ECCOMAS Congress 2016

VII European Congress on Computational Methods in Applied Sciences and Engineering

M. Papadrakakis, V. Papadopoulos, G. Stefanou, V. Plevris (eds.)
Crete Island, Greece, 5–10 June 2016

HIGH COMPUTATIONAL EFFICIENCY THROUGH GENERIC

ANALYTICAL FORMULATION FOR LINEAR SOIL PRESSURE
DISTRIBUTION OF RIGID SPREAD RECTANGULAR FOOTINGS
John Bellos1 and Nikolaos P. Bakas2
1
Assistant Professor
School of Architecture, Engineering, Land & Environmental Sciences, Neapolis University Pafos
[email protected]
2
Lecturer
School of Architecture, Engineering, Land & Environmental Sciences, Neapolis University Pafos
[email protected]

Keywords: Rectangular footing, soil pressure distribution, analytical solution, vertical load-
ing, biaxial bending, eccentricity regions, compression zone, computational efficiency.

Abstract. The primary objective of this paper is to resolve and provide generic analytical
formulas concerning the linear pressure distribution of rigid spread rectangular footings and,
consequently, limit computational costs. All five distinct regions of the eccentricity diagram
are related to five possible forms of footing deformations and five discrete shapes of compres-
sion zone. For each region, the linear soil pressure distribution in soil-footing interface, the
neutral axis position, the maximum pressure and the pressure values at the four corners, are
expressed in closed forms as functions of biaxial eccentricities, mean soil pressure and foot-
ing dimensions. Several special cases are also presented, verifying the correctness and the
consistency of the developed analytical formulas and revealing the physical meaning of the
eccentricity diagram. The explicit expressions for responses and resultants enable algorith-
mic implications without iterations, providing high computational efficiency with low compu-
tational cost when forming envelopes for shear forces and bending moment or optimizing the
design of footing geometry and footing reinforcement, etc. Through developed computer soft-
ware, a provided simple example of a rigid spread footing under variable eccentric loading,
demonstrates how the theoretical content of this article is used to perform numerical calcula-
tions. The software itself comprises 3D visualization technology to facilitate visual examina-
tion and validation of the results.
 John Bellos and Nikolaos P. Bakas

1 INTRODUCTION
It is well known that rigid spread rectangular footings resting on elastic soils tend to de-
form in a linearly elastic manner under the action of vertical load and biaxial bending [1, 12].
Such loading conditions may occur either due to eccentric vertical loading or high horizontal
excitation as those used for structural design in high seismic risk areas. Thus, assuming foot-
ing rigidity of adequately high magnitude, it is reasonable to consider a planar settling of foot-
ing base and a linear soil pressure distribution due to constant ratio of pressure to settlement.
The footing-soil interaction generates compression underneath the entire or part of the
footing base, depending on the area being in contact with the soil. This area, usually called
active area or effective area or compression zone [3, 11, 13], should not be less than a certain
percentage of the entire base area. In fact, when the effective eccentricities, computed by di-
viding bending moments with vertical force, lay outside a certain geometric ellipse called
secondary core, the active area is smaller than half the base area. As strictly as required by
regulations and usually recommended by good practice, such cases are not acceptable in foot-
ing design.
The most important issue concerning rectangular footings is optimal geometric and rein-
forcement design [2, 3, 6, 7, 14]. Initially, a set of suitable values is chosen so that, under ser-
viceability conditions, the maximum pressure exerted at any footing position does not exceed
the soil bearing capacity. Moreover, constraints set for enveloped bending, shearing and
punching should not be violated under ultimate loading conditions, while requirements im-
posed by regulations should be fully met. All the above involve an immense amount of inten-
sive calculations through complex iterative processes. Several procedures were developed in
the past for handling efficiently such footing-soil interaction problems, by using either analyt-
ical [2, 4, 5, 6, 7, 13, 14], numerical [10, 11], graphical [3, 15], or hybrid techniques [8, 9].
Although accurate and useful, some of them are rather computationally expensive, while oth-
ers are not so wide-ranging or not easily implementable in designing footings of such specific
shape, yet participating in a complex structure assembly subject to various loading actions.
On the other hand, the generic explicit formulas presented in this article increase the efficien-
cy of the required design processes by reducing considerably the computational costs.
Another advantage of the presented analytical method is that, instead of relating eccentrici-
ties to external loading conditions, they are evaluated by using reaction resultants at the re-
spective support node as derived from the space frame model of the structure. Hence, these
effective eccentricities express a clear relation of biaxial bending moments to vertical load for
the specific footing, since they are not directly associated with external loading and influences
from neighboring columns, footings and connecting beams.
This paper attempts to enlighten the issues concerning the rigid spread rectangular footings
by providing easily programmable analytical formulas and, through a developed intelligent
software and 3D visualization technology, facilitate results inspection, confirm engineering
intuition, unshed light to foundation settlement and encourage future research in related fields.

