Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct
Short communication
a r t i c l e
i n f o
Article history:
Received 17 February 2011
Revised 25 September 2011
Accepted 3 October 2011
Available online 18 November 2011
a b s t r a c t
This manuscript presents the development of charts that structural engineering practitioners can use to
quickly quantify the restraining effect at the bases of columns and walls supported on shallow foundations. The formulation is derived based on compatibility, equilibrium, and a simplied constitutive relation that uses two basic parameters to characterize the soil-foundation: the subgrade modulus the
stiffness parameter and the ultimate bearing capacity the strength parameter. The model captures
the intrinsic non-linear behavior of the soil with increasing loading, and the coupling between moments
and axial loads. Using the proposed charts, calculated rocking responses for rigid footing models under
combined axial load and moment are found to compare reasonably well with experimental results.
2011 Elsevier Ltd. All rights reserved.
1. Introduction
Seismic rehabilitation standards [1] for buildings located in
moderate and high seismic risk zones require consideration of
the interaction between the structure and the supporting soil.
Studies on soilstructure interaction problems can be traced as
far back as the work conducted by Lamb [2], more than one
hundred years ago, on the propagation of elastic waves induced
by the application of a point load on the surface of an elastic
semi-innite media. Bycroft [3] was perhaps the rst to present a
complete set of solutions for vertical and horizontal translation,
and rotation about a horizontal axis (rocking) and about a vertical
axis (torsion) of a rigid circular plate on an elastic half-space. A
comprehensive presentation of decades the of theoretical and
experimental research on vibration of foundations is presented in
the classic textbook by Richart et al. [4].
Some of the rst studies on soilstructure interaction that used
the analytical and experimental results from elastic half-space theory were carried out by Hall [5], Parmelee [6] and Parmelee et al.
[7]. Although the main limitation was the dependency of the
impedance on the exciting frequency, these became the rst
attempts to establish a bridge between the elastic half space theory
and the mass-spring-dashpot system.
General formulas and charts for the calculation of impedances
(dynamic stiffness and damping) of surface and embedded foundations of any shape were presented by Gazetas [8] and incorporated
into the pre-standard FEMA 273 [9]. The formulas were derived on
the basis of Finite Element and Boundary Element analyses and
Address: 901 12th Ave., Seattle, WA 98122-1090, USA. Tel.: +1 206 296 5901;
fax: +1 206 296 2173.
E-mail address: smithjh@seattleu.edu
0141-0296/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2011.10.005
Kh
GB3
L
0:4
0:1
B
1m
573
K s0 Bdn
1 ndn
@r
@dn
ks dn
K s0 B
K s0 B
r0
1 ndn 2
d hx B=2
1
dn ha
B
2
dn a
where, a = x/B: normalized abscissa. dn d =B: normalized settlement at midpoint of the footing.
The equation of equilibrium that relates increments in the applied vertical load, dP, and moment, dM, with increments in the normalized midpoint settlement, dn ; and rotation, dh, can be written as:
dP=P0
dM=P0 B
2. Soil-foundation model
It is noticed that Eqs. (2) and (3) are dened in terms of parameters that structural engineers can comfortably relate to: (i) the initial
subgrade modulus, Ks0 which is a measure of foundation stiffness
and, (ii) the bearing capacity of the soil, r0 which is a measure of
foundation strength. Using Eq. (2), the tangent subgrade modulus
for any settlement normalized by the initial subgrade modulus is
found to depend solely on the stiffness to strength ratio, n:
kd
kdh
khd
kh
ddn
dh
P0 r0 BL
The terms of the tangent stiffness matrix of the soil-footing system are found by respectively imposing a unit increment in the
midpoint settlement and a unit increment in the rotation and
determining the corresponding resultant axial load and moment
reaction [18]. In terms of the introduced non-dimensional parameters (Fig. 2) this gives:
kd n
ks ada
ks aa 1=22 da
kh n
= P/(BL)
kdh khd n
ks aa 1=2da
10
Ks0B
Test results
P
( / B)
( / B)
0 = lim
/ B
/ B= 0
( / B)
K 5 a
Fig. 1. Concentric load test in terms of subgrade modulus and bearing capacity
stress.
8
<
1
for dn a P 0
1 ndn a
:
0
for dn a 6 0
11
The formulation can be used to calculate the rocking and settlement response for foundations under combined axial load and moment. In particular, the calculated moment-versus-rotation
relationships for constant axial loads of P/P0 = 0.1, 0.3, 0.5, and 0.7
and foundation stiffness to strength ratios, n, from 10 to 1000 are
shown in Figs. 36. The resulting charts are presented in terms of
non-dimensional parameters and thus readily available to be used
as design aids.
574
P/P 0
M
M/(P 0B)
(B) Normalized
(A) Actual
Ks0B
n ()
Hyperbolic
function
ks()
ks()
1
/B
Fig. 2. Soil-foundation model.
Ks0 B
0.050
Plastic Moment
0.045
1000
500
0.040
200
M/(P0B)
0.035
100
50
0.030
30
10
0.025
0.020
0.015
0.010
0.005
0.000
0
0.01
0.02
0.03
0.04
0.05
(rad)
Fig. 3. Foundation momentrotation response for P/P0 = 0.1.
