower Market Street, just to permit making progress and to keep the settle SECOND PART DEUXIEME PARTIE ments as low aB they were, even though they ~ became quite large over this portion of Chairman O. M O R ET TO shield tunnel. It wae a deep tunnel in diffi_ cult ground; probably we might not have need The second part of thiB session will be de­ ed aB much air pressure if the piles had noT voted to deep excavations and will follow been there. The air pressure certainly made the Bame procedure utilized in the firBt the job possible. part so in spite of time I give the word to Dr. Ralph B. Peck who will read his report in the same way as in the first part, for the case of deep excavations Dr. Peck. I suppose as long as I -have the microphone I might comment on Moretto'e example. He evaluated the ratio of net pressure to Bhear strength as 6, which is probably the correct number. I made some rough calculations from General Reporter R. B. PECK the diagram and came out with 10 but the di£ Prof. R. B. Peck's State-of-the-Art report crepancy really doee not much matter. In appears on pp. 225 of the State-of-the-Art any event, there was a high ratio of net-pre^ sure to shear-Btrength and the settlements volume. were quite large. If they differ by a factor of say 2 from thoBe in the diagrams in the Chairman 0. MORETTO State-of-the-Art Report, I think we are in good agreement. Thank you very much Dr. Peck for an interest-

330 MAIN SESSION 4 ing summary of the Becond part of your Gener­ EXPLORATION OF THE SITE al Report related to the State-of-ttae-Art in deep excavation. The success of any deep excavation depends upon the care which is exercised in the ex­ The next B p e a k e r will b e professor Jennings, ploration and in the planning of the work chief of the Civil Engineering Department, right up to the stage where building below of the Witwatersrand University. ground level is completed. Once the plan has been decided, the only changes which should be accepted are those which are neces­ Panelist J. E. JENNINGS (South Africa) sary for the safety of the work. Changes in depth or alterations in the planning of the The city of Johannesburg in South Africa, as building which will fill the excavation with many other cities in the world, is pas­ should not be made. Once started, the whole work should proceed rapidly with a minimum sing through a stage of redevelopment with of delay until the final construction is the replacement of existing building by high completed. rise structures. Many of these have deep basements, some of which may extend as much The preliminary work is as follows: as 90 ft. below the surrounding street levels. (a) A thorough site exploration should be The relevant details of the supporting sys­ carried out. This must define all tems for four such excavations are given in strata to a depth which substantially Fig. I (a) - (d) . STREET 0 STREET

Case (a) 30 in. bored concrete piles held by cablee anchored in rock STREET 0 Case (b) 30 in. bored piles braced back to A n u a horizontal bracing system Bvil di M s STREET ADD 0 / ^.concrete walls caet against / f aae rook anchors

70' Case (c) Wall slabs oast against excavation Case id) 30 in. bored concrete piles held and held by cables anchored in rock by cables Fig. 1 Typical cross-sections on Four Basement Excavations in Johannesburg, South Africa

331 SEANCE PLENIERE 4 •exceeds the depth of excavation. All draining soils. The effects of water pres­ horizontal changes in subsoil should sures on the surfaces of failure must be in­ be located. Local variations should cluded. A depth of tension cracking which also be thoroughly understood, remem­ should not exceed half the height of excava­ bering that the greatest difficulties tion should be taken into account in these will be experienced in situaticns calculations. The $=0 method may give queer where the soil is weakest. All soil results as the excavation approaches a limit­ profiles should be systematically re­ ing depth. corded, noting apparent moisture con­ dition, colour, consistency, soil The procedure suggested by Peck, treating structure, soil type and probable the excavated face as terrace loading has origin of each stratum. Up to this much merit and agrees almost exactly with stage only minimal laboratory tests experience in Johannesburg, i.e. if su is the are required, perhaps only the Atter- undrained shear strength of the material berg Limits and gradings. below the cut depth and1 if N is the stability number, N=YH/su , then: (b) During the site exploration the water situation in the soil should be clearly (a) no significant deformations will occur defined. Piezometers should be instal­ if N O . l t (based on elastic stresses led in those exploration boreholes in the foundation below a terrace load­ where it is judged there is sufficient ing with a vertical face); flow to permit their operation; Cb) movements will occur and these will otherwise 'null-flow1 electrical pore- pressure 'gauges should be used. It is become progressively greater as N very useful to have several such measur­ increases from 3.14 to 6.0; ing devices installed at different Cc) large movements will occur if J[>6.0 levels in a single borehole - this will and there will also be a possibility give a measure of any vertical flow of failure. A value of N=6.0 should gradients in the soil. be taken as defining the maximum depth of excavation in a particular When the information in (a) and (b) above has soil (based on bearing capacity theory been assembled the engineer should consider with 0=0). the whole problem and attempt to find pos­ sible solutions which will give the best Earth PreeBurea on the Support Syatem - Most marriage between the requirements of both support systems undergo a movement into the the structure and the soil. He will now be excavation which is either parallel to the in a position to decide what laboratory or vertical face or a rotation about the top. field tests, if any, are required to allow Hence the earth pressures will be approxi­ his designs to proceed. He will also be mately parabolically or trapezoidally distri­ able to decide from which regions the test buted, i.e. an arching active condition will samples should be taken. He should avoid be developed. It is accepted that the move­ 'over-testing' as it is confusing and even ments necessary to develop such active misleading to have too many laboratory tests (total) pressures are much smaller than those on the wrong samples. necessary to cause a triangular distribution of pressure. Therefore support systems THE DESIGN OF THE SUPPORTING SYSTEM FOR THE which rely on developing active pressures EXCAVATION by rotation about the toe must receive very special attention because of the larger move­ Two basic problems must be considered, ments necessary in such cases. naipely, the overall stability of the excava­ tion and the earth pressures which must be Experience in Johannesburg suggests that the resisted by the system for lateral support. movements of a city basement excavation should be limited to 1J in., otherwise damage Overall Stability - The most common proced­ to street services or buildings across the ure is to conduct a slope stability analysis using circular surfaces with <(>=0 for soils street may be excessive. The following i6 a summary of the observations leading to this possessing plasticity, or plane failure weJges with c=0 for non-plastic, freely conclusion:. Predominant Soil Height of Horizontal Movement Ratio Comment Supported Excavation of Top of xcava- A/H tion, A

Firm fissured clay 45' 3" 1:150 Damage to services in the street and build­ ings across the street Firm fissured clay 45' U" 1: 360 Acceptable movement Firm fissured clay 75' U" 1:600 Acceptable movement Very stiffj* fissured clay 45' 8" 1: 720 Acceptable movement Soft jointed rock 60' 1" 1:720 Acceptable movement

332 MAIN SESSION 4 These data show that as far as damage to WATER PRESSURES IN THE SUPPORTED SOIL street services and buildings across the street is concerned, the tolerable movement As in all retained backfills, water pressure is independent of the depth of excavation. in the supported soil is the major problem. When the excavation is carried out adjacent Excavation below the water table, which is to an existing building the movements may almost an invariable situation with all have to be less. Each case should be con­ deep basements, results in transient flow sidered on its particular merits, taking nets which depend upon the rate of excava­ account of the flexibility of the building tion, the permeabilities and the rate of and the consequences of any damage which replenishment. The effects of water pres­ may be caused. sures are most severe at the early stages of the excavation. Later, when the equilibrium However, the ratio movement/depth is known flow net has been established the required to determine the total support pressure. support pressures are smaller. Nevertheless, It is probable that arching active condi­ even at this later stage, the effect is to tions will be achieved when movements exceed increase the pressures by about 50%. a figure of the order of H/1000. This is fortunate since even with an unusual depth The first stage in any deep excavation should of 100 ft the necessary movement is only therefore be the dewatering programme. In lj in. which is within the acceptable range Johannesburg, filter wells at 25 ft centres when the surface is occupied by a street. around the perimeter of the excavation have The total horizontal force to be resisted been found to work satisfactorily. This has will be the total active pressure and its been somewhat surprising in the firm clays distribution will be approximately parabolic which are being supported and the reason for or trapezoidal. Even if this is an under­ the success is probably due to the fact that estimate, the error is unlikely to be grea­ drainage is taking place along fissyres. ter than 50%. This may be reasonably accep­ The wells are 8 in. boreholes with 4 in. ted as within the margin allowed by the slotted casings and with filter sand in the factor of safety but before he accepts it, annular spaces. Each well is equipped with the designer must be quite sure that the a deep well ejector type pump and the water water pressures in the backfill will be level is controlled at 10-20 ft below final controlled. excavation level. Piezometers measure the effectiveness of the drainage and pumping In most support systems it is unlikely that should be started well in advance of excava­ wall friction or adhesion will be developed. tion. A very convenient way of calculating the total pressure is to consider this as result­ Other steps must also be taken for the con­ ing from the pressure of a fluid with an trol of water. All street services must be equivalent unit weight Ye . The British examined to ensure that the^e are no broken Code of Practice No. 2 for earth retaining water-carrying pipes. This examination structures requires that no permanent wall should be repeated at regular intervals shall be permitted unless it can withstand throughout the construction. Regular crack an equivalent fluid pressure with Ye =30 pcf. patrolling should be carried out around the Considering Peck's Fig. 33 in terms of site and if any crack develops it should be equivalent fluid pressure, it is found that sealed to prevent entry of surface water. for his lower design trapezium, Ye =36 pcf Finally, continual watch should be maintained and for the higher trapezium Ye =72 pcf. It for any abnormal water entry into the exca­ is suspected that some of the higher pres­ vation - if this occurs, then no effort sures included by Peck may have been due must be spared in locating and controlling to uncontrolled water pressures, a subject its source. For example, in the excavation which is dealt with later in this report. shown in Fig. 1(a) a broken water pipe to a Many successful support systems in Johannes­ lavatory in a building across the street burg have been designed using Yjr=30 pcf caused the bottoms of a group of piles to with parabolic distribution of the pressure. move inwards 2J in. This caused much alarm There seems good reason to re-examine the until the source was located and the frac­ position before going to values as high as ture was repaired. those suggested by Feck. CONTROL MEASURES DURING CONSTRUCTION In fact it may even be argued that if a Y£=30 pcf is required for permanent construc­ However well the exploration is carried out tions then a lower Ye should suffice for a and the design executed, it must be appreci­ temporary support. All excavations and their ated that the estimated pressures and other temporary works can be viewed as construc­ soil conditions are at best only approxi­ tions which will be carried out with compe­ mations. Throughout the work, as the exca­ tent engineers in attendance. If proper vation exposes the soil, careful watch movement records are maintained and if the should be kept for conditions which may be possibility exists for avoiding action to be different from those which had been accep­ taken if the movements ^threaten to go out of ted in the design. To guide his under­ control, then a lower design pressure might standing of the behaviour of the system, be permitted. Certainly successful support the engineer controlling the work should systems have been designed and constructed carry out the following measurements: in Johannesburg with Ye as low as 15 pcf. (a) Precise levels on a large number of

333 SEANCE PLENIERE 4 point6 on the perimeter of the exca­ The behaviour of tnese materials depends upon vation. These levels should have an the behaviour along joints rather than on accuracy better than ±1.0 nun. and the strength of the material in between the should be referred to several stable joints. Many rock faces may be cut verti­ benchmarks remote from the site. cally to considerable height without any These 6hould preferably be placed in support but others must be supported as if boreholes using sleeved rods to a the material were a soil. level below the excavation level. It is useful if surface levels on The designs of the support systems are based several linet. at right angles to the on data provided by joint surveys and esti­ excavation also be observed. mates of strength along potential failure surfaces which incorporate the joints. The (b) Horizontal movements of points on the methods of design are dealt with in another perimeter. These should be observed paper to this conference. The economic bene­ with an accuracy of better than ±3 mm. fits of such designs may be very consider­ and should be referred to a sufficient able. number of 'immovable' points remote from the excavation. The procedures THE FACTOR OF SAFETY are sophisticated involving both angle and distance measurements with adjust­ In the calculations of overall stability, ment calculations. A useful procedure based on slip circle or plane failure is to string a taut piano wire just theories, current practice favours a factor inside the excavation and to measure of safety applied to the strength of the displacements of intermediate points material, i.e. an allowable developed from this line. The line itself is strength, s

334 MAIN SESSION 4 CONCLUSION nese cities are required to have internal parking area in proportion to the total floor area. A large-sized Experience with the digging of several deep building also requires several basement floors for m a ­ excavations in the city of Johannesburg has chine rooms to house mechanical equipment, shopping shown that while the design estimates of centers on the first or second basement, warehouses, overall stability and earth pressures are building maintenance and so forth. Consequently, al­ of great importance, the problems of sup­ port do not end with a subsurface investi­ most all large-sized buildings to be built in the center gation and the design of a system. The of cities have three or more basement floors. A con­ mental work continues throughout the whole siderable number of them have five of six basement exercise until the whole of the structure floors. As for depth, not a few cases of construction below the ground has been completed. of the basement as deep as G. L. -25 to 30 meters have been carried out. For example, the new building of Chairman 0 . MORETTO the head office of the Bank of Japan is G. L. -32 meters deep, and the new Tokyo underground station, G. L. -28 meters deep. On the other hand, as cities deve­ Thank you very much Prof. Jennings for your very interesting contribution. lop, buildings equipped with the basement became to be constructed in the area with bad ground conditions The next contribution belong to Eng. T. where large-sized structures had not ever been built. Kuesel whose introducement I already had the Apart from these tendencies cities, the troubles, opportunity to do in the first section. seen in the constructions of underground structures for facilities of factories on very soft ground which is reclaimed by means of dredging, have been rapidly Panelist T. KUESEL increasing. Professor Peck points out in the report of this con­ Mr. T. Kuesel's contribution appears on page ference such problems as lateral movement and set­ 312 of this volume. tlement of the surrounding ground, base failure, and earth pressure. These become very serious problems to be solved when underground construction work above- Chairman 0 . MORETTO mentioned is carried out. Particularly base failure is the most difficult problem. Thank you very much Eng. Kuesel for an in­ This report will describe the observed examples of teresting contribution related to the cons­ deformation as well as the means and methods now truction and behaviour of a special type of adopted in Japan to solve these questions, to supple­ sheeting that is becoming very popular and ment the extensive and systematic report of Professor that you have had the opportunity to see in Peck. Now in Japan, the tunnelling method in soft spread way in the Mexico City subway con­ struction. ground has been actively put into practice both in the tunnel shield process and the mechanical tunnel shield The next contribution will be in charge of process. But this method will be omitted here because Dr. M. Endo, Director of the Takenaka Tech­ it is not long since it was developed and the generali­ nical Research Laboratory of Tokyo. zation is not yet established.

Panelist M. ENDO (Japan) 1. INTRODUCTION

The Japanese archipelago, consisting of four main lands and numerous small islands, is geographically :.HOPSIS mountainous, so that the lowlands are situated mostly on fans and deltas which form, in all, less than one- Excavation of ground can be divided into fifth of the whole land, and most of the lowlands are work that can be planned according to normal covered by alluvial strata. Major cities such as To ­ methods and work which must be studied for kyo, Osaka, Nagoya, etc. are located on the alluvial feasibility of each individual case with a strata near the seashore for convenience of transpor­ basic change in the thinking. The greatest tation. Those cities can be divided into two areas, difficulties are encountered when performing i. e., downtown which is covered by the deep layers of excavation of ground thickly deposited with a alluvial strata and uptown covered by diluvial deposit. very soft stratum of alluvial clay where there Fig. 1 shows the depth of alluvial strata and their dis­ is fear of base failure. Descriptions are tribution in Tokyo, and Fig. 2 those in Osaka. In both given and data presented on a number of exam cities, the top layer of alluvial strata is sand or silty pies of such cases, examples of measurements sand of some meters and the next Layer is very soft of bottom heave due to deep excavation, ar.d or soft silty clay, featuring a constant increase in of effects on buildings in the surrounding shearing strength as the depth goes deeper. The dis­ area of excavation sit»s. tribution of shear strength of these clayey soil of To ­ FOREWORD kyo and Osaka are shown in Figs. 3 (a) and 4, res­ pectively. According to the city ordinances, buildings in Japa­ In contrast with the alluvial strata, the diluvial

335 SEANCE PLENIERE 4

Depth of Alluvial Deposit

10-20m

20—30m

30 -40m

Prefecture

(a)

10 I U ptow n! • f x Silty

j • Clayey

• v AKanto Loam ▲A A *

• • A A x A A A aa • I • A A Av_ A • • A A A K > • A £ 4 A A A A A iA A A__._ A • * a a ‘ A ▲ A

• A

Fig. 1. Depth of Alluvial Deposit in Downtown, Toky o

10 20 30 40 50

Angle of Internal Friction ( ’ ) (b) 4. Fig. 3. Relation between Cohesion and Angle of Internal Friction of Silty and Clayey Soils in Tokyo; (a) Alluvial Soils; (b) Deluvial

Hyogo Prefecture Soils

15

O Allu v al Clay • • Deluv al Clay • • 10

•

Osaka Bay • •

o ° °- "»S O ° o o° o 5 ° *o ** o Depth of Alluvial Deposit o ° o (9o o o o o o 5 - 10m o o 1 0 - 15m

15— 20m 5 10 15 20 25 20- 25m Angle of Internal Friction ( ’ ) 25-30m

Fig. 4. Relation between Cohesion and Angle of Internal

Friction of Clay in Osaka Fig. 2. Depth of Alluvial Deposit in Downtown. Osak a

336 MAIN SESSION 4 strata of uptown indicate rather favourable condition of the surrounding soil, ground subsidence, etc. aie for deep excavation although they contain some soft made case by case. Undermentioned are some ex­ soil, as shown in Figs. 3 (b) and 4. amples of those studies. Consequently, the methods of deep excavation e m ­ Recently, an ordinance for preventing noise was ployed for diluvial strata of uptown (Case A) differ established in the cities, restricting pile driving from those for alluvial strata of downtown. With re­ practice, which produce large noises and vibrations. gard to the alluvial strata of downtown, the excavating Therefore, the construction companies are endea­ method varies in accordance with the thickness of the vouring to develop noise-and vibration-less methods strata: less than 10 m (Case B), less than 20 m and began to employ cast-in-place concrete wall m e ­ (Case C), and more than 20 m (Case D). Problems thod and bored pile method. and difficulties often arise in the last two cases, i. e. , Another prohlem which began to be spotlighted (C) and (D). recently is an excavation in artificially reclaimed In Cases (A) and (B), an ordinary excavating m e ­ lands of littoral industrial districts. Those reclaim­ thod is employed with sheet piles or soldier piles and ed lands are made of sand and silt of the sea bottom planks supported by struts or rakers. As far as this which were carried up by dredgers. In extreme method is concerned, there is few to add to the de­ cases, an 2.0 m excavation caused base failure and tailed discussions given by Professor Peck. One breakage of underlaid steel pipe piles. In such a small thing to add is that the record of soil pressure place, it is a very difficult work to dig 4 - 7 m pits measured in the sand layer is in fair agreement with for a plant building. the soil pressure distribution discoursed by Professor Peck, as shown in Fig. 5. 2. E X C A V A T I O N M E T H O D S FO R GR O U N D OF THICK. In Cases (C) and (D), careful preparatory studies ALLUVIAL CLAY for the prevention of base failure, lateral movement

Tenri Bldg. Nishikawa Bldg. Asahi Bldg. Tohden Bldg.