2 THE MATHEMATICAL MODEL

As we stated previously, a rigid spread rectangular footing is expected to deform in a pla-
nar manner. Therefore, it is reasonable to consider a linear soil pressure distribution under-
neath the base, that is
 x y 
p( x, y )  pmax      1 (1)
 xn yn 
 John Bellos and Nikolaos P. Bakas

where pmax is the maximum soil pressure on the base and xn, yn are the intercepts of the neutral
axis determining the active area of the footing (see Figure 1).

Figure 1: Linear soil pressure distribution on a rigid spread rectangular footing.

The effective eccentricities from the footing center, derived from certain reaction resultants
at the central support node, i.e. the calculated values of biaxial moments Mx, My and vertical
force P (see Figure 1), are expressed as
Mx My
ex  , ey  (2)
P P
Equilibrium of actions and reactions at the base center (xc, yc) requires
P   p( x, y )  dA (3)
A

M x   x c  x   p( x, y)  dA (4)
A

M y    yc  y   p( x, y )  dA (5)
A

Substituting equation (3) into equations (4) and (5), they are transformed into
M x  P  x c   x  p( x, y)  dA  P  x c  e x    x  p( x, y)  dA (6)
A A

M y  P  yc   y  p( x, y)  dA  P  yc  ey    y  p( x, y)  dA (7)
A A

The mean soil pressure on a footing with dimensions lx, ly is defined as

P
pm  (8)
lx  l y
 John Bellos and Nikolaos P. Bakas

If, by convention, one uses absolute values for eccentricities and thus presets the maximum
soil pressure at the lower left footing corner, i.e. po = pmax (see Figure 1). Then equations (3),
(6) and (7) can be written as
pm  lx  l y   p( x, y)  dx  dy (9)
A

l  1 e 
pm  lx  l y   x  ex    x  p( x, y)  dx  dy  pm  lx2  l y    x    x  p( x, y)  dx  dy (10)
2  A  2 lx  A

 ly   1 ey 
   y  p( x, y)  dx  dy  p m  l x  l y2      y  p( x, y)  dx  dy (11)
 A
p m  l x  l y    e y 
2  A  2 l
 y

There is a direct relation of pressure values at actual footing corners with those computed
from at rearranged footing corners, with the order strictly depending on the eccentricities
signs, i.e.
e x  0 & e y  0  p1  p o  p max , p 2  p p , p 3  p r , p 4  p q
e x  0 & e y  0  p1  p p , p 2  p o  p max , p 3  p q , p 4  p r
(12)
e x  0 & e y  0  p1  p r , p 2  p q , p 3  p o  p max , p 4  p p
e x  0 & e y  0  p1  p q , p 2  p r , p 3  p p , p 4  p o  p max

Solving the system of equations (9), (10) and (11) and determining the unknown parame-
ters xn, yn and pmax, then the equation (1) depends on the signs of effective eccentricities and,
therefore, it can be rewritten as
 x y 
p( x, y)  pmax   sign(ex )   sign(ey )   1 (13)
 xn yn 
and by relocating the origin of the coordinate system at the footing centroid, the linear soil
pressure distribution can be written in an equivalent to equation (1) expression, that is
 x' y' 1l l 
p( x' , y ' )  pmax   sign(ex )   sign(ey )   1   x  y  (14)
 xn yn 2  xn yn 

