FOUNDATION DESIGN AND BEHAVIOUR OF TOWER LATIN0 AMERICANA IN MEXICO CITY by LEONARDO
ZEEVAERT,P~.D.
SYNOPSIS The foundation design for the forty-three storey building Tower Latino Americana in Mexico City introduced new and interesting problems in foundation engineering. The Paper describes the general philosophy adopted in the design of the foundation of this building. A detailed description of subsoil conditions and mechanical properties of the lacustrine deposits encountered at the site is given. The ground surface subsidence problem and investigations performed to discover the source of compression of the clay deposits are described, and the way this phenomenon was taken into account when consideration was given to the foundation design. Excavations into the lacustrine volcanic clay deposits in Mexico City produce large heave. The Author describes the procedure used to excavate to a 13-m depth for the foundation structure, and to avoid the heave of the bottom of the excavation and the excessive settlement of adjacent buildings and streets. Settlement observations are reported-of the building, of the ground surface, and other deepseated strata. Piezometric water-level observations during construction, and afterwards, are also dealt with. Finally, a comparison of observed and computed settlements is given in an attempt to predict the future behaviour of the foundation of the building.
Le plan de fondation du batiment de quarantetrois Ctages Tour Latino Americana a Mexico a pose de nouveaux et interessants problemes de travaux de fondations. L’article decrit la philoSophie generale suivie pour le plan de fondation de ce batiment. On y donne une description detaillee de l’etat du sous-sol et des proprietes mecaniques des depots lacustres rencontres sur le chantier. Le problitme d’affaissement de la surface du sol et les recherches faites pour decouvrir l’origine de compression des depots d’argile y sent decrits, ainsi que la man&e dont ce phenomene fut trait6 lorsque fut consider6 le plan de fondation. Les excavations dans les depots lacustres d’argile volcanique Q Mexico produisent de fort soulevement. L’auteur decrit la methode employee pour creuser a 13-m de profondeur afin de mettre en place la structure de fondation, en dvitant le soulevement du fond de l’excavation et le tassement excessif des batiments et rues avoisinants. Y sont rapportees des observations sur le tassement du bbtiment, de la surface du sol et d’autres couches profondes. On traite aussi des observations piezometrique de niveau d’eau pendant la construction ainsi qu’apres. Enfin, les tassements observes sont compares aux tassements estimes dans l’intention de predire le comportement futur de ce batiment.
INTRODUCTION
The forty-three storey building property, La Latin0 Americana Seguros de Vida, S.A., (Fig. 1, facing p. 118) was constructed in Mexico City at the corner of Madero and San Juan de Letran opposite the Palace of Fine Arts. The foundation surface occupied by the building is 1,114 sq. m. The weight, including the foundation structure and 20% live load, is 23,500 tons ; therefore the unit load at the foundation slab elevation is 21.1 tons/sq. m. The building is supported on a rigid reinforced-concrete mat foundation resting on 361 concrete piles driven to a depth of 33.5 m into a firm sand layer where they act as pointbearing piles. The foundation plan and the pile layout are shown in Fig. 2. The depth to the bottom of the foundation slab is 13 m below ground surface elevation. The total depth is occupied by two basements and the foundation structure. The foundation and retaining walls have been waterproofed to obtain effective use of the buoyant forces. In order to take care of the ground surface subsidence (typical of Mexico City) as the sidewalk settles away from buildings on pile foundations, the Author recommended a special design that would facilitate the lowering at any time of the ground floor of the building. The floor was divided into panels supported on wood blocks, permitting the panels to be lowered as required. This practice will avoid in the future the necessity to construct steps into the building as the sidewalk subsidence progresses. The piles were driven from a preliminary excavation 2.5 m deep made in advance to clean the site from old foundations. After the piles were inserted a I‘ Wakefield “-type of wood 115
. 116
LEONARDO
ZEEVAERT
sheet-pile was driven in a single operation to a depth of 16 m. The wood sheet-pile served to create an impervious diaphragm to prevent water entering the excavation. Therefore the water-table in the upper pervious deposits was protected from a strong draw-down that might have initiated a large settlement of the neighbouring buildings. During excavation to the S-m depth the wood sheet-piles were shored from side to side Thereafter, the foundation beams in both the north-south and the east-west directions. were constructed in braced trenches excavated to the full depth required for the foundation structure. After the gridiron of beams was completed the panels between beams were excavated one after another, and the foundation slab resting on the piles was constructed. -4s substitutes for the excavated load, every panel was immediately filled with sand and gravel. After this the foundation was completed and loaded to obtain a reaction on the piles of 1‘2.5 tons/sq. m, equivalent to about half the weight of the building. The erection oi the steel structure then proceeded, and as more load was added the water-table was permitted to rise and exert under the foundation slab an equivalent reaction to the additional
43rd.srcrey
SAN
JUAN
DE
FOUNDATION
LETRAN
AVE.
STRUCTURE
7OUNDAliON ,-,F VI Wood Sheet-P,les
Fig.
2.