Earth Pressure Earth Pressure Earth Pressure Earth Pressure Depth, in (t/m2) Depth, m (t/m 2) Depth, m (t/m 2) Depth, m (t/m2) 4 G.L. 2 4 6 G.L. G.L. G.L. I I I 1 1 1 Z'XWTW W W 0 //AVAV AXSU'AV 0- Fill

Soft Sandy Clay

Medium Soft Sandy Clay 4 4- Clay Loose Silty Sand Dense Sandy Loose Sandy Medium G ravel Sand G ravel

=3 $ oc CQ Dense Sand 8 8 - V ery Dense ® " W—N y Sand H °3 10 .c 00 Very Dense CS k.M Sand Very Stiff a> I Very Stiff Sandy Clay H I Sandy Clay^ 12J Dense Sand Very Dense Sand Very Dense Sandy Gravel j

16 Stiff Sandy S ilt

Very Dense Sand

Fig. 5. Observed Maximum Earth Pressure Distribution at Steady State Condition in Sand Strata

337 SEANCE PLENIERE 4 In such cases as the beforementioned (C) and (D), dimensioned elastic theory for the case of uneven when the depth of excavation exceeds 7 or 8 m, N as height of the ground as indicated by the bold line. pointed out by Professor Peck becomes about 5 or6, Fig. 8 shows the resultant principal stresses also de­ and there arises the danger of increase in deforma­ termined by the elastic theory of the stress transmit­ tion and of base failure. This of course is influenced ted to the ground from the surface of the perimeter by£*°f the ground being excavated and by the manner walls of the caisson, the stress from the weight of in which the excavated area is expanded. In a subway the caisson transmitted from the supporting slab and project at Hibiya, Tokyo, an open-cut work site with cutting edges, and the stress caused by the weight of two stages of struts provided for lagging wedged a- soil assuming that the initial coefficient of earth pres­ gainst inside flanges if soldier piles, a typical base sure at rest is 1. 0. As is seen from the figure, the failure occurred when excavation reached H = 8 . lm, resultant ma x i m u m principal stress is oriented out­ judging by which it is thought that the neighborhood ward immediately below the cutting edges of the cais­ of this value is a limit for alluvial strata in Japan. son, indicating that the weight of the caisson pushes The value of N in this case was about 5, and it is back the intrusion of soil from outside. However, in thought that the conditions were such that failure this case also, the direction of the resultant principal would occur two-dimentionally. stress is gradually changed from outward to vertical with increasing depth. If there had been no hard When excavating this type of ground to depth grea­ ground underlying and the alluvial silt stratum had ter than above, measures must be taken to minimize continued to greater depth, the ground on the outside adverse effects of changes in stresses formed in the would have circumvented the perimeter walls to the ground by excavation. 98.291m-

AN EXAMPLE OF OPEN CAISSON METHOD

As one example, an open caisson construction in the immediate neighbhood of the site of the failure in the subway project will be described.

This open caiason had a weight of 25, 000 tons, plan dimensions as indicated in Fig. 6. The cutting edges underneath the perimeter walls and supporting slabs along the perimeter walls were provided to support the weight of the caisson on the soft soil.

Fig. 7 shows the direction and magnitude of prin­ cipal stresses in the ground determined by the two-

Supporting Slab

Wooden Supporting Slab

10 t/m*

Fig. 7. Principal Stresses in Ground, Calculated by the Two-Dimensional Elastic Theory for the Case of Uneven Height the Ground as Indicated by the Bold L ine

338 MAIN SESSION 4

\ • X X W V n ,

Fig. 8. Resultant Principal Stresses in Ground Prior to First Sinkage Calculated by the Two-Dimensional Elastic Theory

inside and this excavation method would have been inappropriate. This caisson was sunk to G. L. -17. 3m and the results of measurements of the street level on the exterior before sinking and after settling of the caisson are shown in Fig 9. In this case, it was at­ tempted to reduce friction by providing an inverted taper between the perimeter walls of the caisson and the surrounding ground filling the gaps with pea gra­ vel, but apparently this measure was inadequate since in Fig. 9 the settlement immediately outside the wall was considerable due to pulling in of the neighborhood soil, although 6 to 8 m away the settlement became extremely small.

A-A AN EXAMPLE OF TRENCH METHOD

E In the city of Osaka, a project as shown in Fig. 10 <8| was planned by the island method for a typical (C) - I type ground consisting of loose or medium dense sand B-b | down to G. L.-4. 0 m, loose silty sand from G. L. -4. 0 I to -8 . 0 m, and very soft alluvial clayey soil from " f ; G. L. 8 . 0 to -20. 0 m as also shown in Fig. 10. At the stage that some progress was made in the project, c-c cave-ins of 40 to 50 cm occurred at the places indi­ I ?7/

339 SEANCE PLENIERE 4 As a result, since it was judged that the condition was Cohesion (t/m*) unstable two-dimensionally, the suggestion of Bjerrum and Eide (1956) was adopted and safety factors of 1. 45 and 1. 28 respectively were obtained against base fail­ ure for L=B and L”3B when excavating trenches with widths of B = 14 m. Although the safety was tentatively assured, in this case N=4. 5 was indicated which was larger than 4, so that deformation of the surrounding ground was considered a problem. On the other hand, the constructor wished to adopt as large a figure as possible for L for reasons of expediting the work, and consequently, a trial excavation was made at the loca­ tion indicated in Fig. 12, and measurements of move­ ment of the surrounding ground were made. This test was started with L=B, following which L was extended in both directions to see the relationship with move­ ment of the surrounding ground, measurements con­ sisting of the settlement at the ground surface and at G. L. -9.0 m, and of widths of cracks appearing in mortar troweled on the ground surface. Measure­ ments were made at 20 points (ground surface only) parallel to the trench and 5 points (ground surface and G. L. 9.0 m) transverse to the trench. For measure­ ment of settlement, a point 5.0 m farther away from the northwestern comer of the construction site con­ sidered unaffected by this excavation was selected as the reference point and an optical level was used"to Secure a precision of 0. 5 mm . The progress of the

test excavation is given in Fig. 13(b), and the results Fig. 11. Comparison of Cohesion between Test Results and of settlement measurements were as shown in Fig. 13 Calculated Results from Observed Kailure and 14,. The large settlements at Points E and e were

Construction Block B Construction Block A Depth. m Location where Failure Occurred Trial Excavation r ■ ■ \w a \' Slope Construction Block _JConstruction |t |E — B Block A rial Excavation Area r-"...... ’ Observation Area (b) Plan

Construction 50 100 Sheet Piles ^ Sheet Piles Block C Meters Y-Y' Section

Fig. 10. Osaka S. Bldgs. Project, Showing Cave-ins and Fig. 12. Trial Excavation to Observe Movement Lateral Movement of Sheet Piles and Piles of the Surrounding Ground

340 MAIN SESSION 4 due unfortunately to progress of excavation for an ad­ leviate the distortion of ground stress at deeper por­ jacent construction project during the measurement tions through excavation in small divisions. period. The horizontal displacement as measured by cracking ultimately resulted in a total cracking width CASE OF FLOATING ISLAND METHOD of 83 m m , but the movement towards the adjacent con­ struction site could not be discerned so that the amount The floating island method, as indicated in Fig. 16, of movement in the direction of the trial excavation consists of providing soldier pile walls or continuous could not be clearly grasped. In any case, because of concrete walls around the perimeter of a building, dril­ the substantial settlement, it was decided to construct ling holes at the locations of inner columns using large- a foundation and measurements were discontinued. diameter drilling equipment, erecting structural steel The relation between the value of N and the value of of the building columns, assembling the structural the amount of settlement at Point A in this example is steel for beams simultaneously with start of excava­ indicated in Fig. 15, which is in good agre­ tion at the ground surface, and placing concrete to ement with the values indicated in Professor complete the underground floors of the building from Peck's report. Needless to say, this trench method aims to al- Observation Points ^______40.00m —— ------|

Observation Points

g

I

10 —

20

Fig. 14. (a) Observed Settlements at Points a ~ e (G.L. ±0m ) 20 18 16 14 12 10 b (b) Observed Settlements at Points A ~ E (G.L. —9.0m)

4 £

8 &

Fig. 13. (a) Observed Settlements at Points 1—20 (G.L. ±0m ) (b) Profile of Trial Excavation Fig. 15. Relation between N and Observed Settlement . at the Point A in Fig, 12.

341 SEANCE PLENIERE 4

ip 4 I= > © Soldier Pile Walls

or Continuous Concrete W al

Large Drill Hole

© Structual Steel Column

©Concrete Floor

Excavation

© Assembling Structural Beam

© Bottom of Excavation

Fig. 16. Floating Island Method Fig. 18. Inside View of a Floating Island Method

Kinshicho Platform

Railroad

Depth, m y = lJOt/m1 0 ---- — C = 1.5t/m* - i = 20.0*

Struts

7 = 1.71 Loose Sand C = 1.5

Bottom of Excavation

Very Soft Silt

Soldier Files

Active Earth Pressure 7 = 1.60 C = 2.5 Passive Earth Pressure ¿=4.5'

Residual Earth Pressure'

Earth Pressure (t m*)

y 162 C = 4.4 d 6.0* Soft Silt

Fig. 17. Cross Section of Kinshicho Station Bldg. P roject, Showing Equilibrium of Sheeting Piles

342 MAIN SESSION 4 the top downwards, presently being the mainstream that work may be executed at N < 4 or at worst N < 5. method in Japan of constructing deep basements. An underground bowling center was planned under­ Using this method, construction above ground can be neath the baseball field of Tokyo Stadium(thickness of proceeded simultaneously with construction of the very soft alluvial clayey soil stratum: 29 m) and the underground portion, and there is also the advantage original plan for excavation to a depth of G. L. -7.5m of proceeding with construction while applying the at N = 6. 2 was altered to a depth of G. L. -6. 5 m upon weight of the building to the ground at least under­ reconsideration of space for piping, depths of beams neath the foundation. In this method, the principle of and earth cover for the underground structure. Fur­ preventing base failure through rigidity of the peri­ ther, by scraping 0. 5 m of soil from the portions of meter sheeting walls to withstand earth pressures is the ground beyond the slope shoulders affecting base applied. In zones of very soft ground, passive earth failure (width: 20 m) with bulldozers, excavation was pressures in the interior are hardly active and the made possible by open cut at N = 5.0. Even so, when intersecting points of the sheeting walls are moved excavation reached bottom, cracks of 2 to 3 m m were deep into the ground so that extremely large bending noticed to have formed 15 to 20 m beyond the shouldera moments would act on the sheeting walls. In such indicating that the safety factor against base failure cases, the sheeting walls would be designed to have was at the very limit. Whenever permissible it is adequate reinforcement, but also struts can be pro­ most economical to alter plans in this manner. vided diagonally downwards from already completed In the floating island method, when excavating lo­ upper floors for further reinforcement. cations with ground conditions such as (C) and (D), the A case in which base failure was prevented under principle of preventing movement of surrounding conditions for its occurrence through the resistance ground and base failure through rigidity and strength of sheeting walls is described below. of sheeting walls is adopted, but when necessary, the A railroad station building was to be constructed weight of the superstructure can be applied to allevi­ at Kinshicho Station in downtown Tokyo and the toe ate the shearing stresses within the ground. In actual of an embarkment for railroad tracks was to be ex­ practice, by performing excavation in subdivided por­ cavated 7. 0 m as shown in Fig. 17. The ground was tions, the work can be carried out to match given con­ as indicated in Fig. 17, and when combined with the ditions Fig. 16. Fig. 18 shows an example of excava­ loads of trains, N would be about 7 to 8. Excavation tion performed by the floating island method, which would be performed over a length of 120 m and as it enables safe operations except in cases of extraordi­ was considered to be a case of two-dimensional equi­ narily poor ground conditions. librium, various countermeasures were studied. It The present focal point of Japanese technology is was deemed that the trench method or methods de­ to construct the sheeting walls provided in advance pending on soil stabilization were undesirable from around the perimeter of the site to serve as the struc­ the standpoint of the construction period and so sheet­ tural wall of the building. In Japan, where earthquakes ing walls were decided to be used. The active and must be considered, it is required for underground passive earth pressures acting on the sheet piling walls to have the capacity to serve as shear walls; were obtained as values of earth pressure in the tri­ therefore the vertical joints must be capable of trans­ angular Rankine distribution, and on seeking the fixed mitting shear forces. Including waterproofing, schemes point in the ground the result as shown in Fig. 17 was such as indicated in Fig. 19 and Fig. 20 are being im­ obtained demanding exceedingly great rigidity and plemented. Also efforts are being made to improve strength of the piles. Since sheet piles with the lar­ the verticality of the units and besides using the ICOS gest cross-section were still inadequate, a soldier method, Soletanche method, Else method, and C G C F. pile wall constructed with concrete piers 980 m m in method introduced from Europe, several original diameter spaced at 1,060 m m center-to-center and Japanese excavating decices have been developed. reinforced with 40 steel bars 25 m m in diameter were provided as'a countermeasure. In carrying out the 3. MEASUREMENTS OF BOTTOM HEAVE actual work, it was found that compressive stresses CAUSED BY EXCAVATION of 1- to 2-stage struts as measured by earth pressure guages, although perhaps affected to some extent by Generally speaking, when constructing underground the location of the guages near horizontal braces at floors, the stresses in the ground are subject to the the strut ends, were less than half of the reaction influences of removal of load through excavation and forces predicted from calculations. Although cracks application of load through construction of the building. of several millimeters were formed at the slope sur­ In the case of very soft ground, the effects appear as face near the shoulder, there was hardly any settle­ plastic phenomena such as base failure, lateral move­ ment or deformation to hinder passage of trains dur­ ment of surrounding soil and settlement of surface, ing construction. but even if the foundation is made to reach hard ground To suppress base failure through rigidity and preventing these phenomena, when the depth of the strength of sheeting walls in the case of ground con­ underground floors is great, heave-up of the bottom ditions such as (D) where the thickness of deposits of of excavation due to removal of load and settlement very soft soil is large would require extremely great caused by application of load are of a degree which rigidity and strength in order to obtain equilibrium cannot be neglected. In ordinary open cut methods, below the excavated bottom and is excessively un­ the weight of a building is applied to ground which has economical. In such cases, the first consideration completed bottom heave, and construction of the build­ should be to reduce the design depth of excavation so ing is commenced from that point. But in the floating

343 SEANCE PLENIERE 4

0 0 o o o o o o o o o o o o o o o o o o o o 0 0 o 0 0 o o l i 0 0 o o o o o o ------c: o o o o £ e o o o o E «J o o o o J .2 « o o o o 3 ï a > -5 o o o o ^ o. o o o o o _ o o o o =5 1 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o — 91Ctam-

Fig. 20. Vertical Joints of Cast-in Place Concrete W alls Fig. 19. View of Assembled Steel Bars with Vertical

Joints of Cast-in Place Concrete W alls

Observation Points

Stress and Displacement of Steel Column

and Vertical Displacement at Bottom

® Vertical Displacement at Bottom

A Lateral Movement of Soldier Pile

Settlement of surrounding Surface

Dense Sand

Very Dense Sand and Gravel 40 Hard Clay

M eters

Fig. 21. A Building Near Osaka Station; (a) Plan an d Observation Points, (b) Section and Soil Strata

island method, the ground is excavated while the the western half was completed in Ma y 1968 as the building is being constructed, so that the building is first stage of the project. Fig. 22 gives the dispalce- affected from the stage of heave-up. Fig. 21 to 24 ments of the steel building columns observed at the are examples of floating island methods while Fig. 25 center of the first stage building and at the boundary and 26 are examples of open cut methods. Fig. 21 with the second stage portion (cf. Fig. 21), and the Bhows a building 12 stories above ground, 6 stories increases or decreases in the loads acting on the ex­ underground with an excavation depth to G. L. -25 m, cavated surface, as a function of time. In this case, now in its second stage of construction near Osaka the load removed by the excavation was 40 t/m2, Station. In this building all concrete placement of while the load applied by the total weight of the atruc-

344 MAIN SESSION 4

20

Fig. 22. Load Reduced by Excavation and Vertical Displacement at G .L.-25.0m (See Fig. 21.)

