Professional Documents
Culture Documents
Consoli Et Al. 2019 - Ensaios de Placa Jamile PDF
Consoli Et Al. 2019 - Ensaios de Placa Jamile PDF
Abstract: The usage of soil-cement reinforced layers to bear shallow foundations is a viable option in low-bearing-capacity soils. Existing
methodologies that can determine the bearing capacity in such cases tend to consider the cemented layer’s bounds to be infinite, depending solely
on the ratio between the footing width and the layer’s thickness. The present study intends to assess the influence of the reinforcement width on
the load-settlement behavior of a circular steel footing resting on square-edge soil-cement layers bearing on a weak highly porous residual soil.
Static load tests were carried out on footings (diameter of 300 mm) resting on sand-cement reinforced layers with distinct areas (edges of 450,
600, and 900 mm) and constant thickness of 300 mm. The results have shown two distinct failure modes that rely on the cemented layer’s width.
A punching mechanism was observed for the two smaller reinforcement’s layers and they did not fail. The 900-mm square cemented layer, on the
other hand, had a failure that was initiated by the formation of tensile cracks and fissures in the center of its bottommost segment. This study
highlights the importance of considering the soil-cement layer’s width in the bearing capacity estimation of footings resting on treated layers
above weak cohesive-frictional soils. DOI: 10.1061/(ASCE)GT.1943-5606.0002144. © 2019 American Society of Civil Engineers.
Author keywords: Plate load tests; Artificially cemented bounded layer; Cohesive-frictional soil; Lightly bonded soil; Ground
improvement; Failure mechanisms.
Soil Characteristics
Cohesive-Frictional Stratum
The residual soil existing at the experimental field is the product of
in situ weathering processes and exhibits characteristics related to
these processes, such as interparticle bonding and high porosity. As
such, the studied soil possesses a cohesive-frictional behavior and
relatively high permeability (k ¼ 1.1 × 10−5 m=s) when compared
to transported soils with similar grain size distribution. Field and
laboratory tests were employed in order to characterize the soil. The
first encompassed two in situ cone penetration tests (CPTs) per-
formed 10 m away from each other to a depth up to 15 m. Results
are shown in Fig. 1 and reveal the horizontal homogeneity of the
local soil. The soil profile possesses a thin soil crust (depth <0.5 m)
that exhibits a high tip strength (qt ) reaching 4 MPa, followed by a
reduction to 0.8 MPa of the qt in the next 3.0 m. A tip strength gain
is then observed, and it remains at around 1.5–2.0 MPa until the test
limit. The sleeve friction (f s ) followed the same trend as recorded
for the tip. The laboratory tests were carried out on a sample re-
covered from a depth of about 1.0 m, where the average bulk unit
weight was 16.1 kN=m3 , the void ratio (e) was 1.21, and the mois-
ture content was around 33%. The soil presented a liquid limit (LL)
of 42% and a plastic index (PI) of 11%. The grain size distribution
revealed 2.0% medium sand (0.425 mm ≤ diameter ≤ 2.0 mm),
20.0% fine sand (0.075 mm ≤ diameter ≤ 0.425 mm), 22.0% silt
(0:002 mm ≤ diameter ≤ 0.075 mm), and 56.0% clay (diameter ≤
0.002 mm). The soil is thus classified as a lean clay with sand (CL)
according to the Unified Soil Classification System (USCS)
(ASTM 2017). Its unit weight of grains is 26.7 kN=m3 . In order to
assess its mechanical behavior, isotropically consolidated drained
triaxial tests were performed in undisturbed specimens under con-
fining pressures of 20, 35, and 50 kPa. The results are presented in
Fig. 2 and yielded an effective peak friction angle (ϕ 0 ) of 30.5° and Fig. 2. Residual cohesive-frictional soil isotropically consolidated
an effective cohesion intercept (c 0 ) of 13.0 kPa. The deviatoric drained (CID) triaxial tests at confining stresses of 20, 35, and
stress–axial strain curves exhibit soil ductile behavior for effective 50 kPa: (a) deviator stress (q); and (b) volumetric strain (εv ) versus
axial strain (εa ).
confining stresses of 20–35 kPa and strain hardening behavior for
Fig. 5. (a) Load-settlement curves; and (b) pressure-settlement curves of 300-mm-diameter circular plate bearing on natural cohesive-frictional soil,
on 450-, 600-, and 900-mm-edge cemented layer into weak cohesive-frictional soil.
behavior [Fig. 4(b)]. Results led to a peak friction angle of 37.7° the previous load stage and the logarithm of time multiplied by a
and a cohesion intercept of about 105 kPa. time factor. When this time factor is plotted against the applied load,
two linear portions are observed and failure load is determined by
the point where one portion changes to another.