3 THE ECCENTRICITY DIAGRAM

As effective eccentricities, evaluated using equations (2), sweep the rectangular footing
base certain types of deformation occur, thus generating an eccentricity diagram for each
footing, as presented in Figure 2. This diagram, also obtained by other authors using different
approaches [13], constitutes a very strong practical tool for determining the form of pressure
distribution expected under certain loading conditions. In this work, the border lines on the
eccentricity diagram are provided explicitly while special cases reveal their physical meaning.
There are five distinct regions in this diagram, namely A-E, corresponding to different
types of deformation and active areas of unique geometric shapes (see Figure 2). Each region-
al case is subdivided into four subcases, each associated with one primary footing corner of
maximum soil pressure, as determined by eccentricities signs in equations (12). Notice that
there is a rhombus domain, usually called main core, having semi-diagonals equal to 1/6 of
the respective footing lengths and a surrounding elliptic domain, usually called secondary
 John Bellos and Nikolaos P. Bakas

core, bounded by an ellipse with major and minor semi-axes equal to 1/3 of the respective
footing lengths (dashed line in Figure 2). The first contains the entire regions C, while the
second comprises regions B and certain parts of regions D and E. The remaining parts of re-
gions D and E and the entire regions A are excluded, since corresponded to active areas
smaller than half the footing base area. Regulations as well as good practice require that effec-
tive eccentricities should remain inside the secondary core in order to maintain an acceptable
footing design.

Figure 2: The eccentricity diagram of a rigid spread rectangular footing.

4 THE ECCENTRICITY REGIONS

The analytical formulas for all five regions are presented in the subsections below. They
are produced by integrating the equations (9), (10) and (11) over the active area of the footing
base corresponding to each regional case. Nevertheless, for saving space and time, the de-
tailed development process of the formulas is not provided here.

4.1 Region C: Full Compression Zone

Regions C corresponds to four equal right triangles composing a rhombus shape with semi-
diagonals equal to 1/6 of the respective footing lengths (see Figure 3). This rhombus area,
usually called main core, represents small eccentricities generated by vertical forces applied at
points in it and capable of setting the entire footing base under compression. Maximum pres-
sure occurs at the footing corner corresponding to the right triangle enclosing the eccentricity
point.
 John Bellos and Nikolaos P. Bakas

Figure 3: Full compression zone.

The intercepts of the neutral axis are given by

 e ey   e   e ey   e 
xn  l x  1  6  x  6   12  x  , yn  l y  1  6  x  6   12  y  (15)
 lx ly   lx   lx ly   ly 
     
The maximum soil pressure at the primary corner of the footing base (coordinate system
origin O) along with the soil pressure at the rest corners, are provided by the equations
 e ey   e ey 
po  pmax  pm  1  6  x  6  , p p  p m  1  6  x  6  ,
 lx ly   lx ly 
   
(16)
 e ey   e ey 
p q  p m  1  6  x  6  , pr  pmin  pm  1  6  x  6  
 lx ly   lx ly 
   
In this particular case, since the active area is the entire footing base, all footing corners
should be under compression, i.e. po ≥ 0 & pp ≥ 0 & pq ≥ 0 & pr ≥ 0. Since pr represents the
minimum compression value, the last of equations (16) yields the rhombus shape boundary
for regions C shown in Figure 3, i.e.
ex ey 1
  (17)
lx ly 6

4.2 Region A: Triangular Compression Zone

Regions A corresponds to four smaller corner rectangles, each having sides equal to 1/4 of
the respective footing lengths (see Figure 4). They represent excessive eccentricities generat-
ed by vertical forces applied at points inside these corner rectangles, imposing high pressure
to respective triangular compression zones of the footing base. Maximum pressure occurs at
the footing corner associated with the corner rectangle enclosing the eccentricity point.
 John Bellos and Nikolaos P. Bakas

Figure 4: Triangular compression zone.

The neutral axis intercepts are given by

1 e   1 ey 
xn  4  l x    x , yn  4  l y     (18)
 2 lx   2 ly 
 
The soil pressure values at the footing corners are provided by
3  pm
po  pmax  , p p  pq  pr  pmin  0 (19)
 1 ex   1 e y 
8     

 2 l x   2 l y 
For triangular compression zone the intercepts should be confined within the neighboring
sides of the footing base. Hence, equations (18) yield the rectangular shape boundary for re-
gion A (see Figure 4), expressed as the lower and upper bounds in the two inequalities
1 ex 1
0  xn  l x   
4 lx 2
(20)
1 ey 1
0  yn  l y   
4 ly 2

where the upper bounds denote that eccentricities outside the footing base are unbalanced.