Foundation
plan
and layout
“LA
LATIN0
Forty
Three
of piles
LAYOUT
AMLRIGANA” -
5tcw.j
Bwldlng
FOUNDATION
DESIGN
AND
BEHAVIOUR
OF TOWER
LATIN0
AMERICANA
117
load. This procedure was followed until the total load of the building was applied and the water-table was restored to its original elevation. Settlement observations and piezometric water levels were carefully observed during the entire process of construction of the foundation, and thereafter. In order to design the foundation of Tower Latin0 Americana it was necessary to investigate the source of surface subsidence and the index and mechanical properties of the subsoil The results of these investigations are reported in materials at the site of the building. the Paper. SUBSOIL
CONDITIONS
The subsoil condition was investigated from continuous cores of undisturbed samples obtained from a 2.5-m depth to a depth of 70 m from the ground surface. The samples obtained were 5-in.-dia. undisturbed samples in the lacustrine clay deposits and 3-in.-dia. The results of the investigation are shown in soil profile, in the clayey sand and silt deposits. Fig. 3. The stratigraphic column was found as follows : Depth from : 0.0-5.55 m
5.555.70 m 5~70-6+30 m
6~80-6~85 6.85-7.45 7.45-7.55 7.55-9.15 9.15-l
m m m m
1-9 m
1 l-9-12.1 12.1-15.8
m m
15.8-15.85 m 15.85-16.5 m 16.5-21.4 m
21.4-21.50 21.50-22.50
m m
22.50-23.65 23.65-24.30
m m
24.30-27.20
m
27.20-29.10
m
Condition A fill was found of clayey silt and sand with humus. A large content of pottery remains of Aztec origin was encountered in these horizons. The average water content is about 4576. A layer of black volcanic ash with silt and little clay. Deposit of light grey plastic fissured silty clay with root-holes and high content of calcium carbonates, Caliche Barrilaco. The average water content is about lOOo/o. Pumice sand. Grey clayey silt with calcium carbonates. Pumice sand, and gravel. Greyish olive-green fissured clayey silt with little calcium carbonates. Average water content about 90%. Becerra sediments. Lacustrine volcanic clay, containing the mineral montmorillonite, diatoms, and ostracods. Tacubaya Clay I. Black volcanic ash. Brown and reddish brown lacustrine volcanic clay containing the mineral montmorillonite, diatoms, and ostracods. Tacubaya Clay I. Black volcanic ash. Grey clayey silty sand, with root-holes and calcium carbonates. Olive-green lacustrine volcanic clay, montmorillonite, diatoms, and ostracods with lenses of white volcanic glass at 19.75 and 20.80-m depth. Tacubaya Clay II. Brown pumice sand. Grey clayey silt and fine sand with root-holes and calcium carbonates “ caliche “. Brown and reddish brown volcanic clay. Tacubaya Clay III. Grey clayey silt and fine sand with root-holes and calcium carbonates “ caliche “. Olive-green lacustrine volcanic clay, contains montmorillonite, diatoms, and ostracods. Tacubaya Clay IV. Series of lacustrine deposits of volcanic montmorillonitic clay, pumice sand and ostracods sand. Ostracods and iiolites very abundant. Extremely pervious deposit in horizontal direction corresponding to Tacubaya Clay V.
118
LEONARDO
29.10-3350 m 33.5~38.20
m
38.20-41.55 m
41.55-41.95 m 41.95-45.25 m 45.25-47.70 m 47.70-64.50 m 64.50-65.25 65.25-65.40 65.466660 66.60-68.75 6875-70.00
m m m m m
ZEEVAERT
Olive-green lacustrine volcanic clay, containing the mineral montmorillonite, ostracods, and some diatoms. Tacubaya Clay V. Series of alluvio-lacustrine deposits of andesitic sand, clayey silty sand with little andesitic gravel and pumice, root-holes and calcium carbonates in the upper part of the deposit. Tarango Sand I. Olive-green lacustrine volcanic montmorillonitic clay, with diatoms, ostracods, sponge spicules, with a black sand lens at 41.20. Tarango Clay I. Fine sand layer of white clean volcanic glass, wind deposited on the lake. Olive-green lacustrine volcanic montmorillonitic clay with white clean volcanic glass lens at 43.50-m depth. Tarango Clay I. Same lacustrine clay as above, interbedded with numerous thin lenses of volcanic sand. Tarango Clay I. Series of deposits of sand, clayey silt, or silty sand of andesitic origin. Little gravel and pumice grains. Tarango Sand II. Brown lacustrine volcanic clay. Lenses of volcanic sand. Olive-green lacustrine volcanic clay. Fine sand of white, clean volcanic glass. Olive-green iacustrine volcanic clay.