Field Testing Program
The field testing program was carried out at the experimental Test Results and Analysis
site described previously. Plate load tests were conducted using Figs. 5(a and b) show plots of applied load-settlement and applied
rigid circular steel plates 300 mm in diameter and 50.8-mm thick. pressure-settlement curves of 300-mm-diameter steel plate bearing
The setup used for conducting the plate load tests was in accor- on the residual cohesive-frictional soil without reinforcement
dance with ASTM D1194 (ASTM 1994). The load was applied and on 450-, 600-, and 900-mm square-edge artificially cemented
through a system composed of a hydraulic jack, a reaction beam, layers (300-mm thick) built above the same residual soil. It is evi-
a loading platform, and a calibrated load cell. Four dial gauges with
dent from Fig. 5 that higher load and pressures at a given settlement
divisions of 0.01-mm and 50-mm travel were used for vertical dis-
were observed in larger treated areas. Punching failure through the
placement (δ) measurement. Measurements were made directly on
square edges of the cemented layers occurred for the steel plate
the plate, as well as on the top of the treated layer adjacent to the
resting on top of the 450-mm and 600-mm layers. As mentioned
plate. Settlements on the top of the plate were taken at three points,
previously, the collapse occurred through the edges of the artifi-
120° from each other. The gauges were fixed to a reference beam
cially cemented strata, as though these volumes were part of the
and supported by external rods. The load was applied in equal in-
foundation structure. This mechanism of failure was corroborated
crements of not more than one-tenth of the estimated ultimate bear-
ing capacity. For each load increment, measurement of settlements by the field analysis of the cemented layers after the test was
was made at the following fixed times: 0 s, 15 s, 30 s, 1 min, 2 min, completed because no cracks or fractures were observed on the
4 min, 8 min, 15 min, 30 min, and 60 min (ABNT 1991). The fail- retrieved layers (Fig. 6). The sand-cement block, when loaded
ure load (Qu ) and the limit pressure were established at a relative through the steel plate, suffers soil punching shear failure (Vésic
settlement (δ=D) of 3%. This failure (limit) criterion was proposed 1975). For punching failure, as the load increases, the compression
by Consoli et al. (2009) based on a previous study by Berardi and of the soil immediately below the footing occurs, and the continued
Lancellotta (1991) that analyzed the behavior of more than 200 penetration of the footing is made possible by vertical shear around
shallow foundations, having found that maximum displacements the footing perimeter. There is practically no movement of the soil
at working pressures were generally of the order of 1%. Thus, on the sides of the footing. A continuous increase in the vertical
the δ=D criterion (3%) used in the present research to define the load is observed to maintain the movement in the vertical direction.
maximum (or limit) pressure was three times the value of δ=D ob- On the other hand, failure started in the cemented layer for the cir-
served by Berardi and Lancellotta (1991) at working pressure. Ad- cular plate resting on the 900-mm square-edge reinforcement. This
ditionally, failure was evaluated according to two other approaches. failure mode was attested to by the cracks that were found on the
The tangent intersection method, proposed by Mansur and parallelepipedic volume that was vertically cut after the load test
Kaufman (1956), was applied to the results to define the failure was completed (Fig. 7). Consoli et al. (2009) noticed a similar trend
loads. The intersection of the tangents of the initial and final portions for infinite layers with unitary H=D ratios (where H ¼ 300 mm
of the load-displacement curve defines the failure point. The method was the thickness of the treated layer and D ¼ 300 mm was the
proposed by the Laboratoire Central des Ponts et Chaussées diameter of the loading circular plate). The authors verified (in
(Cudmani 1994), which considers a time factor, was used as an the field and also through finite-element simulations) that tensile
alternative. The method defines the settlement at a given time after cracks initiated at the bottom of the cemented soil layer, below
the application of a load increment as the sum of the settlement of the circular plate.
its bottom. Hence, the bearing capacity of the circular plate resting
on this layer could be evaluated as if this layer was infinite.