4.3 Region D: Trapezoidal Compression Zone in x-Direction

Regions D corresponds to four curved trapezoids attached to the x-direction footing sides
in pairs, each having bases equal to 1/3 and 1/4 of the y-direction footing length respectively,
altitude 1/4 of the x-direction footing length and a curved side (see Figure 5). They represent
large eccentricities generated by vertical forces applied at points inside these curved trape-
zoids, imposing pressure to respective trapezoidal zones covering the entire x-direction foot-
 John Bellos and Nikolaos P. Bakas

ing side. Maximum pressure occurs at the footing corner corresponding to the curved trape-
zoid enclosing the eccentricity point.

Figure 5: Trapezoidal compression zone covering the entire side in x-direction.

The neutral axis position, determined by its intercepts and its trace yp on the parallel side
(see Figure 5), is provided by
 e e2   e 
xn  l x  1  6  x  1  12  2x  12  x  ,
 lx l x   lx 

 1 ey   ex2   ex ex2   ex  ex2 
1
yn   l y     
 2  1  12  2  1  2   1  12  2  
 1  4  2  ,
 2 ly    (21)
2 l x   lx lx 
 lx  l x 
   
 l 
y p  yn  1  x 
 xn 
The soil pressure values at the footing corners are
 ex2   2   1 ey 
po  pmax
p
 m 
 2  1  12  2   1  6  ex  1  12  ex    ,
3  lx   lx l x2   2 ly 
      (22)
 l 
p p  pmax  1  x , pq  pr  pmin  0
 xn 
Requiring the x-intercept to lie on the extension of the respective footing side, the bases of
the trapezoidal region are given as the lower and upper bounds of an inequality, i.e.
xn ex 1
1  0  (23)
lx lx 4
 John Bellos and Nikolaos P. Bakas

In fact, the upper bound of the inequality guaranties that the discriminant, appearing under the
square roots of equations (21) and (22), is always positive definite.
Confining the y-intercept within the respective footing side, the lateral sides of the trape-
zoidal region are given as the lower and upper bounds of another inequality (see Figure 5)

yn 1 1  e2  
  1  2  ex  1  12  ex
2 
  e  ey 1
0 1   2  1  12  2x 1  4  x    (24)
2 6    l x2   l x  l y 2
ly lx   lx  
Here, the lower bound of the inequality expresses the curved side of trapezoid while the upper
bound states that eccentricities outside the footing base are unbalanced.

4.4 Region E: Trapezoidal Compression Zone in y-Direction

Regions E corresponds to four curved trapezoids attached to the y-direction footing sides
in pairs, each having bases equal to 1/3 and 1/4 of the y-direction footing length respectively,
altitude 1/4 of the y-direction footing length and a curved side (see Figure 6). They represent
large eccentricities generated by vertical forces applied at points inside these curved trape-
zoids, imposing pressure to respective trapezoidal compression zones covering the entire y-
direction footing side. Maximum pressure occurs at the footing corner corresponding to the
curved trapezoid enclosing the eccentricity point.

Figure 6: Trapezoidal compression zone covering entire side in y-direction.

The position of the neutral axis is provided by

1 e   e y2   ey e y2   e y  e y2 
 l x    x    2  1  12  2   1  2   1  12  2    1  4  2  ,
1
xn 
2 2 l  
 y  ly   ly l y   l y  l y 
     
 ey e y2   ey   ly 
y n  l y  1  6   1  12  2  12  , xq  xn  1   (25)
 ly ly   ly   yn 
   
 John Bellos and Nikolaos P. Bakas

The soil pressure values at the footing corners are

 e y2   ey e y2   1 ex 
po  pmax
p
 m 
 2  1  12  2  1  6   1  12  2    
 ly     2 l ,
3
  
ly ly
  x 
(26)
 ly 
pq  pmax  1  , p p  pr  pmin  0
 yn 
Confining properly the intercepts of neutral axis with respect to footing sides in equations
(25), the curved trapezoid is determined by the following two inequalities

1 1  e y2     e 
2

0
xn
1   2  1  12  2   1  2  e y  1  12  e y  1  4  y   e x  1
lx 2 6  ly   ly l y2   l y  lx 2
      (27)
yn ey 1
1  0 
ly ly 4

The bounds of the first inequality express the lateral sides of the curved trapezoid while the
bounds of the second its bases.

4.5 Region B: Pentagonal Compression Zone

Regions B corresponds to four diagonally placed curved triangles, each having two curved
sides and one flat base, as shown in Figure 7. These curved triangles form part of the second-
ary core, inscribed in its elliptic boundary and surrounding the main core (see Figure 2). They
represent moderate eccentricities generated by vertical forces applied at points inside them,
imposing pressure to respective pentagonal compression zones while preserving a triangular
uncompressed zone. Maximum pressure occurs at the footing corner corresponding to the
curved triangle enclosing the eccentricity point.