H 15 H 16 Upward Movement (mm) Upward Movement (mm)

E Sandy x: Gravel I- ---

"g Sandy iZ Gravel

0 1.0 Sandy Silt Decreased Load at the Final Excavating Depth (kg/cm *)

Fig. 24. Tim e v.s. Upward Movement below the Final Fig. 23. Relation between Decreased Load and Upward Movement at G.L. —25.0m (See Fig. 21.) Excavating Depth (bee Fig. 21.)

ture was 21.4 t/m2, but in Fig. 22, it is shown that the FI 1 m axi m u m load removal was held to 32 t/m2 as there was a load of 8 t/m2 of the building applied at the time inuum jl of final excavation due to the floating island method Depth, m î~: j 0 adopted. This observation was started after excava­ t n t tion was half completed and it is thought bottom heave 20 was actually even greater. In Fig 23 it is shown that 40 the relation between load and displacement of the struc­ Very Dense Sand and Gravel 1 /

ture is plotted to be roughly linear for the various ob­ bO 0 10 30 50 \ l servation points. According to this, bottom heave is shown to have occurred at a rate of 3. 3 c m per 1 kg/ First Stage Portion Second Stage Portion c m 2 of load removal, while settlement was at a rate of Fig. 25. Vertical Movement at the Tim e of Completion of 0. 45 c m per 1 kg/cm2 of load application. It is note­ the Second Stage Construction (See Fig. 21.) worthy that the rate of displacement during load appli­ rounding ground had settled 20 cm. cation is smaller than that during load removal. Fig. Another example is a building indicated in Fig. 26 24 gives the measurement results obtained in the constructed on a tableland with a high-rise portion 17 ground at the center of the second stage portion of the stories above ground and a low portion 3 stories high, building due to second stage excavation after the bot­ the foundation level of the high-rise portion being G. L. tom had been reached. This figure shows that the -16. 5 m and that of the low portion being G. L. -15 m. major cause of bottom heave is the expansion of the In order to preclude adverse effects of the joint be­ clay stratum in the underlying ground with the influ­ tween the high-rise and low portions in construction ence noticeable to a considerable depth. Fig. 25 gives of the building, measurements were made of displace­ the records of displacements occurring in the second ments at the bottom of excavation and at 7 m, 13 m, stage construction at the time of completion of second and 20 m below the bottom when excavating by the open stage excavation. It is interesting to note that the end cut method. Further, the displacements caused by in­ of the first stage structure had been lifted, the amount creasing loads during construction were measured. of heave on the first stage side thought to have occurred The results were as shown in Fig 27 and it was con^ in the first stage had been reduced, and that the sur- sidered that the relationship between increase and

345 SEANCE PLENIERE 4 decrease in load and the displacement, P , of the ex­ depth of -18 m in an adjacent lot. In both of the new cavation bottom could be approximated by the equa­ projects, the best methods available at the present tions given below for the periods of bottom heave and stage of technology were employed to reduce effects of settlement. in the neighboring properties. In the subway project, 1720 the pneumatic caisson method was used, while in the Bottom heave: PH = 12 10 (I - e'™ O'___ (I) adjacent building project, a continuous rigid sheeting 1700 wall 60 cm thick constructed by the ICOS method was Settlement: 540 •( I - é 1^ 0-___ (2) built in advance to a depth of G. L. -20 m with four stages of struts provided during excavation. where t: number of days elapsed since applica­ During the entire period the settlement of the build­ tion o£ O' ing was observed from an immovable point using opti­ cal levels and water levels. The result was a maxi­ As a result, placing of concrete at the joint be­ m u m settlement of approximately 50 m m as shown in tween the high-rise portion and the low portion was Fig. 30. The observations indicated very clearly the deferred until concrete of the high-rise portion had manner in which the influences of the construction reached the tenth floor above ground. work were manifested during each period of the con­ struction. These observations show that there is some 4. EFFECTS OF EXCAVATION ON BUILDINGS • effect on the surrounding area even when methods con IN SURROUNDING AREA sidered to have minimal effects are employed. As in­ dicated in Fig. 30, in the adjacent excavation work The Toko Building is a small-scaled building at with foundation concrete 60 c m thick placed leaving a Hibiya, Tokyo, with one basement floor and 8 floors small sump pit, settlement was not stopped, the set­ above ground, a fairly rigid reinforced concrete tlement ceasing only after waterproofing concrete was building constructed in the 1920's. The'foundation of placed and the excavation bottom was completely seal­ pine piles, as shown in Fig. 28, reaches a stratum ed. This is a phenomenon which raises the considera­ underlying the alluvial deposit, but it is questionable tion that settlement of surrounding ground is comple- whether the foundation work was adequate. A subway project was carried out close by under the road in front of this Toko Building as shown in Fig. 29, while excavation for another building was performed to a

Very Soft Alluvial Silty Clay ■

High-rise Portion Low Portion

Soft Alluvial Silty Clay

Medium Stiff Sandy Silt

Fig. 26. Plan, Section, Soil Strata and Observation Points Fig. 28. Foundation Plan and Section of Toko Bldg.

3.0r

D- 7 (Measurement Depth lm. 7m below the Bottom) D-13 (Measurement Depth 13m below the Bottom) D-20 (Measurement Depth 20m below the Bottom)

10 - + 1 E Load Increased by Building High-rise Portion Load Reduced by Excavation Construction ■s-*0 J -20- Low Portion -301______M ar, 1964 Sep. 1964 Sep. 1965 Sep. 1966

Tim e Fig. 27. Variation with Time of Vertical Movements and Load Reduced by Excavation (See Fig. 26.)

3 4£ 4

MAIN SESSION 4 tely stopped only after pore water pressures are suf­ sir able to change the design to obtain N < 5 , but when ficiently mobilized. Actually, on the whole, there this is not feasible, the only solutions would be to re­ were no adverse effects at all on the Toko Building. duce relative THby applying loads to the areas sur­ rounding the excavation surface, carrying out soil sta­ 5. CONCLUSIONS bilization to increases« and obtain N < 5 or increasing Neb by trench cuts and partial excavations. However, In excavation work, the greatest difficulties are in the last case, there is danger of considerable defor­ encountered when excavations are made by methods mation and settlement occurring in surrounding ground accompanied by danger of base failure in ground de­ and the utmost caution must be exercised. posited with more than 10 m of very soft alluvial clay, With the floating island method now widely used in Case(C), or in ground with thickness of alluvial clay Japan, one proceeds with excavation while construct­ stratum exceeding 20 m, Case(D). In Case(C), sur­ ing the building from the ground surface downward, rounding sheeting walls can be made rigid and strong and other than in cases of exceptionally adverse con­ to cope with the problem, but in Case(D) this method ditions, the method is adaptable to most cases due to becomes extremely uneconomical. ease of limiting lateral movements of the surrounding If possible, it would be most economical and de- ground through the use of rigid walls, partial excava­ tion, and by providing reaction against the building structure. However, when the depth of excavation is great and there is an underlying clay stratum, this method will be subjected to the effects of bottom heave and settle­ ment so that in buildings with high-rise and low-sto­ ried portions or buildings constructed in two stages, it is necessary to study methods of constructing joint sections which will not cause harmful Btresses in the building structures. 1IS- The Lot} V \ When carrying out excavation work on ground of(C) Toko of Excavation \\ or (D) category in the neighborhood of buildings with Bldg. Works of Adjacent Bldg. inadequate foundations, there is fear of adverse effects <-18m)| \\ 10 20 of excavation regardless of the type of excavation m e ­ & 41 Meters thod due to changes in pore water pressures in the ground. However, it may be said that workmanship Fig. 29. Underground Constructions around Toko Bldg. has a great effect on settlement and it is most import-

Fig. 30. Settlement of Toko Bldg

347 SEANCE PLENIERE 4 ant in this case to secure careful work. ing of Architectural Institute of Japan.

ACKNOWLEDGEMENTS ENDO, M., T. KAWASAKI, Y. IKUTA, and Y. KIMU­ RA (1969). "An experimental study on connect­ The author's utmost thanks are due to Messrs. ing properties between continous basement T. Kawasaki, T. Hashiba, U. Ikuta, walls and structures of building," (in Japa­ M. Niwa and T. Morita for furnishing data for nese) , Transaction of the annual meeting of preparation of this report. Also, the author Architectural Institute of Japan, Hokkaido. wishes to sincerely thank Dr. Y. Yoshimi and Messrs. M. Toraono, Y Suzuki and M. Kondo for HASHIBA, T., M. NIWA, and F. YOSHINAGA (1968) their special cooperation in the work. "Measurement of surrounding ground displace­ ment of Osaka Fiber Wholesale Center Bldg., “ (in Japanese), Technical Report of Technical REFERENCES Dept, of Osaka Main Office, Takenaka Komuten Co., Ltd., BJERRUM, L., and 0. BIDE (1956). Stability of Strutted excavations in clay, "Geotechni­ KAWASAKI, T., T. HASHIBA and Y. MENDE (1969). que, 6, No. 1, pp. 32-47. "Discussions on the Takenaka Deep Foundation Process from standpoint of soils engineering, ENDO, M., and U. TAKAHASHI (1962). Measure­ “ (No. 1, No. 2) (in Japanese), Technical Re­ ment of earth pressure acting on the struts port of Takenaka Technical Research Laborato­ of cofferdam, for sandy strata), (in Japane­ ry No. 1268, No. 1293. se) , Technical Report of Takenaka Technical Research Laboratory, June, No. - 590 KITAZAWA, G., T. TAKEYAMA, K.SUZUKI, H. OGA- WARA, and Y. OSAKI (1959). Tokyo Jibanzu ENDO, M., (1968). "Current of recent founda­ (Subsoil maps and profiles of soils in Tokyo) tion construction methods, " (in Japanese) , (in Japanese), Gihodo, PP. 113. Journ. of Architecture and Building Science. Architectural Institute of Japan, Nov., PP. KINKE BRANCH of Architectural Institute of 777-781. Japan, and Kandai Branch of Japanese Society of Soil Mechanics and Foundation Engineering ENDO, M., and U. IKUTA (1965). "A study of (1966). Osaka Jibanzu (Subsoil maps and pro- foundation works of NKK ironworks in Fukuya­ files of soils in Osaka), (in Japanese), Co­ ma, " (in Japanese), Technical Report of Ta­ rona -Sh a. kenaka Technical Research Laboratory, Spt., No. 912. KINKI BRANCH of Architectural Institute of Japan, (1967). "Muon nushindo kisohoho" (Cons, ENDO, M., and T. KAWASAKI (1960). "A Study truction methods of foundation with little of excavation method of Kinshicho Station noise and vibration), Text Book of a training Bldg.," (in Japanese), Technical Report of Ta. course. kenaka Technical Research Laboratory, No., 368, No. 373, No. 400. NIWA, M., and T. MORITA (1969) "Measurement on the surrounding settlement due to deep ex­ ENDO, M., T. HASHIBA, and T. OUCHI (1969). cavation at the No. 1. Bldg. for City of Osaka "An observation on an ultimate bearing capa­ “ (in Japanese), Technical Report of Technical city of silty clay. "Soils and Foundations, Dept of Osaka Main Office, Takenaka Komuten Vol. IX, No. 4, Dec. (Intended). Co., Ltd.

ENDO, M., U. IKUTA and H. NAKAZAKI (1969). OUCHI, T., (1951). "Open caisson process of “Experimental study of volume change behavior Nikatsu International Bldg., "(in Japanese), in partially saturated clayey layer during con Journ. of Architecture and Building Science. struction period, " (in Japanese) Takenaka Architectural Institute of Japan, Oct. PP. Technical Research Report. No. 4. 1 - 10 .

ENDO, M., T. KAWASAKI, and K. YAMAMOTO (1966) OUCHI, T., (1961). "Study of the sinking be­ "Measurement of settlement of Toko Bldg. af­ havior of caisson process in silty soils", fected by adjacent excavation works, " (in Ja­ (in Japanese), Dr. Eng. Thesis, Tokyo Insti­ panese) , Technical Report of Takenaka Techni­ tute of Technology, Tokyo. cal Research Laboratory. Sept. No. 1019 TAKAGI, M., K. SHIMIZUGAWA, and U. IKUTA ENDO, M., T. KAWASAKI, Y. IKUTA, and Y. KIMU­ (1969). "Report of the works by Takenaka RA (1969). "An experimental study of verti­ Basement Wall Method," (in Japanese), Seko, cal joints for continous basement walls," Syokoku-Sha, April 1969, PP. 27-42. (in Japanese), Transaction of the annual meet 348 MAIN SESSION 4 TBRZAGBX, K., and R. B. PERCK (1967). Soil without Struts", Proc. of the 3rd Annual machanlca in engineering practice, 2nd Edi­ Meeting of the Japanese Society of S. M and tion, John tfillay and Son«, Inc., Haw York. F. E., pp 499-502.

APPENDIX ODA, T. and OSANAI, S. (1963), "Measurements of the Earth Pressure and Tie-Rod Loads of List of Typical Japanese papers concerned the* Quay Wall at Kushiro Port" , Journal of Earth Pressure, Movement of Surrounding Soils the Japanese Society of S.M. and F.E., Oct., and Cofferdam in Japanese. No. 68, pp 19-26.

AMIMOTO, K., OKAMOTO, T. and YAMADA, S. (196 YAMAGATA, K., NAKATA, K. and FUKUMOTO, K. (1958), "Measurement of Earth Pressure Acting (1966), "Measurements of the Strut Loads and on Side Walls of Subway Construction", Jour­ Movements of Surrounding Soils During Exca­ nal of the Japanese Society of Soil Mechanics vation" , Journal of the Japanese Society of and Foundation Engineering, No. 34, pp 21-30. S.M. and F.E., Mar., No. 97, pp 29-36.

ABOSHI, T., OZAWA, A. and KUSUMOTO C. (1966), YAMAGATA, K. and YAO, S. (1967), "Variation "On the Measuring Method of the Deflection of the Earth Pressure Acting on the Coffer­ and Stress in a Quay Wall by Using a Inclino dam Consist of Steel Pipe Piles during Exca­ meter", Proceedings of the 20th Annual Meet­ vation (Part-I)", Journal of the Japanese ing of Japan Society of Civil Engineers, 3., Society of S.M. and F.E., May, No. Ill, PP 53. pp 29-38.