Figs. 8(a and b) present the failure load (Qu ) and the equivalent
limit pressure (σe ) as a function of the relative settlement for the
three tested cemented layers, as well as for the 300-mm-diameter
circular plate resting directly on the natural cohesive-frictional soil.
The 900-mm cemented layer presented a peak valued at 118 kN,
observed in Fig. 8(a), which was considered to be the field failure
Fig. 6. Parallelepipedic shape of the 450 × 450 × 300 mm improved load. However, no peak was observed for the smaller treated layer
sand-cement volume (removed from the field). widths (450 and 600 mm). As such, the methods presented sub-
sequently were used to determine their failure load. For the sake
Fig. 7. (a) General view; and (b) lateral view of failure mechanism of the 900 × 900 × 300 mm improved sand-cement volume (above weak cohesive-
frictional soil) that failed below the vertically loaded steel plate.
Fig. 8. (a) Applied load (Q); and (b) equivalent pressure (σe ) versus relative displacement (δ=D) curves of 300-mm-diameter circular plate resting on
450-, 600-, and 900-mm-wide square cemented layer above a weak cohesive-frictional soil.
1994) yielded failure loads of 49, 87, and 104 kN for the 450-, ies carried out in the field). The Hansen (1961) method [Eq. (1)]
600-, and 900-mm cemented layers, respectively. The use of these was applied as if the footing was a single element formed by the
distinct methods led to similar failure load values, ranging within circular steel plate with the cemented layer resting on the weak
7% of the average. The different definitions of failure marginally cohesive-frictional residual soil. The strength parameters of the
affected the analysis. The continuity of this study was then devel- weak cohesive-frictional residual soil were reduced to a lower limit
oped considering the failure values determined at the peak for value of two-thirds in order to agree with Terzaghi’s (1943) recom-
widest treated layer and at a relative displacement (δ=D) of 3% mendations for punching failure mechanisms (Consoli et al. 1998)
(Consoli et al. 2009) for the others.
Qu B
The equivalent pressure (σe ) was defined as the applied load ¼ c 0 N c Sc þ γDN q Sq þ r γN γ Sγ ð1Þ
divided by the cemented layer bearing contact area, and similar A 2
behaviors can be noticed between the 450- and 600-mm cemented where A = area of the cemented layer; Br = reinforcement width;
layers. According to Consoli et al. (1998), the curves of equivalent γ = unit weight of soil; N c , N q , and N γ = bearing capacity factors;
pressure (σe ) versus relative displacement (δ=D) of the steel plate and Sc , Sq , and Sγ = shape factors for square foundations, given by
resting on the natural soil and the 450- and 600-mm cemented
layers should stand one on top of the other since the soil profile N c ¼ ðN q − 1Þ cot ϕ 0 ð2Þ
shows homogeneity for the first few meters of depth (Fig. 1).
The curves do not overlap because of the extra resistance in the N γ ¼ 2ðN q þ 1Þ tan ϕ 0 ð3Þ
interface between the cemented soil (which is buried 300 mm)
and the natural soil. However, it was observed that, at large dis- 0 ϕ0
N q ¼ eπ tan ϕ tan2 450 þ ð4Þ
placements, the interface between the cemented soil and the natural 2
Nq
Sc ¼ 1 þ ð5Þ
Nc
Sq ¼ 1.0 ð6Þ
Sγ ¼ 0.6 ð7Þ
in the strength parameters for punching the failure mechanism; and Acknowledgments
K was determined in accordance with Eq. (9)
The authors wish to explicit their appreciation to FAPERGS/CNPq
1 − ðsin ϕ10 Þ2
K¼ ð9Þ 12/2014—PRONEX (Project No. 16/2551-0000469-2), MCT-
1 þ ðsin ϕ10 Þ2 CNPq (Editais INCT-REAGEO, Universal & Produtividade em
Pesquisa), and MEC-CAPES (PROEX) for the support to the
Eq. (8) was presented by Vésic (1975) to address the situations
research group.
where the involved soils present both cohesion and friction. Exper-
imental works (e.g., Tcheng 1957) showed the failure mode under
such conditions was governed by punching mechanisms associated
Notation
with vertical slip lines through the foundation perimeter. Because
the materials involved in the present study present both cohesion The following symbols are used in this paper:
and friction angle and the observed failures exhibited punching A = area of the footing;
mechanisms, the use of Eq. (8) was possible.