Figure 7: Pentagonal compression zone (triangular uncompressed zone).

 John Bellos and Nikolaos P. Bakas

As shown below, in full agreement with other sources [13], in this case the solution cannot
be expressed in closed form. In fact, the intercepts of the neutral axis are provided by recur-
sive formulas, i.e.
3 3
 l   l 
1  1  x   1  y 
1 e   xn   yn 
xn  4  l x    x   4 4 3
,
 2 lx   lx   ly   lx  lx
1  1    1    4  1   
 xn   yn   xn  x n
3 3
(28)
 l   l 
1  1  x   1  y 
 1 ey   xn   yn 
yn  4  l y    
 2 ly  4 4 3
  1  1  l x   1  l y   4  1  l y   l y
 xn   yn   yn  yn
 
One can easily prove, either numerically or algebraically, that the set of equations (27) consti-
tutes a fast converging iterative procedure. Consequently, once the intercept values found they
are substituted to all consequent equations.
The traces of the neutral axis on the two opposite footing sides, as shown in Figure 7, are
 l   ly 
y p  yn  1  x , xq  xn  1   (29)
 xn   yn 
The soil pressure values at the footing corners are
lx l y
6  pm  
xn y n
po  pmax  3 3
,
 lx   ly 
1  1    1   (30)
 xn   yn 
 l   ly 
p p  pmax  1  x , pq  pmax  1  , pr  pmin  0
 xn   yn 
Since, on its limits, region B degenerates to regions C, D and E respectively, the boundary
of the curved triangle is clearly determined by the set of three inequalities

ex ey 1
 
lx ly 6

ex 1 1  e y2  
  1  2  ey e y2   e 
1  4  y 
   2  1  12  2  1  12  2 (31)
lx 2 6  ly   ly ly   ly 
     
ey 1 1  e2   ex e2   e 
    2  1  12  2x   1  2   1  12  2x  1  4  x 
ly 2 6  lx  
  lx l x   lx 

The first inequality represents eccentricities outside the main core of rhombus region C, while
the rest represent eccentricities inside the secondary zone bound by the curved sides of the
trapezoidal regions D and E respectively (see Figure 2), as derived in the respective chapters.
 John Bellos and Nikolaos P. Bakas

5 SPECIAL CASES
Several borderline cases are represented here, verifying the correctness and the consistency
of the developed analytical formulas, reassuring the regional continuity and revealing the
physical meaning of the eccentricity diagram.

5.1 One-way Eccentricity in Region C

Imposing x-direction eccentricity in Region C by setting ey = 0, as shown in Figure 8,
equations (15) and (16) yield the well-known formulas for soil pressure distribution of single
slope
 e   e 
xn  l x  1  6  x  12  x  , yn   (32)
 lx   lx 

 e   e 
po  pq  pmax  pm  1  6  x , p p  pr  pmin  pm  1  6  x  (33)
 lx   lx 
while inequality (17) diminishes to representing segment ‘kc’, that is
ex 1
 (34)
lx 6

Figure 8: Soil pressure distribution for one-way eccentricity in segment 'kc'.

5.2 One-way Eccentricity in Region C

Imposing x-direction eccentricity in Region E by setting ey = 0, as shown in Figure 9,
equations (25) and (26) yield the well-known formulas for soil pressure distribution of single
slope and limited active area
1 e 
xn  3  l x    x , yn   (35)
 2 lx 
 John Bellos and Nikolaos P. Bakas

2  1 ex 
po  pq  pmax   pm    , p p  pr  pmin  0 (36)
3  2 lx 
while inequality (27) diminishes to representing segment ‘ce’, that is
1 ex 1
  (37)
6 lx 2

Figure 9: Soil pressure distribution for one-way eccentricity in segment 'ce'.

5.3 One-way Eccentricity in Junction of Regions C, E and B

Figure 10: Soil pressure distribution for one-way eccentricity in point 'c'.

Applying x-direction eccentricity in point ‘c’, as shown in Figure 10, by setting

 John Bellos and Nikolaos P. Bakas

ex 1
 , ey  0 (38)
lx 6
the equations (15) and (16) of region C, (25) and (26) of region E, and (28) and (30) of region
B, yield identical results
xn  l x , yn   (39)
po  pq  pmax  2  pm , p p  pr  pmin  0 (40)
In such a case, the soil pressure distribution of single slope has its neutral axis coincide
with the footing base side in y-direction, as illustrated in Figure 10.