Fig. 3 shows the water content profile from which may be seen distinctly the lacustrine bentonitic clay deposits. The first lacustrine volcanic clay deposit corresponding to Tacubaya Clay I-V, assumes a high water content that remains practically constant with depth and reaches a height of 350%. Near the sand lenses the water content in the clay drops on account of higher content of coarser grains in the sediments. The large scattering of the water content appears to be because of the transgression and regression of the sediments as the water level in the lake assumed different elevations. This fact may be recognized also by the variation in the Atterberg limits. The liquid limit was encountered as high as 400% and as low as 260% regardless of depth, and the plasticity index between 264% and 110%. The unconfined compressive strength shows a large variation from O-7-1.4 kg/sq. cm. This variation may be associated with the different salinity of the water in the lake during the process of sedimentation. The minimum value of the unconfined compressive strength varies from 0.7 kg/ sq. cm in the upper part of the deposit to 0.85 kg/sq. cm at the bottom. From the permeability point of view it is important to notice the sandy and silty layers containing calcium carbonates at depths of 15.85, 21.50, 23.65, and 28 m. These horizons define shallow waters in the lake. Particularly important is the series of silt and sand layers with high content of microscopic shells, between 27.20-29.0 m deep. All these materials have a permeability from ten to one hundred times larger than the volcanic clay deposit. From geological considerations these layers may be considered continuous since they appear in the same stratigraphic position in the subsoil in many other places in the heart of the city. Therefore, from the hydraulic point of view, for consolidation purposes, they may be considered as drainage surfaces within the clay mass. Compressibility curves for the volcanic high compressible clay deposits are illustrated in Fig. 4(a). The first hard deposit Tarango Sand I has a variable compaction, its water content varies The upper part of the deposit, because of cementation with clay and calcium from 2570%. carbonates, has in the in situ state a higher strength ; but the strength may be variable in the horizontal direction because of the erratic development of calcium carbonates and clay content. The cohesion may be as large as O-4 kg/sq. cm and the angle of internal friction as high as 36”.
FOUNDATION
DES
GN AND
STANDARD PCN!lTRATlOh BLOWS PER Fool 50 I[x) 150200
BEHAVIOUR
OF TOWER
.ATINO
AMERICANA
FIATLRCONTLNTIN: UNCONFIN COMPRESS If DRYWM-ITOfSOLID:lRENGTUINI
sotoo IS0 200 250 3003so401
iI510 1.5
119
EFFECTIVE PRLSSURL IN KG/CM' 05 1.0I"520 25 30 35
.
m
m -
GRAY SILTYCLAY WITU CALCIUM
* WATER CONTENT
CARBONATES.ROOT-HOLES AND
X UNCONFINED COMPRESSIVE STRENGTH
SAND(CALlCUE)
0 INTERGRANULAR EFFECTIVEPRESSURE
iAND ATTERBERG LIMITS
l
X VOLCANIC CLAY @ SANDY J SILTYCLAY
INCLUDING SURFACE LOAD
4= BREAK INTHECOMPRESSIBILITY CURVE
ss SPECIFIC GRAVITY Fig.
3.
Subsoil
profile
.
120
LEONARDO
ZEEVAERT
PRCSSURC ‘P’ INK/C’
PRCSSLIAC :P’ IN K/C’ Fig.
4.
Compressibility
curves
The second lacustrine volcanic clay deposit corresponding to Tarango Clay I, has an Compared with the almost constant water content of about 190% in its entire thickness. The silt and very fine sand upper volcanic clay deposit the Atterberg limits are smaller. The variations in content is larger and has less content of ostrocod shells and diatoms. The unconfined compressive strength assumes minimum liquid limit are from 260-108%. values of about 0.9 k&q. cm in the upper part of the deposit, Fig, 3. Compressibility curves of this volcanic clay are shown in Fig. 4(b). The second hard deposit, Tarango Sand II, consists of a series of alluvio-lacustrine strata of sand, silt, and clayey silts with gravel, and may be considered in a semi-compact state. The compressibility is low. The cohesion is zero for sand and silt stratifications and as large The angle of internal friction may reach values as 0.67 kg/sq. cm in the clayey sediments. up to 45”. The second lacustrine clay deposit, Tarango Clay II, encountered at 65-m depth has a water content of 150%, liquid limit of 153% and plasticity index of 105%. Compressive strengths are as low as 1.65 kg/sq. cm. HYDRAULIC
CONDITIONS
The investigation of the hydraulic conditions in the subsoil is extremely important in relation to the ground surface subsidence of the area in question and the value of the effective To perform this investigation piezometers were installed overburden pressures in the subsoil.