4. Plastic moment
The plastic moment capacity of a foundation can be estimated
using a simple expression suggested by Meyerhof [16]. This is classically known as the equivalent width concept and allows to relate the axial capacity of a footing under concentric loading to that
of a footing under eccentric loading (on the same soil) by assuming
that in both cases the soil can fully plastify at a stress r0 (Fig. 7). If
the concentric load capacity is P0, then the eccentric load capacity P
for an eccentricity, e, can be calculated as:
2e
P P0 1
B
12
Mp 1 P
P
1
P0
P0 B 2 P0
13
575
0.110
Plastic Moment
0.100
1000
0.090
500
M/(P0B)
0.080
0.070
200
0.060
100
0.050
50
0.040
30
0.030
10
0.020
0.010
0.000
0
0.01
0.02
0.03
0.04
0.05
(rad)
Fig. 4. Foundation momentrotation response for P/P0 = 0.3.
0.140
M/(P 0B)
Plastic Moment
0.120
1000
0.100
500
0.080
200
100
0.060
50
0.040
30
0.020
10
0.000
0
0.01
0.02
0.03
0.04
0.05
(rad)
Fig. 5. Foundation momentrotation response for P/P0 = 0.5.
110
170
0:65
From the same gure, the bearing capacity stress is approximately: r0 = 0.65 MPa. The foundation bearing capacity under concentric loading alone is therefore:
K s0 B
@ r
110MPa
@dn 0
576
0.120
Plastic Moment
0.100
1000
M/(P 0B)
0.080
500
200
0.060
100
0.040
50
30
0.020
10
0.000
0
0.01
0.02
0.03
0.04
0.05
(rad)
Fig. 6. Foundation momentrotation response for P/P0 = 0.7.
P
Pn
P0P0
0.7
100
0.6
80
60
0.4
Hyperbolic fit (R2 = 0.98)
0.3
= P/A (psi)
= P/A (MPa)
0.5
40
P
0.2
20
0.1
B
0
0.00
0.02
0.04
n = /B
0.06
0
0.08
577
18
2000
15
1500
M (k-in.)
1000
M (kN-mm)
12
6
B
500
Measured
3
Design Aids
Elastic Half Space (ASCE 41-06)
0
0.00
0.01
0.02
0.03
0
0.05
0.04
(rad)
Fig. 9. momentrotation for a 305 mm square footing model on gravel.
140
120
M (kN-m)
100
P
= 300kN
4 kips (17.8 kN)
P=
80
60
40
Measured
20
Design Aids
Ealstic Half Space (ASCE41-06)
0
0
0.005
0.01
0.015
0.02
0.025
(rad)
Fig. 10. momentrotation envelope for 1.0 m-square footing on Ticino sand, high density (HD).
K s0 B
E
2G
1 m2 1 m
14
Combining this expression with Eq. (1) and since L/B = 1 (square
footing), the following value is calculated for rotational stiffness:
Kh
K s0 B 3
B 780 kN m
4
Mp
1
0:2941 0:294x18:4kN m 1:9kN m
2
578
45
40
35
M (kN-m)
30
PP== 4100kN
kips (17.8 kN)
M
25
20
15
Measured
10
Design Aids
0
0
0.005
0.01
0.015
(rad)
0.02
0.025
0.03
Fig. 11. momentrotation envelope for 1.0 m-square footing on Ticino sand, low density (LD).
0.25
= P/A (MPa)
0.20
0.15
0.10
0.05
B
0.00
0
0.01
0.02
0.03
0.04
0.05
0.06
n=/
Fig. 12. Concentric load test of 80 mm square footing model on dense Ticino sand [26].
The measured momentrotation response envelope (i.e., connecting the tips of the hysteretic loops) is shown in Fig. 10 for the
HD soil sample and Fig. 11 for the LD soil sample.
In order to use the design aids, the stiffness to strength ratio is
rst calculated. For r0 = P0/1.0 m2 and P0 = 5P, the following is
obtained:
Prior to the application of moment, small amplitude cycles under concentric load were applied in each case. The re-loading subgrade modulus for each soil sample was reported by Allotey and
Naggar [25] as:
579
0.14
0.12
0.10
P/P 0 = 0.5
Pe/B (kN)
0.08
0.06
P/P 0 = 0.3
P
e
0.04
0.02
P/P 0 = 0.1
Measured
Design Aids
Half Space
0.00
0
0.01
0.02
0.03
0.04
0.05
0.06
(rad)
Fig. 13. Momentrotation for 80 mm2 footing model on dense Ticino sand [26].
18
2000
e/B = 1/3
15
12
1500
1000
e/B = 1/6
M (kN-mm), Pe (kN-mm)
M (k-in.), Pe (k-in)
e/B = 1/4
e/B = 1/8
e/B = 1/12
500
0
0.00
0.01
0.02
0.03
0.04
0
0.05
(rad)
Fig. 14. Momentrotation response for 305 mm square footings, constant eccentricity versus constant axial load.
Mp 120 kN m and 40 kN m
580
Kh
K s0 B 3
B 1:22 kN m
4