IOP Conference Series: Earth and Environmental Science

PAPER • OPEN ACCESS

Research on the differential settlements of mat foundations

To cite this article: C-M Ma and Y-Y Chen 2019 IOP Conf. Ser.: Earth Environ. Sci. 351 012018

View the article online for updates and enhancements.

This content was downloaded from IP address 194.110.89.206 on 29/10/2019 at 01:57

ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

Research on the differential settlements of mat foundations

C -M Ma1,2 and Y -Y Chen1

1
Department of Civil and Ecological Engineering, I-Shou University, Kaohsiung, 807, Taiwan
2
Email: [email protected]

Abstract. Mat foundation is a very commonly used foundation type. There are many
advantages in adopting mat foundations including the effects of loading compensation, full
utilization of underground space, lowering foundation settlements and increasing the factor of
safety of bearing capacity for foundations etc. In addition, because of the high stiffness of the
mat foundation and foundation girders the differential settlements could be effectively
decreased, which would in turn increase the safety of structures. The differential settlements of
the foundation will produce extra stresses in the structure. At the worst it will make the
structure members failure. There are many factors which would affect the differential
settlements of the foundation, which include the uneven distribution of structure column
loading, the subgrade reaction of foundation soils (or so called soil spring coefficient), the
combined stiffness of the foundation plate and foundation girders, the shape and size of the mat
plate etc. The main objective of this research is to use the finite element method to simulate the
interactions between the mat foundation and the underneath soils. It will focus on the
systematic research on all possible factors affecting the differential settlements of mat
foundations. The influence of each factor on the differential settlements will be studied and
identified. The influence charts for the differential settlements for each individual factor will be
proposed. Finally a reliable method using these influence charts to predict the differential
settlements of mat foundations will be developed, which will be verified by practical examples.

1. Introduction
Mat foundation is a very commonly used foundation type. Mat foundations would carry both the
loading from the upper structures and also the pressures from the soils underneath. Therefore this is a
typical soil-structure interaction problem. Researches about the mat foundations are still in progress
[1-3]. In the design of mat foundations the foundation plate would carry the moments and shears from
soil reactions as well as the extra moments and shears resulting from the possible differential
settlements of the foundation. Too much differential settlements will at the worst make the structures
failure. Therefore allowable differential settlement for the foundations is specified in many building
codes. The study of differential settlements of mat foundations is therefore very important. There are
several factors that will affect the differential settlements of mat foundations. Those include the
distribution of column loading; subgrade reaction coefficient of soils; the thickness of foundation plate;
stiffness of foundation girders and the shape and width of the foundation plates etc. [1]. Usually the
foundations resting on the clay soils suffer more obvious differential settlements. Therefore this
research will focus on the study of differential settlements of mat foundations on the clay soils.
There are several methods simulating the differential settlements of mat foundations [4]. In the first
simplified method as shown in figure 1, the soil springs are only imposed directly under the

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

foundation columns, and then the soil reaction pressures are acting on the foundation girders together
with the column loads to calculate the differential settlements. In the second simplified method as
shown in figure 2, the soil springs are uniformly imposed directly under the foundation girders. In this
case only the column loads are necessary to be imposed to calculate the differential settlements.

Figure 1. Simulation method 1. Figure 2. Simulation method 2.

Neither of the two above-mentioned methods takes into consideration the influence of the mat
plates on the differential settlements. In order to correctly predict the differential settlements, a finite
element method is adopted to simulate the soil-structure interactions as shown in figure 3 in this
research. A commercial finite element program “SAFE” is used to simulate the interaction behavior. In
this simulation, plate element is used to simulate the mat plate while the beam element is used to
simulate the foundation girders. On the other hand, the soil springs are uniformly imposed directly
under the foundation mat plate to correctly simulate the soil reaction pressures. The soil springs must
be assumed not to take any tensile forces to match the special characteristic of foundation soils [2,3].
However only flat mat foundation as shown in figure 4 is considered in this preliminary research, the
influence of the foundation girders on the differential settlements could be therefore ignored.