ANBIRU,. T. (1961), “Some Consideration Con­ YAMAGATA, K. and YAO, S. (1967), "Variation nected with the Measured Strut Loads (in of the Earth Pressure Acting on the Coffer­ Sand“, Trans, of the Architectural Institute dam Consist of Steel Pipe Piles during Exca­ of Japan, Oct., No. 69, pp 777-780. vation (Part- II)", Journal of the Japanese Society of S.M. and F.E., Jun., No. 112, ICHIHARA, M . , TABATA, K. and KONDO, M. pp 7-16. (1966), "Measurements of the Earth Pressure and Tensile Load of Tie-Rods Acting on the Chairman O. MORETTO Yamano shita Quay before and after Niigata Earthquakes", Journal of the Japanese Socie­ ty of Soil Mech. and Foundation Engineering, Thank you very much Dr. M. Endo for your very interesting contribution. Mar., No. 97, pp 13-22. The next contribution will be in charge of KATAYAMA, A., IGUCHI, H., (1961), "Measure­ Dr. Alberro of the Instituto de Ingeniería ments of Earth Pressure Acting on a Reinfor­ of the Universidad Nacional Autónoma de Mé­ ced Concrete Sheet Pile",Journal of the Ja­ xico. panese Society of S. M. and F. E., Aug., No. 46, pp 4-II. Panelist J. ALBERRO (Mexico)

KOTODA, K., MINOU, A . , ENDO, M. and KAWASAKI T. (1959), “Measurements of Lateral Earth INTRODUCTION Pressure by Strut Load (in Cohesive Soil)", Lors de la construction du métro de la Trans, of the Architectural Institute of Ja­ ville de México, la realisation d^une cam­ pan, Oct., No. 63, pp 689-692. pagne de mesures de chantier a été décidée. Il a été convenu d'installer des appareils KOTODA, K., YAMASHITA, J., (1959), "Experi­ de mesure des déplacements, provoqués dans mental study on the Movement of Sheet Pile le terrain par le processus d'excavation, with Model in Sand", Journal of Japanese ainsi que des vérins plats et des piezome- tres dans les murs latéraux du tunnel afin Society of S. M. and F. E., Aug., No. 34, de connaître la distribution des pressions pp 4-9. horizontales. Les résultats obtenus sont sa tiBiaisante et permettent d'analyser la di¥ KOTODA, K., OKI, N., (1968), "Comparison of tribution des déplacements et des contrain­ Observed Strut Loads with Proposal Value of tes sur le pourtour du tunnel. , Terzaghi and Peck(1967)", Proceedings of the Avant de présenter quelques-uns des re Annual Meeting of the Architectural Inst, of suitate obtenus, il convient de décrire de Japan, pp 641-642. façon euecinte les caractéristiques du sous-sol de la ville de Mexico a^nsi que les méthodes d'excavation utilisees. KOTODA, K., KANATANI, Y., MIYAZAKI, Y. and HANAMURA, M. (1968), "Experimental Study on CARACTERISTIQUES DU SCKJS-SQL DE LA VILLE DE the Movements of Sheet Piles in a Excavation MEXICO

349 SEANCE PLENIERE 4 Le sous-sol de la Tille de Mexico est H est communément admis que le foni^ forme de dépôts lacustres, d'origine^volca­ d'une fouille se soulève du fait de la de-% nique . En sarface, on trouve des dépôts de charge induite par l'excavation. Ce phénomè sable silteux^ou des remblais artificiels ne est à court terme élastique, et la fig 2 de 2 à 5 m d'épaisseur en général, où l'on, en présente un cas, analysé' à l'aide du dig trouve le niveau phréatique, suivis d'une gramme construit par N.M. Newmark . La com­ couche argileuse de 10 à 30 m d'épaisseur, paraison des déplacements verticaux du fond formée de cendres volcaniques trds compre­ de fouille calculés par cette méthode et me ssibles et veinée de minces couches de sa­ surés directement est tràs satisfaisante. ~ ble. La teneur en eau de cette couche argi­ leuse est en moyenne de 2Q0 pour cent et _ ses limites de liquidité et de plasticité sont égales respectivement a 290 pour cent et 85 pour cent en moyenne. Son indice des vides est de 7» H s'agit de plus d'une ar­ gile sensible au remaniement X Indice de sen sitivité égal a 8). La résistance à la com­ pression simple des éprouvettes d'argile, non remaniée est en moyenne de 0.8 k g / c m .

PROCEDES D*EXCAVATION

La profondeur d'excavation varie entre 6 et 8 m en général, sauf pour quelques cas particuliers tels que les croisements avec le réseau de drainage de la ville.^Dans ce cas,,1a construction du siphon nécessaire au rétablissement de la continuité' du co­ llecteur de drainage requiert une excava­ tion de 10 m de profondeur environ. La sta­ bilité des parois de l'excavation a été ob­ tenue suivant les cas au moyenne talus ou de palplanches étayées, soit métalliques Fig 1 Coupe du tunnel. Localisation des ap­ soit de béton moulé dans le sol. La longue­ pareils de mesure ur de fiche des palplanches est variable. Les étais des excavations ont été préchar­ gés, lors de leur mise en place, à 30 to­ nnes environ pour les étais proches de la surface et à 90 tonnes pour les étais pro­ fonds. Ceci correspond a des charges de 6 et 18 tonnes par metre linéaire d'excava­ tion. N L'excavation s'est faite a partir de la surface du terrain au moyen de pelles me caniques sauf rares exceptions. Les filtra­ tions ont été contrôlées au moyen de pompa­ ge et souvent par électrosmose.

JLEPAREH£ d e m e s u r e s B N -C J O .0 0 de prof); NivMU / ' Y Le long de la ligne No. 1 du métro. t Profondeur maxima trois stations de mesure ont été installées. de l'excavation Chaque station comporteNle long de la paroi ______(8 ,5 0 ). verticale du tunnel 4 vérins plats du type Freyssinet, 4 pièzometres pneumatiques pla­ Histoire de l'excavation cés aux mêmes élévations que les vérins — ______= = plats et pour l'une des stations un extenso 12.50 m mètre transversal de la Slope Indlcator, Co. avec 2 éléments sensibles. Le long de la ligne No. 2 du métro, ont été installées 2 stations de mesure du même type que pour la ligne No. 1. Sur la ligne No. 3, actuellement en construction, Vue en plan une^tation de mesure est déjà en place. Le Stolion de mesure km 5 + 7 6 7 .3 0 schéma de l'installation type d'une station est présenté dans la fig 1. , Fig 2 Déplacements verticaux en fonction du De plus, le long du trace des lignes temps ont été Installées des réferences topogra­ phiques qui permettent un relevé des dépla­ Il est aussi, non moins communément ad­ cements du terrain, lors de l'excavation. mis que la surface du terrain aux alentours de l'excavation s'affaisse. Il faudrait ce­ DEPLACEMENTS DU TERRAIN SUR LE POURTOUR DE pendant, à ce propos, préciser, il me sem­ L 'EXCAVATION ble, la durée du processus d'excavation et

35 0 MAIN SESSION 4 da pompage. En effet, à long terme, le pom­ xico,. La fig 4 en présente deux exemples, page provoque une consolidation du terrain tires des mesures faites lors de' l'excava­ et par conséquent un tassement de la surfa­ tion avec des p&lplanches métalliques de ce. Mais, instantanément et en accord avec deux siphons1. De même, la fig 5 résume les les résultats de la théorievélastique, le mesures faites pendant la construction des mouvement doit être un soulèvement /les alen immeubles du Centre Urbain "Présidente Jua- tours de l'excavation. La fig 3 présente la rez". Dans ce cas, la pente du talus de configuration des déplacements de la fron­ l'excavation de 6 m de profondeur et de tière de l'excavation, d'après un exemple 18 m de largeur, était égale à 0.75/1« traité par la méthode des éléments {inis.

Ces soulèvements Instantanés ont été obser­ vés dans la pratiquev lors de 1 'exécution Exponsioo des excavations du métro de la ville de Me­ 100 Prof, excov.

------i x r

E * <000 kg/cm 2 0.30

» 1 Echelle des longueur*

► ^ 9 Echelle des depiocements

Distonce du repère ou bord de rexcovation Argile Profondeur de l'excavation E » 50 kg/cm* y • 0.45 Fig 5 Expansions a proximité des excava­ tions. Centre Urbain Présidente Juarez Ces exemples montrent qu'à court terme -rTT les mouvements d'expansion élastiquetà pro­ ximité de l'excavation ne sont pas négligea Fig 3 Déplacements du pourtour de l'excava­ bles. Leurs effets sont divers et, en ce tion. Méthode des éléments finis qui concerne les excavations dans la ville de Mexico, on leur attribue la naissance de fissures de tension, aussi bien dans la zo­ Deplocement ne du fond de fouille que dans les talus et en cm les zones proches des excavations. 2 - 'Surfoce le —- Surface dé référence Ces fissures à leur tour, peuvent modl I5 /I/6 9 18/1/69 l - fier radicalement la forme des surfaces po­

0 4 6 8 I0 12 tentielles de rupture, qui dans de nombreux cas deviennent des plans passant par le D istance ó I axe de -I5 /I/6 9 Prof.en f l'excavation en m ------2 pied du talus2. Pour restreindre l'ampleur de l'expan­ ■18/1/69 sion élastique du fond de l'excavation on a Siphon Dr. Navarro mLaém souvent utilisé des pieux ancrés à grande profondeur et fichés dans le terrain avant de procéder à l'excavation. Dans ce cas, les mouvements d'expansion sont fortement réduits, 1'argile adhère aux pieux qui a leur tour induisent dans le terrain des efforts tranchants dirigés vers le bas. Ce­ tte restriction des mouvements élastiques due à la génération d'efforts tranchants en tre terrain et pieux influe sur la valeur ~ des poussées latérales.

POUSSEES LATERALES SUR LES ETAIS DES EXCAVA TIONS

Les poussées latérales sur les étals des excavations creusées dans 1'argile, sui vent des lois différentes d'après la vaJ.eur du coefficient de stabilité Nb. Fig 4 Expansions a proximité des Lorsque la résistance du sol, compris excavations entre la surface du terrain et le fond de

351 SEANCE PLENIERE 4 la fouille, n'est pas prise en compte, on démontre théoriquement que, dans le cas

d'un problème à deux dimensions, le terrain Stratigraphie commence à se plastifier près du fond de l'excavation lorsque Nh = 3.14. Pour une va w N.P. 1.50 m leur de Nb égale à 5«14, la rupture se pro­ duit. Pour un problème a 3 dimensions la va Veinée de sable leur de Nb qui correspond à la rupture va­ Racines fossiles rie entre 6.2 et 9.1. Nous pouvons donc con sidérer que pour Nb inférieur à 4 ou 5 le Veinée de soble problème pose" est essentiellement élastique, alors que pour Nb supérieur à 6 le problème Veineé de sable doit être traité par la théorie de la plas­ Veinee de sable ticité'.

1. Excavations dans les argiles avec N b < 4 ou 5» Lit de sable Dans ce cas, il est raisonnable <^e traiter le problème au moyen de ^ a /theorie de l'élasticité. C'est^ce qui a ete fait, pour interpréter les résultats de poussees latérales obtenus avec les stations de mesu re du métro de la ville, de Uexico. Veinéíe de sable L'exemple présenté dans les figs 6 à 9 correspond a une des stations de mesure de la ligne No. 1, placée à hauteur de la rue Medellin. L'excavation de 7 n de profondeur • T. Scissomele de labo. a été découpée dans un terrain essentielle­ o R, Boite de cisoillement ment argileux. La fig 6 présente la coupe * CS, Compression simple stratigraphique du sous-sol à cet endroit. La cohésion de l'argile y est égale en moye Symboles nne à 2.5 t/mz et le coefficient de ^tabilT l,»VI Remblai Gravier t é Nbvaut 3.6.^'excavation, flanquée de, ~ U////À Argile l ' . V I Fossiles deux murs de béton moulé dans le sol, a été r v V 'l Sable vitrifie' étayée à 2.00t 4.00 et 5.50 m de profondeur sut La figp represente la structure du tunnel E57E] Sable terminé et l'on peut y distinguer claire­ ment le caisson intérieur flanqué des deux palplanches en bé,ton armé. La rigidité' des Fig 6 Coupe stratigraphique du terrain. palplanches est égale a 239 x 10“tXmz/m. Station de mesures Medellin Les résultats des pressions totales, mesu­ rées à. l'aide des vérins plats scellés dans les murs moulés, sont présentés dans la fig 7. On peut y remarquer la très faible dispersion des valeurs enregistrées à 4 et - r -L 10 m de profondeur du 7 Juin au 25 Juillet 50

1968. Les valeurs de la pression horizonta­ 40 le, enregistrée par le vérin supérieur, pla­ ce' à 1.80 de la surface du sol, sont beau­ coup plus variables durait de la présence a ce niveau d'une rangee d'étais dont la charge a été parfois considérable. Le 24 Juin, par exemple, la rangée supérieure d'étais a été chargée à 30 tonnes. Il convient de noter aussi la réduc­ tion avec le temps des pressions laterales. Des piâzometres furent installés dans une perforation çlacée à 50 cm du mur moule' et aux mêmes élévations que les vérins plats. Les pressions Interstitielles ainsi mesu- rees ont pu etre décomptées des pressions totales. La fig 9 en traduit les re'sultats. La constance du coefficient de poussée K, calcule' en fonction des efforts effectifs, est remarquable, ce qui implique que le diagramme des pressions effectives est Mesures en cm triangulaire. Beton fc=175 kg/cm2

L'exemple présenté dans les figs 10 à Fig 7 Structure du tunnel. Ligne No. 1. 13 correspond à un cas similaire à l'anté­ Station de mesures Medellin rieur. Il est intéressant de noter cepen­ dant que,bien que la rigidité de la palplan che soit dans ce cas double de celle de

352 MAIN SESSION 4 l'exemple précédent, les valeurs du coeffi­ cient K, calcule' en fonction d ’efforts Pressions horizontales totales en kg/cm2 effectifs, sont sembables dans les deux cas. 0.2 0.4 0.6 0.8

» L L * L P Cohesion, av tot , • W, % k g /c m * t o n / m * Stratigraphie 0 250 500 0 0.5 10 0 5 10

Le 24 Juin, les étois proches de lo surfoce sont chorges 0 30 f La construction est termine^ le lo. Juillet 1968

Poche de sable

trv tôt , Pressions verticoles totales • T, Scissométre de lobo o R , Boite de cisaillement Fig 8 Pressions horizontales totales. Symboles Station de mesures Medellin IM.»J Remblai [îïw \ Srovier Y //// X Argile l'.1.1.1.1.! Fossiles 13-fcl Silt l y . y i SoUle vitrifie' ES3Soble

Fig 10 Coupe stratigraphique du terrain. Station de mesures Buenavista

80 I- - 720 -

800 - M esures en cm Fig 9 Quotient K des efforts effectifs fro­ Be'ton fc = 140 kg/cm2 rizontaux et verticaux. Station de mesures Medellin Fig 11 Structure du tunnel. Ligne No. 2. Station de mesures Buenaviata

353 SEANCE PLENIERE 4 Pression horizontale Pression interstitielle ma. On peut y remarquer que la zone limitée totole en kg/cm2 en kg/cm2 par la courbe correspondant à des efforts tranchants maxima de 3 t/mz (ce qui est une valeur moyenne de l a /cohe/sion de l'argile), est réduite. En conséquence le comportement est essentiellement élastique.

Fig 12 Pressions horizontale totale et in­ terstitielle Fig 14 Courbes d'égal effort tranchant o maximum en tonnes/m

Les pressions totales calculées par ce tte méthode, en supposant de plus que le coefficient de poussée au repos de l'argile est égal à 0.5, et que le pompage, du fait de la rapidité’ de la construction, ne dimi­ nue pas les pressions interstitielles dans l'argile, sont présentées dans la fig 15 ;on y montre aussi les valeurs des pressions to taies mesurées directement à la station ~ d'observation de la rue Merida, dont les ca ractéristiques sont proches de celles choi­ sies pour le calcul. Les expansions du f ond jie la fouille, calculées par cette méthode sont égales à 13 cm, alors que l'expansion obtenue par me sure directe {fig 2) est de l'ordre de 15 cm. Il convient de souligner enfin,l'ac­ cord satisfaisant observé entre les déforma tions du mur calculées et mesurées directe­ ment à la station de la rue Merida. Cette concordance d'ensemble entre le calcul et les mesures directes, confirme la validité du calcul élastique pour des exca­ horizontaux et verticaux vations dont le coefficient de stabilité Cette similitude des valeurs du coeffi est inférieur à 4. cient K est due, semble-t-il, aurait que En conséquence il semble recommenda- le terrain se comporte de façon élastique, ble, pour ce6 cas, de calculer les poussees les déplacements du mur étant faibles. En réelles, soit par la méthode du coefficient conséquence, un exemple type d'excavation a de pression au repos , soit par la méthode été choisi et traité par la méthode dee éle des éléments finis. La réglé donnée par ments finis. Les valeurs des coefficients R.B. Peck suivant laquelle, dans ce cas, d'élasticité des terrains ont été choisis en les poussées sur les étais peuvent être cal fonction des résultats de nombreux essais culées en supposant une poussee laterale tant de laboratoire que de chantierM La réelle p , variable entre 0.2 y H et fig 14 en traduit les »e'sultats, relatifs à 0.4 y H est probablement valable lorsqu'il la distribution des efforts tranchants maxi s'agit d'argiles saturées, mais avec une

354 MAIN SESSION 4 Pressions horizontales totales en kg/cm2 2. Excavations dans les argiles aveo Nb > 5 . Lorsque le coefficient de stabilité Nb de l'excavation est supérieur à 5, il se forme une zone plastique près du'fond de l'excavation, zone plastique dont les dimen sions augmentent lorsque Nb augmente, jusqu’au moment de la rupture du fond de fouille. , Les considérations, propres du cas an­ térieur, ne sont plus valables. Il est nécé ssaire^de se baser, lorsque Nb est supé­ rieur à 5, sur une théorie de la rupture pour calculer les pressions latérales. Par la théorie classique de^ankine, on obtient la valeur de la poussée totale:

v H 2 ^su

P0 étant la poussée horizontale totale H la profondeur de l'excavation Su la résistance au cisaillement non drai­ ne" de l'argile y le poids spécifique du terrain.