FOUNDATION
DESIGN
AND
BEHAVIOUR
OF TOWER
LATIN0
121
AMERICANA
at different depths. The horizons selected to install the porous point of the piezometers were The curve marked B the most pervious strata at 48, 34, 28, 21, 16, 12, 8, and 2-m depth. in Fig. 3 shows the effective overburden pressure computed with the piezometric pressures encountered, and the curve marked A shows the effective pressures with static hydraulic conditions (as if all piezometric water level elevations would reach the water-table found at 1.15 m from the ground surface). The curve marked B shows that the effective pressures increased by the drop in piezometeric pressures, because of downward water flow. The investigation demonstrates that there is a small drop in the piezometric water levels for piezometers installed at 28-m depth or less, but the strong change in the piezometric levels starts at 34-m depth. The semi-pervious layers at 28 m appear to provide sufficient water to maintain, at present, the hydrostatic pressure practically unchanged at this elevation. Therefore, an The seepage important downward hydraulic gradient is established only after 28-m depth. forces have increased the effective pressures in the fifth layer of the upper clay deposit Tacubaya, and in deposit Tarango Clay I, as shown in Fig. 3. From this investigation it was concluded that the source of ground surface subsidence was mainly the compression of Tacubaya Clay V of the upper volcanic clay deposit, and that of the second volcanic clay deposit, Tarango Clay I. Benchmarks ST48 and 9T34 (see Fig. 9) installed at the site at 48-m and 34-m depth, respectively, show the quantitative values of the compression of these two clay deposits and of the total ground surface subsidence with respect to benchmark ABN49 installed at 49-m depth in the Alameda Park, 280 m away from the site, Fig. 7. The location of the benchmarks and reference points used in this investigation are shown in Fig. 8. From observations in the Alameda Park, illustrated in Fig. 9, it will be seen that starting in 1950 the rate of drop in piezometric water pressures has diminished and also the velocity of ground surface subsidence to about half of its value during the period 1949-1950. This phenomenon may be due to the suppression of part of the deep water supply wells in the central part of Mexico City. FOUNDATION
DESIGN
The foundation was designed with piles, covering an area of 1,004 sq. m on the first hard deposit, Tarango Sand I, Fig. 3. This layer was selected to avoid excessively large negative friction on the piles, and the emerging effect of the building from the surface of the ground ; in contrast to a design using piles bearing on Tarango Sand II, which would cause the effects referred to above to be of an unacceptable magnitude. Furthermore, the piles were more economical with a length to reach the first hard stratum. A safe average load of 1.2 kg/sq. cm was assigned to the upper part of Tarango Sand I, taking into account the reduction of pressure because of excavation, the rigidity of the foundation structure, and the distributing effect of the supporting sand layer itself. The weight of the building is 2.10 kg/sq. cm : therefore, to obtain an increment of pressure in Tarango Clay I that could be taken safely, it was necessary to support with uplift water Thus it was decided to place the pressure the balance foundation pressure of 1-O kg/sq. cm. foundation slab at a depth of 13.0 m from the ground surface. The probable settlement caused by the increment of load in the second clay deposit, Tarango Clay I, may be estimated using the following settlement equations, taking into consideration the secondary consolidation :
.
St = s, + -52 Primary
consolidation
S, = L’ mqll. H . p” . t . F,(T,) S, = L’ m,z . H . ;[F(Tm)(t 4*
.
.
.
.
.
(1)
.
.
.
.
.
(2)
t, < t < tc .
.
(32
: -
ta) +
o
F1(T&,l
122
LEONARDO
ZEEVAERT
Secondary consolidation : t, < t < t, + t,
(4)
.
t, + te < t in which : n=m
c
2
WV) = 1 -
-
(2n + 1)P f
(2%
:
E
1)2.
(5)
T” .
.
.
.
.
.
.
.
(6)
n=o
?I=*
F,(T,)
= 1 - -&
u
t, =-
n=O
1)4 .(I
(212 :
T,, . Hz 4cv
-
E
4&l , T, = 112 . t
(2n+ l)W --i+u)
.
.
.
.,
(7)
.
(8)
$ = AP t = rate of loading, considered constant during loading period t, e The values of m,l, mt, c, and T,, are obtained from consolidation tests. The average values of these mechanical properties and the average increment of pressure are reported in Table 1, Fig. 5. The computed time-compression curve of Tarango Clay I for a loading period t, = 0 months is shown as curve A in Fig. 6. However, the rate of settlement is governed by the perimeter friction of the foundation against the upper subsoil deposits and therefore by the speed of settlement corresponding to the ground surface subsidence with respect to the point of the piles resting at 34-m depth on the sand layer. Therefore, the rate TIME IN MONTUS 1
2
34
P-id_
SAND
200
I
i i i
mw mt cv L
I I
1000
I\ 3
hP
//I////////[
C=/K C2/K +tfX fl-f)n#$ K/c”
CLAY
0039 0026 0.00166 3.64 0.47
g
j
68\0
YOLCLUIC 6LW
s 2
CLAY
0.03550.021 o.m2
3.03 0.40 Fig. 6. Compression of Tsrango Clay I, because of weight of building
Fig. 5.
Consolidation
properties,
Tarango
Clay I
FOUNDATION
DESIGN
AND
BEHAVIOUR
OF TOWER
LATIN0
123
AMERICANA
of loading of deposit Tarango Clay I appears to be much smaller than the rate of loading corresponding to the construction period of the building. Calculations made to adjust the observed settlement of the building with computed settlement demonstrate that the fitting of the observed curve with the computed curve marked B in Fig. 6, requires a loading period close to t, = 84 months. The building will not emerge from the ground surface until the rate of compression of the above-mentioned clay deposit is smaller than the rate of ground surface subsidence with respect to the 34-m deep sand layer. The plan and cross-section of the foundation are shown in Fig. 2. The number of piles used is 361 and, under normal conditions, they carry a load of 33 tons/pile. Several pile tests performed at the site showed the elastic limit working conditions of the piles to be 90 tons and the maximum load necessary to force the pile into the sand stratum, 120 tons. However, as the building emerges from the ground surface the compression of Tacubaya Clay V will create a negative friction on the piles because of the relative velocity resulting from the ground surface subsidence and the compression of Tarango Clay I. The total negative friction acting on a pile may be estimated by means of the equation : h . F(-)
=
s
.
.
.
.
.
.