Figure 3. Simulation method 3.

2
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

Figure 4. Mat foundation.

2. Theoretical background

2.1. Vertical coefficient of subgrade reaction

As far as the simulation of the soil behavior, the concept of vertical coefficient of subgrade reaction is
adopted to simulate the foundation soils as soil springs. The factors that influence the coefficient of
subgrade reaction include soil type; foundation pressure; foundation size; embedment depth of
foundation and thickness of compressible soil layer [2,3,5]. In this research, the consolidation theory is
used to calculate the coefficient of subgrade reaction. According to the consolidation theory, the
consolidation settlement  Hc could be calculated as:
Cc  ,   ,
H c  H 0 log( 0 , ) (1)
1  e0 0
where e0; Cc and H0 are respectively the initial void ratio; compression index and thickness of
compressible soil layer, whereas  0’ and   ’ are the initial effective stress and effective stress
increment of the soils due to foundation loads.
The coefficient of subgrade reaction kv is defined as:
q
kv  (2)
H c

where q is the foundation pressure. As far as the foundation pressure; foundation size; embedment
depth and thickness of compressible soil layer are known, then the effective stress increment due to
foundation loads could be calculated. Therefore the consolidation settlements and then vertical
coefficient of subgrade reaction could be obtained. From the consolidation theory it could be seen that
under the constant foundation pressure the consolidation settlement increases as the size of foundation
or the thickness of compressible soil layer increases, which would therefore make the coefficient of
subgrade reaction decrease. On the other hand as the embedment depth of foundation increases, a
smaller consolidation settlement and higher coefficient of subgrade reaction would be obtained
because the net soil pressure increment decreases due to compensation effects of soil excavation. The
above-mentioned conclusions regarding the coefficient of subgrade reaction are in accordance with the
results of relative researches [2,3,5]. From the consolidation theory the relationship between the
foundation pressure and consolidation settlement is nonlinear. Therefore it could be further concluded
that the coefficient of subgrade reaction is not constant and depends on the magnitude of foundation
pressures. Based on the conclusions discussed above, it is reasonable to use the consolidation theory to
predict the coefficient of subgrade reaction of foundation soils. Later a practical numerical example
will be chosen to show how to calculate the coefficient of subgrade reaction based on the
consolidation theory.
2.2. The interaction behavior between the mat foundation and the soils underneath
The governing equation for the interaction behavior between the mat foundation and the soils

3
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

underneath could be written as: [6]

 4w 2 4w  4w q p
 2 2  4   (3)
x 4
x y y D D(xy )
where w is the deflection (settlement) of the mat foundation and q is the soil reaction pressure under
the foundation. Therefore,
q  kv w (4)
p is the column load and D is expressed as:
Ec t 3
D (5)
12(1   c2 )
where Ec;  c and t are respectively the Young’s modulus and Poisson’s ratio of concrete and
thickness of the mat plates.
In this research, a commercial finite element program “SAFE” is used to simulate the interaction
behavior between the mat foundation and the soils underneath [7]. In the simulation, the thick plate
element is adopted to simulate the mat plate, because mat plate is usually pretty thick. The thick plate
element has planar flexural plate stiffness in two directions within the plate as well as the translational
stiffness in the direction normal to the plate. In addition it could take the effects of shear deformation
into consideration [8]. Therefore it is appropriate to use thick plate element to simulate the mat plate.
In the “SAFE” program, the beam element can be chosen to model the foundation girders, while the
linear spring can be taken to model the soil under the mat foundation. Besides the soil springs are not
allowed to take any tensile forces. Furthermore the point load can be chosen to model the column
loads and the uniformly distributed load can be used to model the loads imposed on mat plates.
Therefore the finite element approach using the “SAFE” program could model the interaction behavior
between the mat plate and soils appropriately.