Br = reinforcement width;
From Fig. 10, it is possible to detect that the bearing capacity
c 0 = effective cohesion intercept;
prediction considering the soil-cement layer as a part of the shallow
footing (Hansen 1961) is the correct result for edges of the square D = diameter of the loading circular plate;
reinforcement up to about 750 mm, and the infinite cement-treated D50 = mean effective diameter;
layer with H=D ¼ 1.0 (Vésic 1975) solution is the correct result for f s = CPT sleeve friction;
edges of the square reinforcement of 750 mm and greater. As such, H = thickness of the treated layer;
it can be ascertained that for the geometries, materials, and cement K = coefficient of permeability;
content evaluated in this research, the Hansen (1961) and Vésic Q = applied load;
(1975) solutions plotted on a single graph, as presented in Fig. 10, Qu = failure load;
is a useful tool to predict the failure behavior of reinforced layers q = deviatoric stress;
resting above the weak cohesive-frictional soil. qt = CPT tip strength;
γ = natural unit weight;
δ = vertical displacement;
Conclusions δ=D = relative displacement;
This study evaluated field load tests of a circular plate on square εa = axial strain;
bounded sand-cement reinforced layers with three different widths, εv = volumetric strain;
resting above a weak cohesive-frictional soil. The key outcomes are σe = equivalent limit pressure; and
as follows: ϕ 0 = effective friction angle.
• Two distinct failure modes were observed depending on the
soil-cement reinforcement edge width. The sand-cement layer
is kept intact and vertically punches the cohesive-frictional soil References
or the sand-cement layer shows some cracks below the center of
the plate extending upward from its bottom. ABNT (Brazilian National Standards Organization). 1991. Foundations—
• For the smaller reinforcement widths studied (450 and 600 mm), Static loading tests. NBR 12131. Rio de Janeiro, Brazil: ABNT.
ASTM. 1994. Standard test method for bearing capacity of soil for static
the ultimate bearing capacity could be assessed as if the footing
load and spread footings. ASTM D1194. West Conshohocken, PA:
was an association between the circular steel plate and the re- ASTM.
inforcement layer because no cracks or fissures were observed. ASTM. 2017. Standard practice for classification of soils for engineering
The reinforcement of 900-mm width failed due to excessive ten- purposes (Unified Soil Classification System). ASTM D2487. West
sile stresses that developed in its bottom. Hence, the bearing Conshohocken, PA: ASTM.
capacity of the circular plate resting on this layer could be Berardi, R., and R. Lancellotta. 1991. “Stiffness of granular soil from field
evaluated as if this layer was infinite or continuous. Higher loads performance.” Géotechnique 41 (1): 149–157. https://doi.org/10.1680
and pressures at a given settlement were observed in larger /geot.1991.41.1.149.
treated areas; Burd, H. J., and S. Frydman. 1997. “Bearing capacity of plane-strain foot-
• The plot of the Hansen (1961) and Vésic (1975) solutions in a ings on layered soils.” Can. Geotech. J. 34 (2): 241–253. https://doi.org
/10.1139/t96-106.
single graph was proven to be a useful way to predict the failure
Consoli, N. C., R. C. Cruz, M. F. Floss, and L. Festugato. 2010. “Param-
behavior and bearing capacity of the reinforced layers resting eters controlling tensile and compressive strength of artificially ce-
above a weak cohesive-frictional soil: (1) for soil-cement layer mented sand.” J. Geotech. Geoenviron. Eng. 136 (5): 759–763.
widths below the threshold solution by Vésic (1975), which https://doi.org/10.1061/(ASCE)GT.1943-5606.0000278.
considers an infinite treated layer for a specific H=D, the correct Consoli, N. C., F. Dalla Rosa, and A. Fonini. 2009. “Plate load tests on
result is given by Hansen (1961); and (2) for soil-cement layer cemented soil layers overlaying weaker soil.” J. Geotech. Geoenviron.