La valeur max^ma de la pression latera Fig 15 Poussées,totales sur les murs le apparente, donnes par la régie empirique latéraux du tunnel de R.B. Peck, provient de cette analyse et vaut (yH-4Su). Avec la distribution des nappe phréatique profonde. Lorsque le ni­ pressions apparentes, proposée par veau phréatique est proche de la surface co R.B. Peck, la valeur de la poussée totale Q mme pour le cas des argiles de México il survies étais de l'excavation est donc éga­ faudrait appliquer la formule: le à: 4S P = + 0.4 (y H - ywh) (1) Q =. 1.75 [té y 1^(1 - -y§)] / / avec Il est bien evident que cette poussee totale Q doit âtre supérieure ou au moins yw poids spécifique de l'eau égale à la poussée de 1 1 eau sur la palplan- h différence d'élévation entre le che, ou à plus forte raison sur le mur mou­ foru} de fouille et le niveau le" dans le sol. En effet la perméabilité" du phreatique mur moule" dans le sol est très faible. Il ( y H- y h) pression verticale effective faut donc vérifier que: H w profondeur de l'excavation. 1.75 [té y H2 (1- téyw H2 Il est, en effet,fort peu probable ^u e la poussée en^un point du mur soit inférieu re a la poussée due à l'eau. Or, s'agissant soit 1.75 (1 - fl) * - y - d'argiles peu permeables, la diminution des 7 pressions interstitielles provoquée par le Pour le cas du siphon de Morazan^ pour pompage est à court terme imperceptible. lequel N = 6 et y= 1.2 t/m3 , il s'avère La formule (1) coincide avec celle qui que cette inégalité n'est pas vérifiee. correspond à la méthode du coefficient de La régie empirique de R.B.^ Peck, dans poussée au repos, avec une valeur du dit ce cas, prédit une poussée latérale totale coefficient égale à 0.4. La distribution inférieure â celle due à la simple présence des pressions proposées par R.B. Peck est de la nappe phreatique. Ce résultat n'est cependant très différente de celle obtenue pas digne de crédit, et il serait sans dou­ par mesure directe ou par la méthode du coe te recommendable de modifier cette règle de, fficient de poussée au repos. Il ne faut _ façon à ce que la valeur de, la poussee late pas perdre de vue, à ce propos, que la ré­ raie totale soit au moins égalé à la pous­ gie empirique de R.B. Peck s'applique aux , sée de l'eau. , , charges des étais et non aux pressions late Pour le cas des excavations realisees raies réelles. Le diagramme des pressions dans les argiles d'Oslo et de Mexico, avec latérales réelles du sol contre le mur peut N> 5, les charges des étais ne vérifient être triangulaire, sans que le diagramme pas la règle empirique de R.B. Peck. Les me­ des charges sur les étais le soit. Cette sures effectuées dans ces cas, montrent que apparente contradiction est due aux déplace les réactions des étais sont nettement supé­ ments des appuis des étais pendant la cons­ rieures à celles observées ailleurs. truction. Ces déplacements des appuis provo Cette distribution exceptionelle des quent alors une augmentation ou une diminu­ charges sur les étais peut Stre due unique­ tion des réactions des étais. ment aux déflexions subies par la palplanche

355 SEANCE PLENIERE 4 avant la mise en place des étais, sans qu'il des etançons. soit nécéssaire poux autant de considérer Il convient de noter, de plus, que t une redistribution des pressions latérales sous l'effet des grandes défonnationa la re réelles sous l'effet des déplacements obser sistance au cisaillement non draine' de l'ar ve's. Dans d'autres cas , ceci a été démon­ gile de Uexico diminue de façon notoire, la tré? Pour vérifier cette hypothèse, la pal- sensitivité moyenne de cette argile étant planche a été' analysée comme une poutre con 8. Ceci pourrait contribuer à augmenter tinue, soumise à des butées et poussées cal apprèciablement le coefficient ( 1 - 'rSv ). culées par la formule de Bankine et appuyée y H sur les étais (fig 16); les expressions des CONCLUSIONS charges sur les étais en fonction des dépla cements des appuis, qui résultant de ce cal En conclusion, il convient, a propos cul sont, pour la 4eme étape d ’excavation : des mouvements verticaux enregistrés sur le pourtour des excavations, de ne pas sous- R1= 1.0+ 884 v.j-2006 v2+1407 V y 2 8 9 v ^ estimer les mouvements d'expansion élasti­ que, qui peuvent être d'importance pour le R2= 8.3-2006 Vj+5436 v2-5140 v3+1735 v4 cas des argiles de faible resistance. ]*|n ce qui concerne les poussées latéra R3= 15.6+1407 vx-5140 v2+6894 v 3 ~ 3 2 7 3 v ^ les reelles, pour les excavations a coeffi­ cient de stabilité inférieur à 5, les mesu­ R^= 39*4- 289 Vj+1735 v 2 - 3 2 7 3 v 3 - 1 9 3 2 v 4 res effectuées dans les argiles de Uexico P = 8.7+ 4 vx- 25 v2+ 112 v3- 105 v^ prouvent qu'elles peuvent être calculées par la méthode du coefficient de repos. Te­ Dans ces calculs l'effet#de l'excava­ nant compte de la présence d'une nap^e tion générale de la zone, ante'rieure à phréatique superficielle et de la très fai­ l'exècut^oij de la fouille proprement dite ble perméabilité des argiles, la règle empi n'a pas ete pris en compte. rique de R.B. Peek peut, dans ce cas, être” L'application de ces formules pour les d'une application dangereuse. En effet, la valeurs mesurées des déplacements relatifs poussée totale, calculée d'après cette rè­ des appuis, l'extrèmite inférieur«de la pal gle, peut être inférieure à la poussée de planche étant prise comme référence, four­ l'eau. f % nit les résultats pre'sente's dans la Table I Lorsqu'il s'agit d'excavations a coeffi cient de stabilité supérieur à 5, les dé­ Table I flexions subies par la palplanche avant la mise en place des étais sont très importan­ Re'actions des appuis en fonction de tes pour le cas des argiles de la Ville de leurs déplacements relatifs O LL Cohesion * LP kg /cm 2 Deplacement Reaction Reaction • W,% Appui relatif en calculée mesurée 0 250 500 0.5 IX) cm ton ton ‘.V.: Sable et gravier 1 - 2.5 - 3.9 6.5 N.F. 4.00m 2 - 5.4 14.2 15.2 3 - 9.0 8.8 8.0 ÉVeinée de sable 4 - 12.3 9.9 6.5

Veinée de sable " Il est evident diaprés ces résultats que les charges mesurées sur les étais de Poche de sable tête sont supérieures à celles données par - le calcul. En ce qui concerne les étais plus profonds la concordance est acceptable. Veinée de sable Il faut donc admettre que la distribu­ i tion des poussées latérales qui a servi de ////, base au calcul n'est pas correcte, surtout pour la zone superficielle. Lors de la mise 1 en place de la première rangée d'étançons, Veinée de sable la (pre'char^e (de 12 tonnes qui a été' appli­

quée a généré un état de butée dans le te­ • T , Scissométre rrain. VA de labo. o R , Boite de Cependant le calcul montre que la char cisaillement ge de 15 tonnes mesurée dans les étais de Symboles » CS , Compression simple la deuxième rangée est due uniquement aux ftVxVl Remblai Gravier dénivellations des appuis des étais. Il est probable par conséquent, que l'enveloppe pro A rg ile Fossiles posée par R.B. Peck pour le calcul des char Sable vitrifie ges sur les étais ne Rapplique pas au cas de Uéxico, du fait des déplacements excep- tionellement grands de la palplanche et des Fig 16 Coupe stratigraphique du terrain. concentrations résultantes des reactions Siphon de liorazan

356 MAIN SESSION 4 Mexico. Il a été prouve que les concentra­ this purpose I will divide my statement in twe tions de charge sur les étais sont dues, parts 1 1) State of stress at rest in preconsolidated tout au moins partiellement, à ces défle­ xions, sans qu'il soit nécéssaire pour au­ soils and 2) Stability of the bottom of excavations tant de supposer une variation des poussées in preconsolidated clays. réelles du terrain en fonction des déplace­ ments. De plus, ces grandes deformations State of stress at rest in preconoolidated clays provoquent sans doute, une diminution notoi re de la résistance au cisaillement non At rest, the initial state of stress in the ground drainé de ces argiles sensitives. Cet effet contribue à l'augmentation du coefficient de is defined by a principal stress ratio Kg, whose poussée. ^1 est souhaitable de considérer value depends on the geologioal history of the soil. la dite reduction de resistance pour le^cal­ This ratio is not necessarily a property of the cul de la poussée latérale maxima d'après la material, but rather the product of a state of règle proposée par B.B. Pack. A propos de deformation and, therefore, it may be smaller, equal cette règle, il faut souligner que la pous­ or larger than one, depending on the geologic sée totale qui résulte de son application, peut encore être inférieure à la poussée to process that led to the formation of the deposit. tale de l'eau, lorsque la nappe phreatique” est superficielle. Existing information indicates that in normally oon solidated clays Kq - 0.6. It has been shown REFERENCES (Skempton, 1961) That its value increases in olayB preoonsolidated by the load of deposits that were 1. Sistemadle Transporte Colectivo "Mesures later eroded, in proportion to the ratio of preoon- effectuées lors des excavations des Si­ phons de Dr. Olvera et Dr. Navarro", eolidation stress and present overburden stress. Rapport preparé par Solum, S.A., México, Since unloading due to erosion develops under a 1969 condition of one dimensional expaneion, the 2. Reséndiz, D. 7 Zonana J. "The Short-Term prinoipal vertical stress varies without a Stability of open Excavations in Mexico proportional change in the horizontal prinoipal City Clay", Volumen Carrillo. Proyecto stress, because the material oannot expand Texcoco, 1969 3. Marsal, R.J. y Mazari, M. "El subsuelo horizontally! a state of preoompression remains de la Ciudad de México", Facultad de In­ that rises the original to Valued that may be­ geniería, UNAM., México, 1959 come larger than one. On- the oontrary, if preoon- 4. Reséndiz, D., Nieto, J. y Figueroa, J. solidation was due to partial desicoation, with the "The elastic properties of saturated development of capillary etresBes that induced an clays from field and laboratory measur­ all around hydrostatio condition, during this ements", III Panamerican Conference on prooess, moves toward one, beoause its value be- Soil Mechanics and Foundation Engineering Vol. I, pp. 443-466, Caracas, 1967 Gfc 5« Tschebotarioff, G.P. "Soil Mechanics, comes equal to K0 ------where G . is the Foundations and Earth Structures", &Z + G k Me Graw Hill, pp. 488, 1951 6 . Skempton, A.W. y Ward, W.H. "Investigat­ ions concerning a deep cofferdam in the average all arouiv. uyai-o6tatic oapillary stress at Thames Estuary clay at Shellhaven". depth z. As it dries, ths soil contracts and, Geotachuique, Vol. Ill, pp. 119-139, 1952 depending on the degree of deeiccation attained, it 7. Rodriguez, M. y Flamand, C. "Strut loads may or may not orack, as indicated in fig. 1, where recorded in a deep excavation in clay", tho mechanism of preconsolidation by desiccation is Proceedings VII Int.Conf.on Soil Mech. and Found. Engineering, Vol. II, pp. 459, represented. México, 1969. Should the soil become saturated again, an expanmin is produced that may not compensate the previous Chairman O. MORETTO contraction, as indicated graphically in the above mentioned figure. Upon saturation, the capillary An interesting contribution that I am sure streesee disappear and the eoil expands trying to it brings in consequence an interchange of recover its original state. However, since the r£ ideas between you and Dr. Peck. bound of the material is much smaller than its compressibility, it cannot recover the lateral As I have made in the firet part, I am goipg compression undergone previously and the result is to read a little discussion on this subject. that the vertical stress practically does not change, because the thickneee of the depoeit varieB I would like to make a fev remarks on the stability very little but, on the contrary, the horizontal of the bottom of exoavations in preoonsolidated stress may likely become highly relaxed, as only olaya to bring forward the difference in behaviour part of the contractions may be recovered. In that may develop depending on whether preconsolidat^ thie event, the value of Eq would be not only ion was reached by a load that was later eroded or smaller than one but, furthermore, smaller than derives from capillary stresses due to drying. For that corresponding to the same material nornally

357 SEANCE PLENIERE 4

MECHANISM OF PWECONSOUDATlON BY DESICCATION

initiai HATt SATUWATIP HOWHAl T COWSOUPATiO P.,»K

MOMWU DtSlCCATiOM WITHOUT ntSURlNO tTHOWO OUKXATlOW WITH FlttuWlMO

£ I ■ C . : G. > K#1G',

m i m i * L im

rni «iv uni ■¿¡U

A* sCy*K*»G‘t L _ r

Fig. 1 - Mechanism of Preconsolidation by Desiccation

consolidated, before the desiccation process The observations made recently in connection with the started. excavation required to seat the cut-off part of the core of a dam may be typical. The oore is resting ¡stability of the bottom of excavations in precon-. on soft silt and clay rock with an unconfined solidated clays. compreeEive strength of about 15 kg/cm^, a formation that the subsoil investigation made for design Since in preconsolidated clays there is no danger purposes showed to be highly impermeable with sample of a real bottom failure, the susceptibility of recoveries that, after the few upper meters, reach the bottom of an excavation in these soils depends one hundred per cent. only on the process that led to preconsolidation. The excavation removed some 10 m of alluvium and If this process yielded a value Kq larger than one, cut a trench 5 m deep in the soft rock to enter into then the stresB release produced Tjy excavation may the intact non wheathsred material. In some shift the state of stress nearer to the failure sections, several hours after excavation had condition, as indicated in the Mohr diagram of the finished, the bottom started to rise and a set of right hand side of the fig. 2 . fissures opened up in the surface. Grouting to seal the fissures showed them to extend to a depth Should a failure condition be reached, only a rather exceeding 5 m below the bottom of the excavation. email deformation is needed to release the horizontal stress and ease the soil. Because of this reason, When the process that led to preconsolidation is the consequences for the stability of the bottom are the one shown in fig 1 , the stress release that usually minor, as they commonly lead only to an derives from sxcavation shifts the state of stress opening of fissures and joints, coupled sometimes nearer to a hydrostatic condition, as indicated in with a local bulge. the Uohr diagram of the left hand of fig. 2, and

STRESS PELEASE DUE TO EXCAVATION IN HIGHLY PRECONSOLIDATED CLAYS

Fig. 2 — Stress release due to excavation in highly preconsolidated days

358 MAIN SESSION 4 the soil in the bottom, away froc the foot of the saturated clayB. We are then dealing with cut, tends toward a more stable condition than it total stresses. Now if there happen to be had before excavation started. This may explain some sand layers carrying hydrostatic pressure, the feeling of ease that some soils or some fissures carrying free water in the clay, it is taken for granted that’there is preconsolidated by desiccation give when observed enough initial drainage to remove the water at the bottom of excavations performed on them. pressure from these free-draining elements in Typical is the city of Buenoe Aires, whose town the soil behind the cut. But, after the ex­ area is underlain by very deep deposits of clay and cess pressure is bled out of these pervious silty soils highly preconeolidated by desicoai-ioa elements, there still remains a large mass Excavations over 20 m. in depth have never shown of soil whose shear strength controls the b£ haviour of the material behind the excavation. any type of distress on their bottom, where the This soil is essentially impermeable. Hence, soil appears to be easily stable. Deformations the pore pressures do not change significant are known to be very email with a slight bottom ly, or at least the water content does not ri Ea c;’ an "elastic" nature. change significantly during the period of construction. Therefore we should be talking only in terms of total stresses and undrained REFERENCE shear strength. Under these circumstances there is no theoretical reason why the earth Skempton A.ff. (1961) "Horizontal Stresses in an pressure must be as greet as the water pres­ Over-Consolidated Eocene Clay", Proc. 5tl1 Internat. sure. It can, indeed, be less than the water Conference on Soil Mechanics and Foundation Eng., pressure; that is, the water pressure that one would consider if there were no drainage Vol. II, Paris, 1 9 6 1 . from the joints or from the pervious zones. The earth pressure that we experience is a Chairman 0. MORETTO function of the shear strength of the clay, and if the shear strength is great enough, the total lateral pressure as related to the It is 12:30 and if you have any objections, undrained shear strength of the clay can in­ we can follow with the session for a while deed be less than the hydrostatic pressure and then proceed to the PannelistB discus­ that would exist. One should always drain sion, which I think will be as interesting off this hydrostatic pressure. If one can as it was in the first part. Anyway, we not drain off the hydrostatic pressure in the will reach as maximum until one o'clock. free-draining material, then by all means the Taking your approval as granted, I give the bracing has to be designed for water pressure word to the General Reporter who will lead as well as earth pressure. the last part of this session. Of course I do not need to explain that the time had Prof. Jennings on one hand has suggested been tyrant with us and naturally we have no that the pressures, in terms of hydrostatic possibility to invite the audience in order equivalent fluid pressures indicated by the to provide their oral contributions, unfor- full trapezoids are too high. I am not quite tunatelly they must present them in writing. sure how he has calculated the 36 to 72 Dr. Peck, please. lb/ft3 density. I should point out again that the trapezoidal diagrams, being enve­ lopes from which strut loads are computed, General Reporter R. B. PECK always refer to considerably more earth pre£ sure than would really act on any given sec­ tion. Perhaps the high values given by Prof. Gentlemen, I would like to compliment the Jennings have their origin in failing to con members of the panel for producing extremely sider that these trapezoids are supposed to interesting.information, good factual data indicate envelopes for all strut loads; they that we can add to our storehouse of knowl­ are not really pressure diagrams. edge. I would also like to acknowledge the assistance I received from Mr. Harvey W. However I would have some miBgivings, I Parker in the preparation of the second part think, about designing the bracing for very of the State-of-the-Art Report. deep cute for a 15 pound fluid pressure. I am willing to admit the point that Prof. There have been two reiatea objections to the Jennings cuts Btand up, because obviously General Report which perhaps .1 could lump to­ they do, but before we can have much discus­ gether. The report must not have been as sion about these numbers I would think we clear as it should have been or otherwise the would have to get from Prof. Jennings some questions would not have arisen. The rules measured values of pressures or loads, and for the trapezoidal diagram relating to strut if they come out to be bo low I will be the loads, when we are talking about saturated happiest man in the world to add points fur clays with "N" valueB greater than about 4, ther to the left in the diagram. ~ should of course take into account the shear strength of the clay because, when we get to I believe we certainly owe Prof. Moretto N values greater than about 4, we have passed our thanks or pointing out that the stiff­ from an elastic state to a state in which the ness of clays has different origins and that strength of the clay iB mobilized. These we certainly get different reactions depend­ rules are undoubtedly applicable only when we ing upon whether this comes from desiccation are talking about undrained conditions and or precompression by an external load. I