(9)
0 Sda
in which S is the unit shearing strength of the remoulded clay along the shaft of the pile. The following approximate value may be assigned to the start of shear flow : s=~K,tan$,.*,
.
.
.
.
.
.
.
(10)
PALACIO
CINCO
DE MAYO
u
AVE.
I
PI8
GU4RDlOLA ILIILUWCU PIUS
PI7
J UARLZ
AVE. z
ll--llI?
Fig.
7.
General
location
of building
and
benchmarks
Pi6
I
MADERO
AVE
124
LEONARDO
ZEEVAERT
MADERO AT8 “i’
Q
Fig. 8.
K0 = ratio of horizontal
ds =
angle of internal
The vertical effective equilibrium equation :
@Ta
mw~tti
““’
I
Li >
Pn - = PV
AVF 0112
Reference points at the site
to vertical friction
effective
of remoulded
stress in the clay deposit. clay.
pressure within a group of piles may be computed
~+Jx.pv=~
.
.
.
by the following
.
.
.
(11)
in which : P,=j(z)=
ver t ica 1 e ff ec t ive pressure in the because of negative friction. N = a constant
= 2.1 d
-n. K, 'tan
clay
deposit
not
affected
by reduction
&
. in which : d = diameter
of pile shaft
n = number of piles per unit area Therefore, the value of the effective vertical pressure at depth z when negative acting on the pile may be computed by the following expression :
s
aP0 + c p, .cNZ= &--a.2 &
1
.
.
.
.
.
friction
is.
,
The value of 9, may be expressed by an approximate function of z, and knowing the boundary conditions, the value of C, may be determined. In the case of Tower Latin0 Americana, it was found that the centre piles may take an approximate load of 19 tons/pile because of negative friction, the piles on the sides about 22.5 tons, and the corner piles of the order of 27 tons. Therefore, when the building emerges from the ground surface the most heavily loaded piles will be those at the corners. However, the total load including the load induced by negative friction is well below the ultimate elasric load of 90 tons found from pile tests. The earthquake effect increases the load at the edges of the foundation to about 6 tons/pile. Therefore, the average coefficient of safety against pile-point penetration in the sand is of the order of 2, and against elastic behaviour, 1.50.
FOUNDATION
DESIGN
Fig.
9.
4ND
Surface
BEHAVIOUR
subsidence
OF
with
TOWER
respect
LATIN0
AMERICANA
125
to ABN49
EXCAVATION
The foundation design adopted (Fig. 2) called for a deep excavation into the volcanic clay. Current practice in Mexico City for excavations up to 6-m depth have shown that heave may be very important depending on stratigraphic conditions. A large heave may be observed when the excavation cuts into the lacustrine volcanic clay deposits. Therefore, in order to perform the excavation required for this building a special design was necessary to ensure the minimum possible heave and disturbance in the clay deposit. On the other hand, it was undesirable to produce a large water-table draw-down in the neighbourhood of the excavation, because of the very large settlements that would be induced and consequent damage to the street and neighbouring buildings. In order that the heave of the bottom of the excavation and settlement outside should be unimportant it was theoretically necessary to avoid a large change in the prevailing effective stresses in the clay mass during the excavation process. Following this philosophy a special hydraulic system was designed. An accurate knowledge of the stratigraphical subsoil conditions, as explained before, was imperative in the design of such a system. The area to be excavated was surrounded by a “ Wakefield “-type wood sheet-pile to a depth of 16 m. The wood sheet-pile upon saturation swelled to form a practically impervious membrane impeding the entrance of water in the excavation, and protecting from a strong draw-down of the water-table in the outside area surrounding the sheet pile. The pressure in the sand layers was maintained by injecting clean water under pressure in the subsoil by means of eight wells placed as shown in Fig. 10. The injection wells were perforated at depths of 12,16,21, and 28 m to feed water to the sand lenses located at these depths. The water-table was maintained in the upper pervious deposits with an absorption ditch provided with absorption wells to a depth of 9 m. The piezometric water levels and water-table around the excavation could be maintained with a reduction in water levels that was not detrimental to the public utilities and old structures surrounding the building under construction.
126
LEONARDO
ZEEVAERT
The heave because of excavation was avoided by producing a strong reduction in the piezometric water levels inside the wood sheet-pile to keep effective pressures essentially the same, or greater, as excavation proceeded to a depth of 8 m. This practice at the same time produced positive friction in the upper part of the piles previously driven from an excavation 2.5 m deep from the ground surface. The above-mentioned phenomenon was created using four deep-well water pumps installed to a depth of 35 m, located as shown in Fig. 10. The water pumps were operated to reduce the piezometric water levels inside the wood sheet-pile diaphragm driven 16 m deep. The water obtained from the wells in the interior of the sheet-piling was injected under pressure in the injection wells in the exterior of the sheet-piling. When the hydraulic system, as already described, was working under normal conditions excavation proceeded from 2.5 m to 8 m depth. Thereafter, trenches were excavated to construct the foundation beams. The reduction of piezometric water levels during the operation of the hydraulic system described are shown in Fig. 11, for piezometers installed inside and outside the wood sheet-pile respectively. The corresponding piezometric water pressures for normal and minimum conditions during the performance of the hydraulic system are plotted in Fig. 12. The total pressures for various steps in the excavation are given in Fig. 13, for one point at the centre of the excavation enclosed by the sheet-pile and another point 2 m away from the sheet-pile in the outside Fig. 10. Hydraulic system layout loaded area. From the total pressures, the piezometric water pressure readings have been subtracted obtaining the effective pressures for the various cases. The effective pressures with excavation to 2.5-m depth inside the sheet-pile are plotted in Fig. 13(a), curve marked A. Curve marked B shows the effective vertical pressures with excavation to 8-m depth and the maximum reduction of the inside group of piezometric levels as obtained while the interior hydraulic system was working. Curve marked A Outside the sheet-pile the effective pressures are shown in Fig. 13(b). shows the effective pressures with the 2.5-m deep excavation inside the sheet-pile, and curve marked B with the excavation inside sheet-pile to 8 m depth and injection wells working outside sheet-pile. In both cases of curves i3, Fig. 13, the absorption ditch and absorption wells were working and represent the minimum piezometric water pressure conditions observed inside and outside the sheet pile during excavation respectively.