3. Numerical Simulation Of The Mat Foundation

A systematic approach will be taken to understand the factors that influence the differential settlement
of the mat foundation in the following. As mentioned above, there are several factors that will affect
the differential settlements of mat foundations. Those include the distribution of column loading;
subgrade reaction coefficient of soils; the thickness of foundation plate; stiffness of foundation girders
and the shape and width of the foundation plates etc. In this research, the distribution of column
loading is taken as column spacing of 8 meters both in vertical and horizontal directions. Each column
would take a load, which could be calculated by the floor area shared by the column multiplied by a
uniformly distributed load of 20 ton/m2 acting on the floors.
As far as the thickness of mat plate is concerned, 50; 70; 90; 110 and 130cm would be used in the
simulation. The vertical coefficients of subgrade reaction would be considered as 30; 50; 100 and 200
ton/m3. A square mat plate of 16m x 16m (L=B) is considered first in this simulation. The relationship
between the ratio of differential settlement to the average settlement and the thickness of the mat
plates under different coefficients of subgrade reaction is shown in figure 5. In this figure the average
settlement of mat foundation could be calculated by the uniformly distributed load of 20 ton/m2 over
the coefficient of subgrade reaction. It could be seen in figure 5 that the differential settlement
decreases as the thickness of mat plate increases. Furthermore the ratio of differential settlement to the
average settlement of mat plate increases as the coefficient of subgrade reaction increases. This is
ascribed to decrease of average settlement of mat plate as the coefficient of subgrade reaction
increases. Actually the differential settlement of mat plate does decrease as coefficient of subgrade
reaction increases.

4
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

0.60 0.60
t 0.50
0.50
n
e
m
el 0.40
Diff./Avg.Settlement

0.40 tt Kv=200
e
Kv=200 S 0.30
.g Kv=100
0.30
Kv=100 v 0.20 Kv=50
A
./f
0.20 Kv=50 if 0.10 Kv=30
D
Kv=30 0.00
0.10 0.5 0.7 0.9 1.1 1.3
Thickness of plate (m)
0.00
0.5 0.7 0.9 1.1 1.3
Figure 5. Differential
ThicknessSettlement
of plate (m)L=B=16. Figure 6. Differential Settlement L=2B=32m.

In order to understand the influence of the shape of the mat plate on the differential settlement,
three other shapes of mat plates with the sizes of 32m x16m (L=2B); 48m x 16m (L=3B) and 64m x
16m (L=4B) are analyzed and the simulation results are shown in figures 6, 7 and 8. The results shown
in figures 6-8 are similar to those in figure 5, but it could be seen that the differential settlement of mat
plate increases as the shape of the mat plate becomes longer. This result is in accordance with previous
research [1]. The above results could illustrate the influence of foundation shapes on the differential
settlements.

0.60 0.60
Kv=200
t 0.50 0.50
n Kv=200
e tt
m
el 0.40
Kv=100
e 0.40
tt S. Kv=100
e Kv=50
gv 0.30
S0.30
.g Kv=50
v Kv=30 A/ 0.20
A 0.20 .f
./f fi 0.10 Kv=30
if 0.10 D
D
0.00
0.00
0.5 0.7 0.9 1.1 1.3
0.5 0.7 0.9 1.1 1.3
Thickness of plate (m)
Thickness of plate (m)
Figure 7. Differential Settlement L=3B=48m. Figure 8 : Differential Settlement L=4B=64m.

Then the influence of the size of mat plates on the differential settlements is analyzed. The size of
the mat plates is taken respectively as 24m x 24m; 32m x 32m; 40m x 40m; 48m x 48m and 56m x
56m, and the differential settlements for each size of mat plates mentioned above is respectively
studied and then compared to the previous results of mat size of 16m x 16m. Let define B* = 16m as
the basic width of the mat plate. The differential settlement of mat plate of 16m x 16m is therefore
taken as the basic differential settlement. The ratio of differential settlement of the mat plate with the
above-mentioned five different sizes to the basic differential settlements is here defined as size
modified coefficient for the differential settlement  and also defined the width of mat plate to the
basic width of the mat plate as mat width ratio B/B*. The relationship between the size modified
coefficient  and mat width ratio B/B* for different thickness of mat plates under constant kv = 30
ton/m3 is shown in figure 9. Similar relationships of size modified coefficient  under kv = 50; 100
and 200 ton/m3 are respectively shown in figures 10, 11 and 12. The influence of the size of mat plates
on the differential settlements is therefore understood.