5.4 Biaxial Eccentricity in Junction of Regions A, D, E and B

Figure 11: Soil pressure distribution for eccentricity in point 'a'.

Applying biaxial eccentricity in point ‘a’, as shown in Figure 11, by setting

ex 1 ey 1
 ,  (41)
lx 4 ly 4
the equations (18) and (19) of region A, (21) and (22) of region D, (25) and (26) of region E,
and (28) and (30) of region B, yield identical results
xn  l x , yn  l y (42)

po  pmax  6  pm , p p  pq  pr  pmin  0 (43)

In such a case, the soil pressure distribution has its neutral axis coincide with the respective
diagonal of the orthogonal base, as illustrated in Figure 11.
 John Bellos and Nikolaos P. Bakas

6 APPLICATION AND DISCUSSION

A simple example is provided here to demonstrate how the generic analytical formulas of
this article are used to perform numerical calculations via developed computer software, thus
offering high computational efficiency by limiting computational costs. The software itself
comprises a 3D visualization module, facilitating visual examination and validation of the re-
sults. Note that all figures and tables of this example have been produced through this soft-
ware and can be reproduced by rerunning it, practically offering an infinite number of
variations to the researcher/engineer.
The structural model concerns a 0.25mx0.25m square column of height 3.00m fixed to a
1.00mx1.00m rigid spread square footing of depth 0.50m. The material used is concrete
C30/37 reinforced with steel B500C, with safety factors 1.50 and 1.15 for concrete and steel
respectively, and applied with covering of 0.03m for the column and 0.065 for the footing.
The central footing studied for variable eccentric loading required by building codes. Eu-
rocode 1 (EN1991, EC1), Eurocode 2 (EN1992, EC2) and Eurocode 8 (EN1998, EC8) ap-
plied in our case, complemented by Greek National Annexes [16]. Accordingly, 33
combinations of actions required in total, as presented in Table 1, where G stands for dead
loads, Q for imposed loads and E for seismic loads to primary directions with accidental ec-
centricities to perpendicular directions.

A = 1.35G + 1.50Q, I = 1, 2, 3, 4
i = 1.00G + 0.30Q + 1.00Εx±eccy + 0.30Ey±eccx,
Β iC = 1.00G + 0.30Q + 1.00Ex±eccy - 0.30Ey±eccx,
iD = 1.00G + 0.30Q + 0.30Ex±eccy + 1.00Ey±eccx, iΕ = 1.00G + 0.30Q - 0.30 Ex±eccy + 1.00Ey±eccx,
iF = 1.00G + 0.30Q - 1.00Ex±eccy - 0.30Ey±eccx, iG = 1.00G + 0.30Q - 1.00Ex±eccy + 0.30Ey±eccx,
iΗ = 1.00G + 0.30Q - 0.30Ex±eccy - 1.00Ey±eccx, iΙ = 1.00G + 0.30Q + 0.30 Ex±eccy - 1.00 Ey±eccx,

Table 1: The 33 combinations of actions required.

ex ey pm p1 p2 p3 p4 Vxf Vyf Mxf Myf

Comb. (m) (m) (kPa) (kPa) (kPa) (kPa) (kPa) (kN) (kN) (kNm) (kNm)
A 0.000 0.000 23.2 23.2 23.2 23.2 23.2 8.7 8.7 1.6 1.6
1,2,3,4B -0.076 -0.003 17.2 7.2 22.5 27.1 11.8 9.1 9.7 1.8 1.8
1,2,3,4C -0.076 0.003 17.2 11.8 27.1 22.5 7.2 9.1 9.7 1.8 1.8
1,2,3,4D -0.003 -0.076 17.2 7.2 11.8 27.1 22.5 9.7 9.1 1.8 1.8
1,2,3,4E 0.003 -0.076 17.2 11.8 7.2 22.5 27.1 9.7 9.1 1.8 1.8
1,2,3,4F 0.076 0.003 17.2 27.1 11.8 7.2 22.5 9.1 9.7 1.8 1.8
1,2,3,4G 0.076 -0.003 17.2 22.5 7.2 11.8 27.1 9.1 9.7 1.8 1.8
1,2,3,4H 0.003 0.076 17.2 27.1 22.5 7.2 11.8 9.7 9.1 1.8 1.8
1,2,3,4I -0.003 0.076 17.2 22.5 27.1 11.8 7.2 9.7 9.1 1.8 1.8

Table 2: Effective eccentricities and resultants for the square central footing for all combinations.