359 SEANCE PLENIERE 4

think he is absolutely correct in hie inter­ Chairman O. MORETTO pretation of the significance with respect to bottom heave. Thank you very much. I believe Dr.Jennings has something to Bay. I think I have said enough to let my amiable colleagues find further points to disagree with. Panelist J. E. JENNINGS

Chairman 0 . MORETTO I would like to put a point to Prof. Peck. I believe that, if overall stability is con­ I have inmediately one point, which I think cerned, one must agree this is an undrained is directly connected with your diagram, condition, a total stress condition, because that is: which is the factor of safety you we are dealing very largely with the materi­ have to use to calculate your struts when als below the depth of excavation. Apart from you use the trapezoidal diagram. many other reasons, the excavation can happen very quickly. But when calculating earth pressures (classical earth pressure theory General Reporter R. 8. PECK ib an effective earth pressure theory, I believe), one wants to differentiate between the effective pressure and the water pressure The strut load that you calculate from the and theBe two must be taken in combination pressure diagram is the biggest load that with each other. Now it does seem to mix up you might ever get in a strut at a given things, but this is the way to approach the élévation in a given cut. There will be real behaviour and to take account of the many strut loads which are smaller than these differences which exist when you can drain values for Btruts at the given level, but and when you cannot drain. there should not be any bigger ones. In fact there might not actually be one as big as calculated, but the calculated one should be Chairman O. MORETTO the absolute maximum strufioad that would develop. Now since you are alBO a professor in reinforced concrete and structures, I Any comments Dr. Peck? leave it to you to take it from there. If I give you a load that will not be exceeded, General Reporter R. B. PECK you may decide what structural safety factor you would like. When we want to know the earth pressure for the design of a permanent structure, I think Chairman 0 . MORETTO we can say we have almost no information to go on, because we actually know very little I probably would like a factor of safety about how the state of stress becomes altered just equal to one or just a little over one, when we take out the temporary bracing and if I want io be safer than normal. put in the permanent structure. When we have a quasi-permanent structure, aB we have in a goo.ij many Blurry-wall types of construction, General Reporter R. B. PECK we probably are actually interested in the earth pressure on the outside of the wall That is obviously the right direction. itBelf. At least we are as interested in these pressures as we are in the strut loads. Yet, we certainly can not get any information Chairman 0. MORETTO about the design of the wallB themselves on the basis of the trapezoidal rulee, I think Prof. Jennings is quite right and we must go Hae any one further commente?, I think Dr. back and try to subdivide the stresses as Alberro haB something to eay. beBt we can into the pore pressures and the effective streBBes. The trouble, as I see Panelist J. ALBERRO it ie that we have, as yet, virtually no field data to see whether we have any reason able way of doing this. Je suib bien d ’accord avec le Professeur Peck pour admettre que quand il'y a des pressions We are dealing with a sort of mixed'problem: dues ïi l'eau derrière le mur il faut les it iB not really undrained and it is not prendre en compte. Je voudraib signaler, seu really drained, and we do not have enough lement, que bien des fois ceci est perdu de data to Bee whether our gueBsee are satiBfa£ vue. En conséquence, je suggère que dans le b tory. You may have noticed in the pressure cas où l'on considère que la pression interjî diagrams that Hr. Kuesel showed concerning titielle dans l'argile est égale à la pression one of the deep cuts on BAHT that it is a hydrostatique initiale (quand il s'agit des sort of a hybrid pressure diagram. There in argiles trSs imperméables) l'on ajoute à cette deed the water pressure was taken as water pression la pression due au terme de pression pressure because, as a matter of fact, drain effective, et qu'on redistribute ensuite cette age was not going to be permitted near this somme Boue forme de diagramme trapèzoidal. cut. The effective stresses were given an

360 MAIN SESSION 4 arched distribution that looks something point. The observations that indicate the like the arched distribution you get from general equivalence of settlement and lateral the trapezoids. This was done because it movement are observations that have been seemed the most reasonable thing to do, but made where the movements are generally fairly I believe there is no field evidence yet large and where some of the settlement is that it is the right thing to do. It just due to lateral movement and Bome to heave looks reasonable and until we get some data of the bottom. This in fact is the key to­ this is probably all we can do. We wind up ward one method of reducing settlements. We with an unsatisfactory situation with res­ do have procedures, of course, for establish pect to that problem. I believe this is the ing stiff walls and for establishing the State-of-the-Art at the present time. cross bracing before the general excavation is made, for the sole purpose of keeping the side walls in their initial position and re­ Chairman O. MORETTO ducing the settlement to that which is due only to what base heave remains. This is a Does Dr. Ward has something to say about it? proper procedure for reducing settlement, and you are quite right in bringing up this point. I also should have mentioned that Panelist W. H. WARD Prof. Alberro'b comments about the general rise of everything around an open cut are perfectly in order; the rise does take place, I think Prof. Peck said that outside an open while the soil is in an elastic state, and excavation the settlement he found was ap­ everybody has known for a long time that they proximately equal to the lateral deformation take place to a large extent in Mexico City of the supporting structure, is that correct? as compared to most other spots so he has had Well this is not always the case; if I un­ the opportunity to observe these movements derstood Prof. Alberro correctly he in fact much more clearly than the rest of us. was getting heaves adjacent to the excava­ tion and presumably the lateral deformation is very small in his case. But certainly in Chairman 0 . MORETTO a stiff fissured clays even if one restricts the lateral deformation of the walls to very small orders which one can do with some of Dr. Jennings has also something to say about these slurry trench method, one has to remem these movements. ber that the bottom heave has to come from somewhere on an elastic basis and it can only Panelist J. E. JENNINGS come from outside. Some of you might have seen some of the information I published a few years ago about the big Shell Center ex­ We have also observed these movements up­ cavation in London which was a very wide ex­ wards and outside the excavations but they cavation several hundred yds wide and 40 are very small and they appear to be of an feet deep. I have no precise information elastic nature, quite insignificant in compar •about the settlement outside the area and I ison with the settlements which occur in the- have no precise information about the lateral ground. yield but I do know what was been happening at levels between 4 and 33 ft beneath and outside the excavation. The ground rose Panelist J. ALBERRO beneath the excavation; it even rose under­ neath the walls and it rose a little outside the walls. But I certainly know the surface Il est évident, je crois, que l'ampleur des settled; further outside the ground settled mouvements élastiques est favorisée ici ^ so that if you take the load on the main area Mexico, du fait que le module d'élasticité you are certainly going to get downward move de l'argile est trèB bas. ment outBide, whether there is lateral yielïï or not, and these effects are certainly im­ Pour répondre aussi au Docteur Ward, je portant in some cases. I might add that in voudrais signaler que, par la méthode des the case of that particular structure which éléments finis, le volume qui est déplacé mainly unloaded the ground, the whole area aussi bien vers le fond de fouille, que dans is uplifting still. Even a 10, 12 storey la zone proche de l'excavation qui est buildings are being up lifted and being déplacée ver6 le haut, se retrouve compensé partially held down by the piles which are quand le coefficient de poisson est de l'or­ supposedly to be holding them up; and what dre de 0,5 par un affaissement loin de l'ex­ i6 more, the area that went down outside the cavation. excavation is now swelling. In other words, the swelling is spreading outside the exca­ C'est-a'dire que la compensation du volume vated area and this haB been going on for ma s'effectue, mais s'effectue loin de l'excava ny, many years. The total movement is going tion. to be many inches underneath thiB structure. Chairmen 0 . MORETTO

G e n n l Reporter R. B. PECK Is there any other comment? Well, if not I would like to put one question which I think You are of course quite right about that is very important. The tie-back anchor me-

36I SEANCE PLENIERE 4

thod has become very popular. Does the panel Chairman O. MORETTO have any experience about plastic flow of tie-back anchors in stiff clays? Does anybody else have any more comments? Panelist J. E. JENNINGS Prof. Jennings.

We have quite a bit of experience of support ing excavations with tie-badk anchors. I do Panelist J. E. JENNINGS not like tie back anchors which are not an­ chored into rock or something solid. These things do yield and, in yielding they create a great deal of trouble because they usually ride up to the top of the excavation. If you M r . Chairman at some stage earlier, Bomebody make a practice of checking the stresses in talked to me about the question of supports the anchors at regular intervals of the or­ for rock excavations and excavation of hard der of weekly to fortnightly, and if you use soils which are jointed. The behaviour is anchors which are fully anchored and stressed along-the joints. Certainly when it comeB back to at least one and a half times the to rock excavations, these are probably more anticipated pressure which you expect from difficult than the support of soil excava­ any of the methods of calculation, these tions. We know more about Boils than we do seem to behave well and retain their pres­ about rocke, but with rocks the behaviour is sure. predominantly along joints. One can find ma­ ny cases where you have no support at all or other cases where you have to provide support Cheirman O. MORETTO which is quite equivalent to that which you would provide for a soft clay or a firm clay. One has got to make an examination of the Is there any other comment? joints in the rock. In an aBsesment of the pressures, I have used largely the coulumb type, but again water pressures on the Panel iS W. H. WARD joints are very important indeed.

I have very little information to add to Chairman 0 . MORETTO this question which Mr. Moretto raised. All I can eay is that tie-back anchors are being promoted in the London flat Area. The only Coming to an end of this session I think we thing ib that we do not know how much they have to thank the members of the panel for creep. I suspect they are all going to creep a most interesting discussion on a very up- considerably but there is no data available to-date problem. Thank you very much for on this to my knowledge at the moment. your attention.

WRITTEN CONTRIBUTIONS WRITTEN CONTRIBUTIONS

G. BALDOVIN (Italy) a pu contrôler que ces tassements corres­ pondaient dans une bonne mesure à la perte de terrain à l'excavation. J'aimerais bien signaler au Rapporteur Général que l'expérience acquise pendant Une première réduction des tassements a la construction de la ligne 2 du métro de été obtenue par injection des coulis Milan a permis d'obtenir de remarquables d'argile et ciment faite à partir de la résultats en ce qui concerne la réduction route, mais l'amélioration la plus impor­ des tassements par l'avancement du bou­ tante a été obtenue en remplaçant les vides clier. laissés par l'épaisseur du bouclier par du gravier calibré injecté immédiatement à Le terrain de Milan est un dépôt alluvion­ l'arrière du bouclier môme. Par cette solu­ naire incohérent constitué par du sable et tion les tassements ont été réduits à 3 cm gravier de compacité variable et les pro­ maximum; on prévoit d'obtenir une améliora­ blèmes des tassements pour décompression tion plus poussée par l'emploi,danB les a donné de grands souais à l'administration. endroits les plus difficiles des injections chimiques. On a employé un bouclier ouvert, profilé à fer & cheval, de 6 m de largeur environ. Je regrette de ne pouvoir pas ici illustrer On a travaillé & une profondeur sous la en détail les procédés adoptés: en tous cia, route entre 6 et 12 métrés en proximité je désire signaler que les résultats de ces des bâtiments. On a constaté; au départ, études ont été rassemblés et publiés par la des tassements de la route de 12-13 cm et on Associazione Geotecnica It-iliana.

362 MAIN SESSION 4 nett addition of the superincumbent loading. L. DECOURT (Brazil) The internal diameter of the tunnel is 10.3 m and the total cover is 7-8 m decreasing to In his remarkable report, Prof. Peck has about 6 m at the south end. The tunnel is lined presented some data concerning surface move­ in rings of 27 precast concrete segments, cast ments associated with tunneling operations. to fine tolerances and abutting along convex to Construction procedures, type of soil, ground convex Joints to allow free articulation. The water conditions, geometry and depth of the tunnel are of great importance. For practic­ lining is 30 cm thick and each ring is 60 cm al application of Prof. Peck's recommend­ wide, with the ring stressed against the ground ations, it iB fundamental to estimate the by jacks at axis level, subsequently replaced maximum settlement of the ground above the by concrete. tunnel ( S'max.) under normal conditions, accidents excluded. Analysis indicated that the excavation for the tunnel should cause some small overall uplift It seems to us that the data presented are insufficient to permit correlations of S’ and that there would be a small degree of max. with all the above-mentioned factors initial and long term elongation of the vertical except, perhaps, depth. diameter of the tunnel. In view of uncertainties in the oarameters of the ground used for such We attempted a regression analysis, assuming calculations difficulty in assessing the loss of linear relationship between o' max. and Z, ground into the face of the tunnel, notwithstanding further imposing the condition that the straight line pass through the origin: the that the tunnel shield was generously epuiooed resulting equation was S' max. = 0,0025 Z. with platform rams and face ram s, the tunnel Repeating the same analysis*but without the was instrumented to record settlements, defor­ above imposition, we came .out with the equa­ mations and loads in the lining. tion: o' max. = 0,146 + 0,0034 (Z - 64). It is felt that despite the large scatter, The tunnel passes beneath Runways 5 and f, these equations are helpful for practically the latter only rarely used and encountered by representing the data presented in the report, but strictly respecting the limits the tunnel, before the form er. Since the orinci_ of the range within which the data have been pal concern was the measure of movement analysed. caused to a runway over the tunnel. a ring of lining underneath Runway 6 was selected for study. The scheme comprised:- A. M. MUIR WOOD (EngtanU

a) Surface levelling ooints tranverse to the Heathrow Cargo Tunnel is a two lane highway tunnel line, immediately above the trial ring, tunnel Unking the Central Area of Heathrow were recorded for movement vertically and ho - Airport, London to the new Cargo Terminal rizontally. Area. b) On the tunnel centre line, three 5 cm diarfle The tunnel is in London Clay beneath 3-6 m ter holes were drilled in which probes were in_ of Taplow Terrace travels and brickearth. serted to measure vertical movement at the Economics of tunnel approach gradients dicta­ bottom of the hole relative to the surface. The ted the advantage of setting the tunnel as high probes were anchored 50cm below the tunnet as possible. Investigations indicated that the invert, 90cm above the crown and 75cm below level of the surface of the clay varied only the top of the clay respectively. between reasonable limits and that, as is usual beneath alluvial deposits, the weathering of the c) 15cm diameter rigid PVC tube lined boreho­ top of the London Clay was confined to a narrow les sunk fOcm clear of and to each side of the band. tunnel and horizontal movement measured by means of a plumb bob constrained by a The tunnel design was based on residual clay pantograph mechanism to travel centrally, strengths for stability, on strain moduli and supported by wire from a drum and head - gear the estimation of initial horizontal loading in system for measuring depth and inclination. the ground for determining short term and long term deformations, using an appropriate method d) Deformation of the tunnel lining was measured of analysis1 . The clay was highly fissured and by Invar taoe with a specially designed portable slickensided, and while the pattern of horizontal straining head, with provision made for loading in the ground, ko, determined from measurements to continue during the life of samples appeared generally sim ilar to that the tunnel. previously found in the locality^, the maximum horizontal loading near to the surface was e) Loads In the ring were measured by pairs taken in design as the limiting Rankine passive of photoelastic load cells mounted at crown, loading at the time of denudation of the clay invert and at axis level at each side of the surface, based on residual strengths, with the tunnel. The type of load cell was selected for

363 SEANCE PLENIERE 4 Its ruggedness, long term reliability and Ground surtoce because the stlfTness could be made compara­ ble to that of the replaced concrete.