5
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

16.00
d=130m
12.00
d=130m
12.00 d=110m
8.00 d=110m
d=90m a d=90m
a 8.00
d=70m 4.00 d=70m
4.00
d=50m d=50m
0.00
0.00 1.5 2.5 3.5
1.5 2 2.5 3 3.5
B/B* B/B*

Figure 9. Size mod. Coeff.  kv=30 ton/m3. Figure 10. Size mod. Coeff.  kv=50 ton/m3.
d=130m 4.00
6.00 d=130m
d=110m
d=90m
3.00 d=110m
4.00 d=90m
a d=70m
d=50m
a 2.00
d=70m
2.00
1.00 d=50m
0.00
1.5 2 2.5 3 3.5 0.00
B/B* 1.5 2 2.5 3 3.5
B/B*

Figure 11. Size mod. Coeff.  kv=100 ton/m3. Figure 12. Size mod. Coeff.  kv=200 ton/m3.

As discussed above, for any arbitrary mat plate (shape; size and thickness of mat plates) and
coefficient of subgrade reaction, the differential settlement could be obtained by interpolation from the
results in figures 5-8 once the shape and thickness of the mat plate and coefficient of subgrade reaction
are given. Then the size modified coefficient  could be calculated by interpolation from the results
in figures 9-12 as the size of mat plate is known. Finally the actual differential settlement of mat plate
is obtained by multiplying the differential settlement calculated above by the size modified
coefficient  . Later some numerical examples will be used to illustrate the detailed procedures to get
differential settlements of mat foundations.

4. Verification by numerical examples

In order to verify the above-mentioned procedures to calculate the differential settlements of mat
plates, three numerical examples will be shown in the following. In the first example, kv = 50 ton/m3
and the size of mat plate of 32m x 80m (L=2.5B) and the thickness of mat plate of 110cm are assumed.
According to kv = 50 ton/m3 and L=2.5B and thickness of mat plate of 110cm, the differential
settlement could be obtained as 2.56cm by interpolation from the results in figure 6 (L=2B) and figure
7 (L=3B). But the actual width of the mat plate is 32m, the mat width ratio is therefore 2.0. From the
results in figure 10, the size modified coefficient  could be obtained as 1.75. Therefore the actual
differential settlement for this mat plate is 4.47cm. The “SAFE” finite element program also simulates
the mat foundation in the first example, and the model-predicted differential settlement is 4.27cm. The
deviation of the differential settlement according to the procedures suggested in this research and that
by model prediction of “SAFE” program is within the error of 4.5%.
In the second example, kv = 100 ton/m3 and the size of mat plate of 40m x 80m (L=2B) and the
thickness of mat plate of 90cm are assumed. According to kv = 100 ton/m3 and L=2B and thickness of
mat plate of 90cm, the differential settlement could be obtained as 1.99cm from the results in figure 6
(L=2B). But the actual width of the mat plate is 40m, the mat width ratio is therefore 2.5. From the

6
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

results in figure 11, the size modified coefficient  could be obtained as 1.81. Therefore the actual
differential settlement for this mat plate is 3.60cm. The “SAFE” finite element program also simulates
the mat foundation in the second example, and the model-predicted differential settlement is 3.20cm.
The deviation of the differential settlement according to the procedures suggested in this research and
that by model prediction of “SAFE” program is within the error of 11.1%.
In the third example, kv = 200 ton/m3 and the size of mat plate of 48m x 72m (L=1.5B) and the
thickness of mat plate of 70cm are assumed. According to kv = 200 ton/m3 and L=1.5B and thickness
of mat plate of 70cm, the differential settlement could be obtained as 2.67cm by interpolation from the
results in figure 5 (L=B) and figure 6 (L=2B). But the actual width of the mat plate is 48m, the mat
width ratio is therefore 3. From the results in figure 12, the size modified coefficient  could be
obtained as 1.02. Therefore the actual differential settlement for this mat plate is 2.71cm. The “SAFE”
finite element program also simulates the mat foundation in the third example, and the
model-predicted differential settlement is 2.64cm. The deviation of the differential settlement
according to the procedures suggested in this research and that by model prediction of “SAFE”
program is within the error of 2.6%.
From three numerical examples above, it could be seen that the differential settlement evaluated
according to the procedures suggested in this research is quite close to that predicted by the finite
element analysis. Therefore it could be concluded that the evaluated differential settlement of mat
foundation following the procedures suggested in this research is pretty accurate.