Since no imposed loads exist (Q=0) and no accidental eccentricities considered in this ex-
ample (eccx=eccy=0), only 9 out of the 33 combinations are different. The effective eccentrici-
ties ex, ey, the mean soil pressure pm, the soil pressure at the corners p1, p2, p3, p4 and the
maximum shearing forces Vxf, Vyf and bending moments Mxf, Myf, are presented in Table 2. It
must be underlined that the effective eccentricities have been calculated using equations (2),
 John Bellos and Nikolaos P. Bakas

the soil pressure values via the explicit formulas applicable in different eccentricity regions
along with mapping equations (12), while the rest of the resultants through algebraic integra-
tion of soil pressure distribution provided by equation (14). Notice that all combinations, but
‘A’, impose effective eccentricities despite the fact the footing itself has no structural eccen-
tricity. This was expected since ‘A’ is the only combination formed exclusively by vertical
loads while only horizontal loads can generate rotational reactions in central footings. The soil
pressure distributions for different combinations and their envelope are illustrated in Figure
12, while the resultants presented in Table 2 and their envelopes are also shown in Figure 13.

Figure 12: Soil pressure distributions for different combinations and envelope.

Figure 13: Resultants for different combinations and envelopes:

(a) Soil pressure, (b) Shear forces, (c) Bending moments.

The footing is also studied for variable structural eccentricity with respect to the column.
The enveloped maxima of bending moments, namely Mxd, Myd, are presented in Figure 14 as
a function of structural eccentricity in x-direction. They are also called design bending mo-
ments since they are associated with footing design. Notice that only for central footings the
two bending components are identical. In fact, by increasing structural eccentricity bending is
affected substantially in the direction of structural eccentricity and slightly in the perpendicu-
lar direction. This reflects directly on the calculated two-way reinforcement Asx, Asy, shown in
 John Bellos and Nikolaos P. Bakas

Figure 15 as a function of structural eccentricity in x-direction, where required reinforcement

increases substantially in the direction of structural eccentricity.

Figure 14: Design bending moments in relation to structural eccentricity in x-direction.

Figure 15: Calculated two-way reinforcement in relation to structural eccentricity in x-direction.

By analyzing the above outcomes, it is clear that there is a substantial amount of calcula-
tions involved in the design of rigid spread footings geometry and reinforcement, even if a
limited number of combinations is taken into account. Of course, the numbers increase dra-
matically when optimization is attempted via iterative procedures or when one seeks result-
ants for multi footings foundations (see Figure 16). In such cases, the explicit formulas
revealed in this article are certainly preferable over numerical methods or similar techniques,
since they undoubtedly limit computational costs. Furthermore, the proposed here analytical
 John Bellos and Nikolaos P. Bakas

method is applicable to either a single footing case or a multi footings and connecting beams
assembly. Only resultant reactions retrieved from the space frame model are used as input in
equations (2), and therefore the process is independent from linked neighboring components.

Figure 16: Soil pressure distributions and envelopes for a multi rigid spread rectangular footings foundation.

7 CONCLUSIONS
This article adds value in studying rigid spread rectangular footings resting on elastic soils
by offering:
1. High computational efficiency through developed explicit formulas for soil pressure
values in different eccentricity regions.
2. A generic, robust and effective solution, independent of the foundation layout, by using
as input the reactions computed at the supports instead of actual loads on the footing.
3. An eccentricity diagram per footing, as a very practical qualitative tool for predetermin-
ing the form of soil pressure distribution and footing deformation expected.
4. An integrated virtual reality software, facilitating visual examination and validation of
the results.
The innovative and straightforward research approach followed here will assist future re-
searchers by providing a valuable reference through a powerful research software tool. Poten-
tially, it could be extended to accounting for more generic footing geometries and non-linear
soil pressure distributions.