To» ot London Clay The results of the Instrumentation (figure 1) Indicated that ground movement on the tunnel centre line, a measurement that was repeated at intervals along the full length of tunnel cross - hatched in figure 1 , amounted to about 11mm in aggregate and varied little from a pattern which indicated about half the total being developed immediately above the shield cutting edge. Immediately above the shield a considerable reversal of movement occurred during passage of the shield and the ftill extent of about 14mm settlement was attained S cilt at the rear of the shield. At the invert, the » Ground Movement Around Tunnel M|i|i||||i|niiuiiMiimiiiii feet . . . . , . . . clay rose by about 3mm before the cutting iiclii* Heathrow Airport London Cargo'TunneI. edge of the shield prevented further measure­ The load cells indicated (figure 3) that, while ment. Generally the pattern of movement ocu_ initially the load at invert was the same as rred radially into the tunnel with the overall the load at one side of the axis while the load magnitude of loss of ground indicating an axial at the other side of the axis was equal to the movement of the tunnel face of about 15mm. load at the crown with a difference of about (figure 2). 15%, after about 600 days, the difference bet

GRAVEL TOP OF CLAY PROBEb ' C UY CROWN PROBE

INVERT PROBE Distance AHEAO of shield Distance 8EHIND shield 50ft 50 ft J ___ I___ I I I I J____ I____ I____ I____ I____ I____ L o .H 0.2 0 . 3 - 0 . 4 - 0 5 - 05

< - I __ O.l O U w 0.2 £ o i Q3 > ° 0 . 4 - ° 0.0- Ï 3 o — 0 . 5 - Ol 6—

50ft 50 ft o w ±± nr 00 J____I____I____L J ___ I___ I___ L J____I____I____L _l___ J____L J _ 0 . 1 - — O.l “ £ -(/) UJ 0 . 2 - - 0.2 X 0 . 3 - — 0 3 * 0 . 4 - — 0 4 o — 0 5 — _ — 0.5 a: (INCHES) (INCHES) (INCHES) 0 .6 -J — 0.6

0 . 6 - 0 .5 - o cr UJ 0 .4 - X 0 3 - 0 .2 - 0.1 - Fig. 1. Ground Movement With Passage of Shield 50 ft Heathrow Airport London Cargo Tunnel.

364 MAIN SESSION 4 ween the invert load and the crown loa'd was References about 30%, with the crown load about 15% greater than the weight of overburden at this 1 . Morgan, H.D. A contribution to the level. It is therefore apparent than considera­ analysis of stress in a cir ble shear forces exist between the ground and cular tunnel. Geotechnique, the extrados of the lining. Vol 11 N o. 1 (M arch 1961) 70,------1------r - | I r r m ------1...... 1------1 II I T pp 37 - 46.

2 . Bishop, A.W . Undisturbed samples of Webb, D .L. & London Clay from the Lewin, P.I. Ashford Common shaft: strength-effective stress relationships. Geotechnique, Vol 15 No • 1 (M arch 1965) pp 1 - 31.

3. Ward, W.H.& The development of earth Th om as, H . S . H ,. loading and deformation in tunnel linings in London Clay. Proc. 6th Int. Conf. I 2 3 4 5 6789» 20 40 60 80I00 5 00 0 0 0 Soil Mech. and Found. OAYS X Crown Gouge Eng. Vol 2 pp 432 - 436 4- Invert Gouge A Left Side Axis Gouges (1965). n Right Side Axis Gouoes

Fig. 3. Ring N° 111: Load Cell Measurement NEFTALI RODRIGUEZ CUEVAS (Mexico) Heathrow Airport London Cargo Tunnel.

The records of tunnel deformation (figure 4) On the atata-of-the-art raport presented to the Confarcnce, there are few theoretleei eapecta on indicate, oontrary to expectation, some small which author would cornent. Throughout the genarel increase in horizontal diameter amounting to report, end explicitly Mentioned et the conclualona an average, for six measuring points, of about of the first part of the paper. Peck stressed the 1 .5mm in the first year. This result merits Influence of Inelastic behavior around a tunnel, and comparison with the relatively deep level expe contemplates procedures to take Into consideration rimental tunnels for the Victoria Line.3 This the lnelaatlc behavior of aoll, through stress-straln time relationships, to Improve the generel under» - phenomenon does not appear previously to have tending of tunneling, ea well aa to Improve construc­ been explained in terms fully corresponding to tion proceduree. observation at the time of construction. It is suggested that at Heathrow the principal cause Nevertheless, the general report aupports the use may lie in the relatively higher horizontal of an error curve, as the only evallabia aolutlon to perm eability of the clay leading to differential axpleln the behavior, without further coimenta on the poaslble approach to underetend the behavior of the rates of consolidation around the tunnel. If this soli surrounding a tunnel. Searching through lltera> is correct the trend should oe reversed in due ture. the error curve wae first mentioned by Lltvinl- course. szyn' ', aa the result of a stochastic approech to the problesi; It Is applicable to fractured and granu- 1er materiel, assuming elements of ground to be of the same size, end of rigid characteristics. A msthe mat leal operator waa derived, and the expreaalon to ~ define subsidence at the surface, comes out to be the error curve.

It seems difficult to author to understsnd how the Inelastic behavior of soil can be teken Into conalde. ration following that epproach, end how stress-strsln time constitutive laws can be uaad. So far there Is another stochastic epproachO), of limited uae, that :an take Into consideration time effecta; the results obtained Justify the use of the vlacoelastlc behavior of a continuum, to represent the stochastic process.

Therefore, It would be possible to study s vlsco - elastic contlnuun around an opennlng, to repreaent the stochaatlc processes Involved; this possibility was explored by author™', using as constitutive laws an elaatlc volunetrlc cooponent, end a devla - 4. Deformation of Lining *iti: Tune torlc component expressed by différentiel operators, Heathrow Airport London Cargo Tunnel. representing s Maxvell-Kelvln unit.

365 SEANCE PLENIERE 4 A continuum with a circular excavation of radlous Measured field data, as wall as theoretical raeulta, rQ, locatad at a dapth H balow tha aurfaca of tha indicated vertical dlsplacemente ten to ¿waive tines medium «a* studied, to daflna tha displacement field those meaaured Instantaneously et a given aactlon, around the opennlng, ualng atrass-strain-time con* - three months after the ehleld had passed through that tltutlve lava above mentioned. A vlacoaleatlc solu­ section. Therefore, It seems to author that values, tion «as obtained, enabling the possibility to define reported by Peck on tables III end IV, should be teken all tha dlsplacemente, the etreasee, and tha strains ss Instantaaaoua valuea, due to their megnltude. Inside the continuum, In order to understend the be - havlor of soil around a tunnel.

Subsidence et the eurfaca vaa computed, obtaining a curve almllar to that proposed by Pack, although a little bit flatter, with characteristics dependent on meterlel parameters, as well as on geometric fea­ tures of the opennlng. Position of Inflection points Is a funtlon of lengtha - r\ being thle length a characteristic parameter of the problem. It wes ■1so found that shape of the subsidence curve, de pende on time, Indicating an Increase In subsidence ss time elspses.

Fig 3 Vertical displacements of the surface of the ground .over the siphon , for =2.58 Another feet cen be obtained from the vlscoelsstlc solution! the existence of horlrontel displacementa et the eurfece and Inside the contlnuuaa. At the aur- face, tha displacements are represented by vectors peeelng through the center of the opannlng, with mag- nltudea dependent on time, as well as on materiel cha racterlatlce, end geometric relationships.

Fig 1 Vertical displacements at the surface, due to a ) Vertical displacements u at *-~2 the opening 2H H C H 2H * ' > i i ___ [ -

-0.1 v =0.2 / 1/ = 0.5 02 uE i pH b)Horizontal displacements v at x = -w- 2H H C H 2 H ------1...... 1

■0.1 o CVJ II 0.2 v ~ 0-5

03 / vE i ■------^ pH Fig 2 Surface subsidence due to an opening , with H/2r0 = 3.10 Fig 4 Verticol and horizontol displacements at x=H/2 , due to the opening

Experience gained during the construction oi. cvo tunnela et Mexico City, using the ahleld technique, Inside the continuum, and around the opennlng, ver- on aandy soils laying over a deep clay stratum. In­ tlcel end horlzontel displacements cen be computed dicated vertical dlaplacaments et the surface, simi­ for different Instants, giving a cleer Idea of tha be­ lar to those defined by the viscoelastic eolutlon, havior. Theae theoretical findings had been aubatan- and an increase In displacements as time elapsed. elated by field data, uslnn vertical inclinometers

366 MAIN SESSION 4 nearby tunn*l locations, shoving • similar response Conference, an important requirement for a as that given by theoretical results. satisfactory tunnel is that its construction should not damage excessively the adjacent The viscoelastic solution can also give a vhola and overlying buildings. For the design of picture of atress, strain and dlsplacaasnt distribu­ the Sao Paulo subway, presently under con­ tion inside the continuum. Once the tunnel location struction, some of his records and criteria has been defined confutations following a program de­ on settlements, caused by shield construction veloped for a digital coa*>utar can be performed, in were interpreted and applied by Promon Enge- order to deflns atress and atrain fielda; a clear de­ nharia, in order to help specify the special pendence, on the retio H/r0 has been found, as well precautions, treatments and underpinnings ss on material characteristics. required by the buildings. Exlatanee of compressive stresses at the surface, It is easy to evaluate the numerous problems on a central zone extending a distance vH - r* , at connected with interference on traffic, exist both aidea of the point at the surface over tne center ing utilities, foundations and even under­ of the tunnel, was detected; tensile stresses develop­ ground floors, that had to be considered for ed at the rest of the surface, being their magnitude the construction of a subway in the center dependent on the ratio H/r0, as well ss on the volu­ of a city of seven million inhabitants, near metric weight of the continuusi. Tensile stresses st and under its hundreds of skyscrapers, and the surface, might be reaponsible for dsmeges and through its most important bank and business fractures, when deep tunneling is performed. streets. The 1.2 Ion of the first line of the subway, that is being designed by Promon Engenharia , (*) is at the most central point of the town, with two large stations, and with the two tunnels following beneath a 16 m wide bank­ ing street, surrounded by skycrapers, then going below the oldest historical quarter of the 400 years-old town, and then below sever­ al blocks of high and also important buildings Vertical normal stresses Horizontal normal stresses (Fig. 1). The soil profile at the site can be broadly described as a remarkably heterogeneous soil (Fig. 1), with interconnected and erratic layers and lenses of clayey sands (loose to dense) and sandy clays (soft to stiff), under lain in part of the area by a medium to coarse sand. Some of the clayey layers are preconsoli dated, but there are numerous and erratic lenses of soft clays, and also of loose sands. Water level varies from 3 to 10 meters depth. Two tunnels will be constructed at depths up to 25 meters, as also indicated in Fig. 1. Fig 5 Distribution of normol stresses parallel to the t axis.and to the y axis.inside the medium One of the problems that had to be solved was Finally, the vlacoelastlc solution could be used to the establishment of rational criteria to define pressure distribution on linings of tunnels, serve as guides for the decision on special for different stiffnesses of the lining, being an in­ precautions, treatments and underpinning teraction problem that could be solved by cosputers, buildings. Besides other criteria already in Indicating a dependence on time; this possibility haa use, Peck's records and criteria, presented not been conteiq>lated so far, by any stochastic solu­ at this Conference, were interpreted and used tion slready known. for this important decision. REFERENCES I Peck's Fig. 9 of Z/2R vs. i/R presents the 1. Peck, R. B.I "Deep excavations snd tunneling In average bands for the different types of soils, soft ground”. Seven International Conference of based on the data recorded in Table VI. The Soil Mechanics and Foundation Engineering. Mexi­ specific curves representing the same data co, 1969. p. 225 to 281. were used for the present design study. 2. Lltvlnlszyn, J.i "Displacementa in Loess Bodies ss In order to estimate the settlements, however, Stochsstic Processes". Bull, de L'Acsdemle Polonal the records given in his table VI had to be se des Sciences, 195S, III 4. analysed and interpreted. It may be observed 3. Csbrlelsen, B.L.I "Stochaatlc models for viscoels^ that the most important design parameter tic meteriels". Rilera, "Msterlsls snd Structures". should be, for a start, the estimate of the July-August, 1968. p. 319 to 326. probable maximum crown settlement, depending 4. Rodrfgusz N. "A simplified model for tunneling snd principally on the depth and dimensions of linings". Privste repftrt presented to SCT. 1969. the tunnel and the nature of the soil. Since in this point no indications were furnished, the data were investigated in search for some E. B. SOUTO SILVEIRA and N. GAIOTO (Brazil) (*) Promon's Soil Mechanics Consultant is Prof. Victor F. D. de Mello; Promon's Consult As was mentioned by Peck in his remarkable ant on shield design is "Sir William Halcrow State-of-the-Arts Report, presented to this and Partners", London.

367 T if.2 - M a t fitting fitting t a M - if.2

DEPTH (m itiri) 20 30 30 50 40 10 0 I 3 4 3 2 I 0 etvim « depth v«. voium« ment SOFT I __ ETE N VLM ( / ) ?/m (n VOLUME ENT SETTLEM — E.OSIF HARO STIFF MEO.TO n i n p n n e v r u c LY — CLAY- of settle­ of ----- 0 20 0 40 m 0 0 5 400 300 200 100 1 FNE LNE 4 E PLENIER SFANCE 368 Fig. 3 - Interpretation of settlements above above settlements of Interpretation - 3 Fig. VE RY DE N S E LOOSE TO MED TO LOOSE E S N DE RY VE ( D - Lateral distribution of settlements, no no settlements, of soele no distribution 2,(T) Curve fig. Lateral of - D ( O LT IL S 02 . 06 . 1.0 0.8 0,6 0.4 0.2 0 •o*l*, aocording to Peek's Peek's to aocording •o*l*, OL / PH <3 ) 2 ( ) m / m

a. MAIN SESSION 4 empirical correlation. Obviously, after esti­ Son diamètre est de 10 m. et la couverture mating the crovm settlement, the settlement de terre passe de 16 M. à 6 M. distribution may be derived from the curves of Fig. 9. L^s tassements produits sont dûs à deux Fig. 2 shows the best fitting curve of Vs causes principales : la fa±fl.e tenue du (volume of settlement) vs. depth, and Fig. 3 front d'attaque et la difficulté de remplir le vide annulaire de 18 cm à l'arrière du shows its logical interpretation, i.e., the bouclier. best fitting error curve of the unit Vs alorg depth, Fig. 2 being its accumulated curve. After many trials at more complete correlation Pour ausculter le teirain avant et après pasBage it was concluded that the data did not permit du bouclier, nous avons employé divers appareils s insertion of the possible influences of tunnel diameter and type of soil, the present corre­ - le pénétromètre; lation being therefore simply one connected - des repères de tassement en profondeur with the depth as a single intervening para­ consistant en un lestage, par une masse meter. de 200 kg, des tiges placées au pénétromè­ tre ; Interpreting the curve established one would - un inclinomètre de la Slope Indicator Cy; conclude that: i) the greatest settlement - des cellules Gltttzl, placées à 1 m. au­ occurs at about 2 5 meters depth; ii) at small tour du futur tunnel, verticalement, ho­ depths, up to approximately 10 meters, the rizontalement et & 45°; settlement is almost negligible; iii) below - des géocells Ménard, placées au dessus approximately 40 meters, any increase in et au niveau du diamètre horizontal du depth of the shield will have an almost neglji bouclier; gible effect on the settlements. Of course, - des cellules Geonor, scellées dans les these conclusions, as all the others connected voussoirs en béton ; with the subject, should be revised through - des extensofflètres Geonor dans les arma­ a thorough statistical analysis of the shape tures des voussoirs et dans les tirants; and equation of the best fitting curves (for - des piézomètreB Warlam. the present assumed as an error curve), which will only be possible with more data than Il y a eu deux zones d'essais, l'une dans those made available at present. 1'Yprésien, l'autre dans le Bruxellien. Tous les appareils cités ci-avant y sont With the two mentioned graphs, the lines of représentés en nombre surrisant pour per­ equal settlement along the two tunnels were mettre une interprétation statistique des calculated, for the different positions and réBultatB. depths of the tunnels, and different soils. The results served as one of the bases for Voici lee résultats obtenus : the decision of the special treatments to 1) les tassements produits à la surface et be conducted, and of the underpinnings of en profondeur montrent un affaisement the buildings, depending upon their struc­ centré sur l'axe du tunnel. tural material, shape and foundations. An On a pu en tirer la zone d'influence du intensive program of measurements of settle­ bouclier. ments of several points of the buildings is On voit nettement que le terrain se sou­ programmed, so that these criteria may be lève à l'avant du bouclier, puis qu'il checked and revised. These same criteria s'affaise en laissant une dépression pei'- were also helpful for programming the field manente au droit de l'axe du tunnel; dans observations to be carried out. ce terrain qui contient des bancs gré- zeux horizontaux, on constate que l'am­ plitude des tassements n'est pas fonction C. VINEL (Belgique) de la profondeur. 2) 1'inclinomètre met en évidence les phé- Les éléments qui 6uivjnt complètent les in­ nomène suivants : le eol commence par formations présentées par C. VINEL et être refoulé à l'approche du bouclier; A. HERMAN dans le rapportfréliminaire eoue après le passage de ce dernier le ter­ le titre "Tunnel dans le sable de Bruxelles rain supérieur reflue vers les zones dé­ par la méthode du bouclier". comprimées; à la base il y a glissement vers l'arrière, facilitée par la présen­ Le terrain au droit du tunnel comprend de ce de la nappe phréatique. haut en bas : 3) les celluleB GIBtzl montrent l'augmenta- - une couche de remblai d'épaisseur variant tion de pression initiale qui correspond de 2 à 6 m . ; à l'approche du bouclier. Une chute bru­ - une couche d'environ 17 m. de sable fin tale de pression se constate après passa­ Bruxellien, décalcifié dans sa partie ge de ce dernier. La remontée de pres­ supérieure ; sion ultérieure correspond aux injections - une couche importante de sable argileux et à la remise en place naturelle du Yprésien. terrain. La nappe phréatique Be situe à l'intersec­ 4) les cellules Ménard, comme les cellules tion de l'Yprésien et du Bruxellien. GIBtzl, montrent que c'est la cellule placée au dessus du tunnel qui atteint Le bouclier traverse toutés ces couches. la première son maximum; sa pression est,.