FOUNDATION
DESIGN
AND
BEHAVIOUR
OF
TOWER
LATIN0
AMERICANA
127
The application of the hydraulic system turned out to be a success, since the settlement of the ground surface outside the sheet-pile caused by the deep-seated compression of the clay deposits, Fig. 13(b), did not affect the neighbouring buildings or public utilities. The upheaval of the bottom of the excavation did not take place since, during the excavation period, the upper part of the clay deposit was under an average increment of pressure of about O-4 kg/sq. cm, Fig. 13(a), which introduced a positive friction load on the piles. After the hydraulic system was suspended, Fig. 11, the pumping and injection wells were sealed with cement to re-establish hydraulic initial conditions. In order to provide means to correct any tilting of the building because of the non-homogenity in the compressibility of the volcanic clay deposits, injection wells were designed at the four corners of the foundation. At any one of these wells the pressure could be raised or lowered if necessary to produce an important difference in uplift water pressure at the corners of the building. Therefore, a counteracting tilting moment can be introduced that may help to force the building back to its vertical position. This hydraulic system has not been in use, however, since the building has not shown any sign of tilting. The deep foundation design undoubtedly has contributed very effectively in absorbing any difference in compressibility properties of the volcanic clay deposit, Tarango Clay I, consolidating under the load of the building. SETTLEMENT
OBSERVATIONS
Settlement and piezometric observations have been carefully carried on by the engineering staff of La Latin0 Americana since the beginning of construction. The most representative observations are reported in Figs 14 and 15 taken with reference to a fixed benchmark, ABN49, established in the Alameda Park at a depth of 49 m.
25M+aH+l3OW-_(
18P16 +
IP21
0
8PZ8 CPl6
38P51
Fig.
l
11.
49.43
Piezometric
water
levels
during
excavation
128
LEONARDO
ZEEVAERT
UYDROSTATK PRESSURE IN T/M~
Fig. 12.
Piezometric
water pressures
!E IN TON/ M* 0
during excavation
40
70 80 93 loo 110 120
CFFECT~LCAIJSE
INSIDESHUT-PILE Fig. 13.
Vertical effective pressures
OUTSIDE SI-IET-PILE before and during excavation
EX-
FOUNDATION
I949
I
IQ50
DESIGN
I
1451
I
AND
BEHAVIOUR
1952
1953
Fig.
14.
Settlement
OF TOWER
1954
LATIN0
1956
AMERICANA
1
1957
129
1958
observations
When excavation proceeded in 1949 the benchmarks at the site showed the following average settlement with respect to ABN49 : P8 9T34 ST48
.. .. ..
.. .. ..
.. .. ..
27.3 cm/year 15.6 cm/year 6.5 cm/year
Sanborn’s sidewalk at the site at the site
Therefore, the added compression of Tacubaya Clay V and Tarango Clay I at the site was 20.8 cm/year, the compression of Tarango Clay I, 9.1 cm/year and the compression of Tacubaya Clay V, 11.7 cm/year. During excavation and because of load relief in Tarango ,Clay I the compression of this deposit stopped from November 1949 to November 1950 until the load of the building was large enough to start again the compression of this layer (Figs 14 and 15), thus showing that at present Tarango Clay I is consolidating because of the load of the building. The rate of consolidation is governed by the rate of the ground surface subsidence. The positive friction acting against the foundation walls and piles holds the building against a faster settlement. This condition may continue until the rate of consolidation of Tarango Clay I is smaller than the rate of ground surface subsidence with respect to the 34-m sand layer when the positive friction will turn into negative friction. The settlement of benchmark 9T34 at the site compared with the columns of the building is shown in Fig. 14, demonstrating that there has been no penetration of the piles in the so-called hard layer Tarango Sand I at 34-m depth. The above-mentioned philosophy assumed
130
* E i 2o
L
5 IO .c :
2
LEONARDO
EXCAVATION
_---
TO:
-2.5t+BM+l3.0M-j
/‘=iiiii‘
ZEEVAERT
RATE
/y
t
%IDRAULIC
/
SYSTEM WORKING
I
6TB 3T2
@ x
IT2
o
8T48 YT34
v t
P8 Pfl
’ +
PI7
0
1950
1
1951
1
1952
Fig. 15.