ACKNOWLEDGEMENTS
The authors would like to express their gratitude to pi-Systems International S.A. for the
constant financial support provided during the project, the Institute of Structural Analysis and
Seismic Research of National Technical University of Athens for the contribution in develop-
ing the computational algorithms for the space frame model and all associates involved in the
production of the integrated virtual reality software.
 John Bellos and Nikolaos P. Bakas

NOTATION
A Active area of footing
P Vertical force reaction on footing
Mx, My Rotational moment reactions on footing in x and y directions
p(x, y) Soil pressure distribution
p1, p2, p3, p4 Soil pressure values on footing corners
po, pp, pq, pr Soil pressure values on rearranged footing corners
pmax Maximum soil pressure value on footing
pmin Minimum soil pressure value on footing
pm Mean soil pressure value on footing
lx, ly Footing dimensions in x and y directions
ex, ey Effective eccentricities from footing center in x and y directions
xc, yc Coordinates of footing base centroid
xn, yn Neutral axis intercepts on x and y axes
x, y, z Coordinate system with origin on footing corner
x’, y’, z’ Coordinate system with origin on footing centroid
sign( ) Sign of a real number
Vxf, Vyf Maximum values of shear forces on footing in x and y directions
Mxf, Myf Maximum values of bending moments on footing in x and y directions
Mxd, Myd Design bending moments of footing in x and y directions
Asx, Asy Calculated steel reinforcement of footing in x and y directions

REFERENCES
[1] J. E. Bowles, Foundation Analysis and Design, 3rd Edition. McGraw Hill, New York,
1977.
[2] R. Jarquio, V. Jarquio, Design of Footing Area with Biaxial Bending. Journal of Ge-
otechnical Engineering, 109, 1337-1341, 1983.
[3] W. H. Highther, J. C. Anders, Dimensioning Footings Subjected to Eccentric Loads.
Journal of Geotechnical Engineering, 111, 659-665, 1985.
[4] D. M. Vitone, A. J. Valsangkar, Stresses from Loads over Rectangular Areas. Journal
of Geotechnical Engineering Division, 112, 961-964, 1986.
[5] R. Irles, F. Irles, Explicit Stresses under Rectangular Footings. Journal of Geotech-
nical Engineering, 120, 444–450, 1994.
[6] H. M. Algin, Stresses from Linearly Distributed Pressures over Rectangular Areas. In-
ternational Journal of Numerical Analysis Methods in Geomechanics, 24, 681–692,
2000.
[7] H. M. Algin, Practical Formula for Dimensioning a Rectangular Footing. Engineering
Structures, 29, 1128-1134, 2007.
 John Bellos and Nikolaos P. Bakas

[8] J. P. Smith-Pardo, A. Bobet, Behavior of Rigid Footings on Gravel under Axial Load
and Moment. Journal of Geotechnical and Geoenvironmental Engineering, 133, 1203–
1215, 2007.
[9] N., Yamamoto, M. F. Randolph, I. Einav, Simple Formulas for the Response of Shal-
low Foundations on Compressible Sands. International Journal of Geomechanics, 8,
230–239, 2008.
[10] M. H. T. Rayhani, M. H. El Naggar, Numerical Modeling of Seismic Response of Rig-
id Foundation on Soft Soil. International Journal of Geomechanics, 8, 336–346, 2008.
[11] A. Ivan, Pressures Distribution for Eccentrically Loaded Rectangular Footings on Elas-
tic Soils. Proceedings of the 2010 International Conference on Mathematical Models
for Engineering Science, 213-216, 2010.
[12] N. S. V. Kameswara Rao, Analysis of Footings on Elastic Foundations, Chapter 5,
Foundation Design: Theory and Practice. John Wiley & Sons (Asia) Pte Ltd, 165–202,
2011.
[13] G. Özmen, Determination of Base Stresses in Rectangular Footings under Biaxial
Bending. Teknik Dergi, 22, 5659-5674, 2011.
[14] A. L. Rojas, J. G. F. Herrera, R. A. A. Vallejo and M. A. C. Alvarez, Design of Isolat-
ed Footings of Rectangular Form Using a New Model. International Journal of Inno-
vative Computing, Information and Control, 9, 4001-4021, 2013.
[15] J. A. Rodriguez-Gutierrez, J. D. Aristizabal-Ochoa, Rigid Spread Footings Resting on
Soil Subjected to Axial Load and Biaxial Bending. II: Design Aids. International
Journal of Geomechanics, 13, 109-119, 2013.
[16] A. Konstantinides, J. Bellos, Earthquake Resistant Buildings made of Reinforced Con-
crete: Static and Dynamic Analysis according to Eurocodes, Volume B. Alta Grafico,
Athens, 2013.