369 SEANCE PLENIERE 4 en effet, limitée par le soulèvement des to the excavation process were measured .by means terreB en Burface. of a W ilson type Slope Indicator, recording dally 5) les extenBomètres enregistrent nettement readings with an approximation of 0.0009 radians. les flexions de l'anneau du tunnel dans son plan et la flexion du plan lui-m8me The soil profile Is sim ilar In both cuts: superficia­ sous l'effet des vérins d'avancement du lly there Is a layer form ed by a sandy slit having a bouclier. On y remarque que les tensions thickness of about 4.00 m . Underlying this layer maximales restent acceptables et que la there exists a soft clay deposit with high com press ­ tension de stabilisation correspond aux valeurs calculées. ibility down to a depth of about 30.00 m. Figures 1 and 2 show the soil profiles at each of the cut sites; 6) la déformée du tunnel, mesurée topogra­ for sim plicity sake number 1 has been associated to phiquement , fait apparaître une asymé­ the 5.30 m wide siphon and number 2 to the 3.50 m trie. Celle-ci ns se justifie pas a wide siphon. These figures also show the values of priori dans la partie rectiligne du tun­ the natural water content (w ), plasticity lim its, n a­ nel, le terrain y étant composé de cou­ tural unit weight ( i ) and shearing strength as de­ ches horizontales homogènes. termined by unconflned com pression tests (qu/2), direct shear tests (Sp) and undratned trlaxial tests (c). Par ailleurs, les diagrammes montrent Figu res 3 and 4 show graphically the values of the l'efficacité des tirants métalliques qui

maintiennent l'anneau en-béton, malgré SHEAR STRENGTH J l'absence de butée latérale dans ce (ton/n?) (ton/m5) terrain très compressible. 1.1 It H 1,4 H La présente communication devait être illustrée en séance par une série de dia­ positives. Comme celles-ci n'ont pu être projetées, l'auteur prie le lecteur inté­ ressé de lui réclamer les diagrammes & l'adresse suivante s

Bureau d'Etudes ELECTROBEL, 1, place du Trône Bruxelles 1 BELGIQUE.

0 UNCONRNED COMPRESSION TEST ♦ S f , DIRECT SHEAR TEST 4 C , UNDHAINED TRIAXIAL TEST

Fig 1 . Cross section, soil profile and soil proper ­ ties at the site of the cut for siphon N o. 1 . J. M. RODRIGUEZ and R. LOPEZ PEREZ (Mexico)

SHEAR STRENGTH f . (ton / nr) (ton / m ) The experiences herein b riefly described are BORINS o - «00 400 « , kf. »Mi-A9 referred to cuts which have been excavated for the construction of two siphons required at the cross­ «_n±5<2____ * = 4 • ■or*tr i ings of sewage collectors and the M exico City sub­ Ì f t l « Q ____ « = * . s * way. —4- Z A\ A « r i » ____ > • Both cuts w ere supported longitudinally by a steel ' Ui P < sheet piling, hammered to a depth of about 10.00m; w - 7.60 . _ . t r * " - . - and by four levels of struts. One of the cuts, 5.30 '-LOAD dor CELL 1 m In width and 35.00 m In length reached a maxi­ A • -IO mum depth of 9.80 m; the other, 3.50 m width and t d f l. 9 0 ____ A i*o a 32.50 m In length reached a maximum depth of H t- ? 10.50 m. A SHEET •in • P ILE ^Z -, -14- 1 • cto, The strut loads were measured dally, by means of V o LIQUID LIMIT ° 1u /2 .UNCONFINED COMPRESSION TEST A PLASTIC LIMIT ♦ S f .D KEC T SHEAR TEST Freyssinet jacks assembled In closed circu it with AC , UNDRAINED TRIAXIAL TEST a 140 kg/cm2 capacity manometer; the accuracy of the jacks Is approxim ately 1 .0 ton. Fig 2. Cross section, soil profile and soil proper- The horizontal deform ations of the soli associated ties at the site of the cut for siphon N o. 2.

370 MAIN SESSION 4 apparent total earth pressure and of horizontal soi l deformation Increments up to a few hours before deformations with respect to depth for several Installing the fourth strut level; It can be seen how excavation stages at each siphon. The earth pressu­ both curves Intercept each other at approximately re was computed from the measured strut loads. Elev. -4.50 m In siphon No: 1 and at -5.50 m In To sim plify the presentation and Interpretation of siphon No. 2. results, a sequence cf numbers In arithmetic The average and maximum ratios between horizon­ progression was assigned to the days elapsed from tal (O' ) and vertical ( (Ty) total pressures observ ­ the date of the Initial measurement of the Slope ed at tie excavation stages being analyzed were com Indicator, labeled 1st day. These figures show puted and Che results are shown In figures 6-a and four types of envelopes of the computed earth pressu 6-b with respect to the ratic between the depth at the re at different excavation stages. The solid Une cut H and the critical depth Hcrlt (4 c/ff)- A line ar represents the measured values and the other three relationship of such factors has been considered wtlh lines the theoretical ones among which the broken out making and Important mistake since the range of line with long dashes was obtained from Peck (1967)^ variation of the ratlcj (T^/ (fv is very narrow; the criterion considering a reduction factor "m " of 0.4 , fact that such relationship Is approximately linear which Is the value recommended for the case of implies that the reduction factor "m " varies In México City, provided the stability number (N) be the opposite way to the one proposed by Peck when greater than 4 In siphons Nos. 1 and 2, the N value It Is correlated to the stability factor "N"(H gure resulted 4.5 and 5.3 respectively* the curve with a 6-c). dash and two dots is the one proposed by Brlnch Hansen (1953)2 after reducing the undralned shear strength by the specified safety factor of 1.5. It Is worth mentioning that the rotation of the sheet pll Ing did not take place around the first strut level as assumed by Brlnch Hansen, but Instead It occurred around a point located between the second and the third strut level, which corflrgns the statement of Rodríguez and Flam and (1969) , referring to the fact that the second strut level becomes In general the most heavily loaded.

30 t» doy fro « 33 rd to 33 tli doy

i r wot jo r u 38 th doy ~ 3 € J i_ d o y D W T J H "OOWOUtZ A W FLAMAND

Fig 6. a) Relationship between the ratio H/Hcrtt> > 59 til Joy and the ratio ((¡V -/ C”v ^av e' b) Relationship between the ratio H/Hcf

From the above mentioned discussion, it was concludes that an earth pressure envelope, sim ilar to the one proposed by Peck (1967) in which the horizontal pressure is defined by the total vertl- It cal pressure (d\y) and by the average ratio G^/CJ v obtained from Figure 6-a, provides acceptable HORIZONTAL SO IL HORIZONTAL SOIL results which only dlfTer in ± 20% from the m easur­ DEFORMATIONS, » . DEFORMATIONS, c«l ed ones, with the only exception cf those correspond («) -n doy - (b) Ing to the second strut level of the siphon reporte d Fig 5. Variation of horizontal soil deformations by Rodrfguez and Flamand (1969). This envelope for different Installation dates: a) at siphon is shown In Figures 3 and 4 by the broken line with No. 1; b) at siphon No. 2. short dashes.

In Figure 5 the solid line shows the increments In ACKNOWLEDGEMENTS horizontal soil deformation with depth occurred during the time elapsed from the Installation of the Thanks are due to M essrs. Rogelio López F . and second strut level to Immediately before the placin g Salvador Peredo for their valuable assistance. of the third strut level and the dotted line shows the

37I ELEVATION , m. ELEVATION, a. . ELEVATION, a. ELEVATION ONS, . B « , S N IO T A M R O F E D ON STASE S A T S N IO T A V A C X E T Î N I F ) < ( ( c ) T H IR D E X C A V A T IO N S T A t E t A T S N IO T A V A C X E D IR H T ) c ( OIONA ML M NTAL NORIZO DcromAnoM a FRT XAAIN STASE EXCAVATION FIRST (a) ONS , . » , S N IO T A M R O F E D ZONTAL SOIL O S L A T N O IZ R O N ONS, L M , S N IO T A M R O F E D RD EXCAVATI E S A T S N IO T A V A C X E D IR H T ) c ( F lg 4 . V a r ia tio n o f a p p a ren t ea r th p r e s s u r e a nd o f f o nd a e r u s s e r p th r ea t ren a p p a f o n tio ia r a V . 4 lg F f o d n a e r u s s e r p th r ea t ren a p p a f o n tio ia r a V . 3 lg F ,

ma for differ va on s for si r o f s e g a st n io at av c x e t n re e f f i d r o f s n o i at rm o f e d i s r fo s e g ta s n tio va a c ex t n e r iffe d r fo s n tio a m r fo e d ------r t S — ------IT F SYMBOLS OF LIST IT F SYMBOLS OF LIST op* pr*p*sod o s * p * r p * p lo e v n E neo* p r Envelop* o*o**w# ) v « W ( * l t * r * tb tijftM ENE LNE 4 E PLENIER SEANCE el r t . t . f l f « ro f d lM te b • «1 4 v * t * f th * t * t * l l * t * t * th f * t * v oser) f r * « flg. flg. « * r f 372 H M r m n o m h *4 te 0 by v ) ) v 6 h o r izo n ta l s o il il o s l ta n izo r o h h o r izo n ta l s o il il o s l ta n izo r o h h No.2. 2 . o N n pho b SCN ECVTO STASE EXCAVATION SECOND (b) n on h p OIOTL IL O S HORIZONTAL ONS, . * e , S N IO T A M R O F E D Z AL IL , O S S N TIO A M R FO E D L TA N IZO R O H UT EXCAVATI E S A T S N IO T A V A C X E OUNTN F ) 4 ( (b) E I A T ( N IO T A V A C X E N T N U O F ) 4 ( ZONTAL SOIL L A T N O IZ R O N ZONTAL SOIL O S L A T N O IZ R O N ONS, . » , S N IO T A M N O T M CFORMATI , . a « , S N IO T A M R O F DC EOD ON StASE S A t S N IO T A V A C X E SECOND N o .1 . .1 o N

H T R A E L A T O T T R E R A P R A , E R U S S E R P m* /m m h SESSION 4 vatlons made to construct a siphon at two cross­ REFERENCES. ings of sewage collectors and the Mexico City subway. 1.- TERZAGHI, K. and PECK, R. B. (1967). "Lateral Supports In Open Cuts". Soil Mecha­ The dimensions of the excavation for siphon No. 1 nics In Engineering Practice. John Wiley and are 35 m in length and 5.30 m In width with a Sons, Inc. U.S.A. pp. 403- 413. maximum depth of 9.30 m and those corresponding to siphon No. 2 are 32 and 3.50 m respectively 2 .- BRINCH HANSEN, J. (1953). "Practical Earth with a maximum depth of 10.50 m. In both exca­ Pressure Problem s. Earth Pressur-e Calculat­ vations a sheet piling was driven to a depth 8.0 m ion". The Danish Technical Press. Denmark, below the bottom thus restricting the horizontal pp. 220 - 222. flow of ground water (see Figs. 1 and 2).

3 .- RODRIGUEZ and FLAMAND (1969). "Strut Loads Recorded In a Deep Excavation In Clay". Proceedings of the Seventh Interna­ tional Conference on Soil Mechanics and Foundation Engineering, Vol. 2. pp. 459 - 467.

J. M. RODRIGUEZ and C. A. MELGOZA (Mexico)

The theoretical analysis of this phenomenon was performed by Juárez Badillo and Rico Rodriguez (1967)'* and It Is explained In a sim ple way as follows: let us consider a clayey material having a horizontal surface with ground water level at the surface. A soil with these characteristics will have the very well known distribution of total, effective and pore pressures.

If It Is assumed that a very rapid excavation of Infinite length Is made to a depth "h", then to­ tal pressure will be decreased In the full depth by a factor ( ( h),where' J Is the unit weight of the so il. Due to the fact that a clay is being consider— ed, the effective pressure before and Immediately after the excavation will remain the same; then A PLASTIC LIMIT ° DIRECT SHEAR TEST to maintain the sam e ratio among total, effective X UN CONFINED COMWSIION TEST and pore pressures It will be necessary that the later decreases In value by the same amount h t IXCAVATION which will create tensions In the water below the fV *' bottom of the excavation and the null pore pressure S E C T IO N A-A' will be located at the certain depth under such bottom. As a result of having performed a rapid excavation of Infinite length. It has been possible to lower the pore pressure In the soil located under the bottom of the excavation.

Actually there are no excavations of Infinite

length and the total pressure will only be altered METERS down to a certain depth, below which the total pressure existing prior to the excavation will remain the sam e. Fig. 1 Layout of the excavation for siphon No. 1, soil profile and soil properties. Under the abo ^ mentioned conditions the water will flow towards the bottom of the excavation Inducing the ground water level to return to its The soil profile at both sites Is sim ilar, as shown Initial position after a certain time has elapsed, In figures 1 and 2 and It consists of sandy silts unless the flow of water Is restricted by pumping and silty sand from 0 to 4.3 m In depth. or by sheet piling. Underlying these m aterials and down to a depth of 30 m there are deposits of a highly compressible In order to study this phenomenon plezometrlc soft clay of high plasticity with thin layers of sand level measurements were recorded at the exca- and silt.

373 SEANCE PLENIERE 4

BOOMS

SILTY SAND

CLAY

Fig - 3. Variation of the piezom etric level with time for different excavation stages at siphon No. 1 .

SILT

O LIQUID LIMIT • TORVANE A PLASTE LIMIT 0 DIRECT SHEAR TES T X UNCONFINED COMPRESSION TEST PIEMNEÍÍItl? *

Fig 4. Variation of the piezometric level with time for different excavation stages at siphon No. 2.

to the pore pressures measured in the two ptezo­ m etric stations Installed at each siphon. O 10 t o 10 M ETI R S The slight variations shown in the figu res 5 and 6 have been atributed to the increm ent of the pore Fig 2. Layout of the excavation fo.- siphon No. 2, pressu-e caused by the shearing strength assumed soil profile and soil properties. to be Induced by the horizontal soil deform ations. On the other hand the theory of Bousslnesq is not fu lly applicable to the case being analysed since ThDse figu res also show the values of the natural the sheet piling produces a discontinuity in the so il. water content (w ), the plasticity lim its and the shea" strength determined from Torvane Measurements reported by Rodriguez and Flamand measurements and from direct shear and unconfl- (1969)2 for other siphon did not register the effec t ned compression tests. Two ptezometric stations of the reduction in the pore pressures probably due w ere installed at each siphon with readings taken to the fact that the horizontal soil deform ations prior to the excavation process and during its w ere very Important, with a maximum value of different stages. The variation of the ptezom etric 20 cm . level with time is shown in figu res 3 and 4 for siphons No. 1 and 2 respectively, being referred It w ill be interesting to corraborate this phenome­ to the excavation progress. non in other excavations where the horizontal flow of ground water Is restricted as well as to investi­ The curves of theoretical pore pressure were gate such an effect in excavations without this evaluated considering the discharge prodi^ced by the excavation as proposed by the theory of restriction. Bousstnesq; such cu~ves are shown in figures 5 ACKNOWLEDGEMENTS. and 6 for the different excavation stages for siphons N os. 1 and 2 as well as the values corresponding The authors wish to thank M essrs. Luis Méndez P.

374 MAIN SESSION 4

POM PRESSURE , I* lM /a POM PRESSURE , it (M/a* I 4 ( I 10 I t 14 1«

Fig 5. Theoretical and measured pore pressures at the two piezom etric stations during different exca­ vation stages in siphon No. 1 .

PORE PRESSURE , it t»»/a' Fig 6. Theoretical and measured pore pressures at one plezom ttrlc station during successive exca­ vation stages In siphon No. 2. and Andris Tenlente for their Invaluable assis­ de Suelos", Vol. II, Ed. Revista Ingeniería. tance . 2. J.M . Rodriguez and C .L . Flamand (1969), "Strut Loads Reco-ded In a Deep Excavation in REFERENCES. Clay", Proceedings Seventh Int. Conf. on Soli 1 . E. JuSlrez B. y A. Rico R; "Mecanica Mech. and Found. Eng., Vol. 2, pp. 459 - 467.

375