1
1953
1
Settlement
19%
1
1955
1
1956
i
1957
i
observations
had a fairly uniform surface subsidence. A reference point P 11 on the Palace of Fine Arts with weight of 1.2 k&q. cm (Fig. 9) shows that this building is not settling any more with respect to the Alameda Park unloaded areas. However, the area west of San Juan de Letran comprising the Alameda Park and Palace of Fine Arts is settling with respect to the La Traza area east of San Juan de Letran. The clay deposit under La Traza area, heavily loaded since the 16th century, is less compressible than those outside. This fact may be observed in Figs 9 and 14 from surface reference points PS and ABN3. Building Guardiola across the street is a twelve-storey building constructed in 1940 on 1,156 wood piles with an average load of 7 tons/pile. The foundation and basement are placed in an 11-m-deep excavation : therefore, it may be considered that the movement of Guardiola Building with respect to ABN49 benchmark, as well as the movement of this building relative to the ground surface, may be considered representative of the ground surface subsidence phenomenon and of the normal compression of Tarango Clay I in this area. The settlement curve P-17 of Guardiola Building, Fig. 15, may be taken as the approximate origin for the additional compression of the Tarango Clay I deposit under the load of Tower Latin0 Americana, as shown in Fig. 16. CONCLUSIONS
The foundation of Tower Latin0 Americana is behaving as predicted by the investigation of the foundation design. The rate of settlement of the building has been uniform and at present is essentially the same as that of the ground surface subsidence observed by reference point P8. From Fig. 16 it can be seen that from December 1950 to February 1953 the clay
FOUNDATION
DESIGN
;j~D,,,,,
5YSCt-l
AND
BEHAVIOUR
OF TOWER
LATIN0
AMERICANA
131
WORKING
-COMPUl~O StlTLtMtNTOr L.A.TOWtRWIT!-l RESPfCT ABM49 BCNCUMARKINALA~ hPAR l OBSERVED SCTTLCMCNT or L.A.T~wCRWITH REsPrcT A8N49 BCNCUMARKINALAI RPAR + OBStRVCD SCTTLCMFNT OF GUARDIOLAWITURtSPtCT ABN49 0tNCUMARKIN ALIl IAPA' . OBSEQV~DSCTTLEMENT Or SANBORN'S SIDCWALK(P8) ----RrrERENcc LiNr ~~~COMPUTED XTTLCMENT.(DARALL~LTO GUARDIOLA +I IQ49 1950 1 1851 ] 1952 1 IQ53 Fig. 16.
1954 1 IQ55 1 1956 1 1957 1 IQ581 (9.591 IQ60
Predictions
iQ6i 1 I962
for future settlement
deposit supporting the building did not compress, although practically all the load of the building had been applied. The settlement is parallel to Guardiola Building on piles. After February 1953, the Tarango Clay I deposit started to compress on account of the load of the building that was gradually transferred to the point of the piles as the foundation was permitted to settle because of the compression of the upper clay deposits. An attempt to estimate the future net settlement of the building is shown in Fig. 16, using computed settlement curves from Fig. 6, for loading periods tc = 0 and tc = 84 months. The origin of compression of the lower clay deposit Tarango Clay I was taken approximately in February 1953. The observed settlement lies between the computed curves. Therefore, the lower clay deposit supporting the building still has to compress theoretically, in 6 years, an estimated value of 12 cm. On the other hand, extrapolation appears to indicate that the building will start to emerge slightly from the ground surface within the next 2 years. However, although in the future the building may emerge from the ground surface because of the ground surface subsidence produced by the compression of Tacubaya Clay V, the ground floor resting on movable supports may be lowered as required to follow the sidewalks at the building. Provisions in Mexico City are under way to reduce the rate of fall of the piezometric water
132
LEONARDO
ZEEVAERT
levels. One item has been to stop pumping from water wells in the heart of the city. The effect may be already noticed by observations made at the Central Park and also at the building site. Another item has been to drill absorption wells in open areas to inject clean water. However, this measure is still in observation. Apparently in some places wells have not been properly sealed in the upper part of the clay deposit ; thus they serve also to drain the upper part of this deposit above the 28-m depth. This practice, if continued, may create a strong reduction of piezometric water levels in the upper part of the clay deposits and consequently a large compression of them. Already this fact has been observed by the Author in several places adjoining injection wells. At the site of Tower Latin0 Americana the drop is noticeable already in the 28-m-deep piezometer. The maximum limiting condition would be when piezometric water levels in piezometer 9P34 have dropped to a depth of 33 m, and at a rate of 0.54 m/year shown by this piezometer. This may take place within the next 30 years. Observations plotted in Figs 14 and 15 show that with respect to the Alameda Park benchmark ABN49 the rate of settlement taking place in the last 2 years has had the following value : 12.9 cm/year Ground surface . . P8 Tarango Sand I . . Guardiola 9.7 cm/year 2.4 cm/year Tarango Sand II . . 8T48 Therefore, the compression of Tacubaya Clay V is at present only 3.2 cm/year and Tarango Clay I is compressing at the rate of 7.3 cm/year. Assuming that conditions will remain as they are observed to-day, then in the next 30 years the building may emerge from the ground surface roughly 100 cm. On the other hand, it will be seen from settlement observations since 1949, that there is the tendency to reduce the rate of drop in the piezometric water levels, and correspondingly the rate of compression of the high compressibly volcanic clay deposits (Figs 14 and 15). If this favourable situation continues, then the differential settlement between building and ground surface will be smaller than the above-mentioned estimated value. ACKNOWLEDGMENTS
The pile driving, excavation, and construction of the foundation was under the direction of Mr Adolf0 Zeevaert, Civil Engineer, Chief Engineer of La Latin0 Americana. The consulting engineering during construction, design of the foundation structure, and the soil mechanics investigation were performed by the Author. The Author wishes to extend his appreciation to his co-workers: Mr H. Vogel, Civil Engineer for laboratory work; to Mr Heriberto Izquierdo, Civil Engineer, who was directly in charge of the calculation of the foundation structure; to Mr Jaime de la Peza for preparation of figures and computations included in this Paper; and to the staff of the engineering department of La Latin0 Americana for careful settlement and piezometric water level observations. Without this information the behaviour of the foundation of the building could not have been estimated and controlled during construction. The Author wishes also to extend his appreciation to the Life Insurance Company, ,La Latin0 Americana, for all the efforts this company made toward the solution of this interesting foundation problem, even though the early investigations appeared to be only of academic value. BUISMAN, A. S. K., 1941. FOREMAN, FRED.,
1955.