5. Application of a practical example

In the following, a practical example would be used to illustrate how to calculate the differential
settlement of mat foundation as well as the coefficient of subgrade reaction as shown in the following:
q=20 ton/m2 Water Table

10m Mat Foundation 24x60m Thickness 90cm

Sand Unit Weight 2.0 ton/m3

e0=1.0, Cc=0.20
10m
Clay Unit Weight 2.0 ton/m3
Hard Stratum

After calculation for the above example, the initial effective stress of the center of the clay stratum
 0’=15 ton/m2 while the effective stress increment   ’=14.5 ton/m2. The thickness of the clay stratum
H0 =10m. Substituting these values into Eq.(1), one could obtain that the average settlement of the mat
foundation  Hc = 29.4cm. Then with the use of Eq.(2), one could get the coefficient of subgrade
reaction kv = 68.1 ton/m3. With the actual size of the mat plate of 24m x 60m (L=2.5B) and thickness
of mat plate of 90cm, the differential settlement could be obtained as 2.26cm by interpolation from the
results in figure 6 (L=2B) and 3.76cm by interpolation from the results in figure 7 (L=3B). Therefore
the differential settlement of the mat plate with L=2.5B should be interpolated as 3.01cm. But the
actual width of the mat plate is 24m, the mat width ratio is therefore 1.5. By interpolation from the
results in figure 10 (kv=50 ton/m3) and in figure 11 (kv=100 ton/m3), the size modified coefficient 
could be obtained as 0.619. Finally the actual differential settlement for this mat plate is 1.86cm.

6. Conclusions

7
ACEER 2019 IOP Publishing
IOP Conf. Series: Earth and Environmental Science 351 (2019) 012018 doi:10.1088/1755-1315/351/1/012018

From the discussion above, the following conclusions could be drawn as:
(1) A method to predict the differential settlement of mat foundation has been proposed in this
research, which is pretty accurate and easy to estimate.
(2) A method to estimate the coefficient of subgrade reaction of foundation soils based on the
consolidation theory has been proposed in this research.
(3) The differential settlement of mat foundation predicted following the procedures suggested in
this research is quite close to that obtained from the finite element analysis.
(4) Finally in this research a practical example of mat foundation is used to illustrate the detailed
procedures to estimate the differential settlement of the mat plate and the coefficient of subgrade
reaction of foundation soils.

References
[1] ACI Committee 1987 Suggested Design Procedures for Combined Footings and Mats, ACI
Committee 336 Report
[2] Bowles J E 1988 Foundation Analysis and Design 4th ed (New York: McGraw-Hill)
[3] Das B M 1997 Principles of Foundation Engineering 4th ed. (Boston, MA: PWS Publishing Co.)
[4] Chen C C, Chay Y C and Hong H K 1994 Investigation Procedures for Structure Design of High
Rise Buildinds Project Report of Architecture Research Institute of the Ministry of Internal Affairs
MOIS830016
[5] Vesic A S 1961 Bending of beams resting on isotropic solid J. Engineering Mechanics Div. ASCE 87
35-53
[6] Timoshenko S and Woinowsky-Krieger S 1959 Theory of Plates and Shells (New York:
McGraw-Hill)
[7] SAFE User‘s Manual 1998 (Berkeley, USA: Computers & Structures Inc.)
[8] Ibrahimbegovic A 1993 Quadrilateral finite elements for analysis of thick and thin plates, Computer
Method. Appl. Mech. Eng. 110 195-209