BIBLIOGRAPHY ” Grondmechanica ” (“ Ground mechanics “). “ Study
of two cores from lake sediments
Geological Society of America, 66 : 471-530.
MARSAL, R. J., F. HIRIART, and R. SANDOVAL,
settlement of the City Mexico, 5 : 14-49.
of Mexico “).
IV
Waltman, DeZft.
of Mexico
City Basin.”
Bulletin of
1953. “ Hundimiento de la ciudad de Mexico ” (“The Centenario de la Universidad National Autdnoma de
FOUNDATION
DESIGN
AND
BEHAVIOUR
OF TOWER
LATIN0
AMERICAN.4
133
SKEMPTON, A. W., 1955. “ Foundations for high buildings.” Proc. Instn Civ. Engrs, 4 (3) : 246. SKEMPTON, A. W., and ‘D. H. MACDONALD, 1956. “ The allowable settlement of buildings.” Proc. Instn Civ. Engrs, 5 (3) : 727. “ Forty-storey building in Mexico.” STACE, F. N., 1955. New Zealand Engineering, 10 : 215. TERZAGHI, K., 1943. “ Theoretical soil mechanics.” Wiley, New York. TERZAGHI, K., and 0. K. FRBHLICH, 1936. “ Theorie de1 setzung von tonschichten ” (“ Theory of settlement of clay strata “). De&eke, Vienna. TERZAGHI, K., and R. B. PECK, 1948. “ Soil mechanics in engineering practice.” Wiley, New York. ZEEVAERT, LEONARDO, 1944. “ Conceptos y experimentos fundamentales que se aplican al diseiio de cimentaciones en arcillas saturadas ” (“ Application of soil mechanics to the study of foundations in saturated clay “). &vista Mexicana de Ingenieria y Arquitectura, 20 : 366-375. ZEEVAERT, LEONARDO, 1949. ” An investigation of the engineering characteristics of the volcanic lacustrine clay deposit beneath Mexico City.” University of Illinois. ZEEVAERT, LEONARDO, 1953(a). “ Compresibilidad de la arcilla volcanica de la ciudad de Mexico” (“ Compressibility of the volcanic clay of Mexico City”). IV Centenavio de la Unniversidad National Autonoma de Mexico, 5 : 50-57. ZEEVAERT, LEONARDO, 1955(b). “ Estratigrafia y problemas de la ingenieria de 10s depositos de arcilla lacustre de la ciudad de M&co ” (” Stratigraphy and engineering problems of the lacustrine clay deposits of the City of Mexico “). IV Centenario de la Universidad National Auto?zdma de Mexico, 5 : 58-70. ZEEVAERT, LEONARDO, 1953(c). “ Outline on the stratigraphical and mechanical characteristics of the unconsolidated sedimentary deposits in the basin of the Valley of Mexico.” IV Congres International du Quaternaire, Rome-Pise. ZEEVAERT, LEONARDO, 1953(d). “ Pore pressure measurements to investigate the main source of surface subsidence in Mexico City.” Proc. 3rd Int. Conf. Soil Mech., 2 : 299-304. ZEEVAERT, LEONARDO, 1953(e). “ Ecuaci6n completa de consolidaci6n para depositos de arcilla que exhiben fuerte compresidn secundaria ” (” Complete consolidation equation for clay deposits showing strong secondary compression “). Revista Ingenieria, April-August, 1953. School of Engineering, U.N.A.M. ZEEVAERT, LEONARDO, 1953(f). (Contribution to discussion.) Proc. 3vd I&. Conf. Soil Mech., 3 : 50.-57, and 129-131. ZEEVAERT, LEONARDO, 1956. “ Heavy and tall buildings in Mexico City.” Proc. Amer. Sot. Civ. Engrs, 82 : 917-923.
