ANALYSIS OF UNDERGROUND PROTECTIVE STRUCTURES

By Paul Weidlinger,1 Fellow, ASCE, and Eve Hinman,2

Associate Member, ASCE

ABSTRACT: A procedure is presented for analyzing underground protec-

tive structures subject to conventional weapons effects. A radiation
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

damping term is added to a simple mass-resistance system to include

structure-medium interaction effects. This methodology, which is based
on a decoupling procedure for finite element computations, is applied to
the deformation and the rigid body response of a single-degree-of-
freedom model. Numerical examples are given for both applications.
The deformation response is shown to be in good agreement with finite
element computations that have been previously shown to compare well
with experimental data.

INTRODUCTION
Underground reinforced concrete structures are used for military instal-
lations protected against the effects of conventional weapons. Usually
such structures are box shaped, partially or fully buried, and covered by a
concrete "burster slab" to protect against penetrating weapons (Fig. 1).
Above-ground structures serving the same purpose are often modeled
using single-degree-of-freedom (SDOF) systems. This formulation offers
an efficient method of analysis for preliminary designs, optimization
studies, and concept evaluations.
Modeling structures that are below ground in this manner is complicated
by the presence of the surrounding soil. The success of the SDOF model
depends, in part, on our ability to predict the dominant mode shape of the
response, which is not generally true for below-ground structures sub-
jected to the ground shock of an explosion. Here the response consists of
structure-medium (i.e., soil) interaction (SMI) which, by its very nature,
does not lend itself to a modal decomposition approach (Wolf 1985).
Consequently, the choice of the appropriate SDOF model and loading
function is far from obvious. Intuitive arguments regarding relevant
physical phenomena, such as dynamic arching and added mass, are used in
the literature (Biggs 1964; Design of Structures 1964). Another approach
uses a modification of the free-field stress in the soil to approximate the
interface pressure loading (Fundamentals 1984). These approximations
can be "tuned" by adjusting numerical values of model parameters, to give
a response that is consistent with experimental data points, but usefulness
for preliminary explorations is limited because they give unreliable results
in some ranges, as shown by the writers (Hinman and Weidlinger 1987).
SMI analysis is performed on mainframe computer systems by means of
large FE models that include both the structure and a portion of the
surrounding soil mass (Nelson 1978). Recently, it was shown (Wong and
'Partner, Weidlinger Assoc, 333 7th Ave., New York, NY 10001.
2
Sr. Res. Engr., Weidlinger Assoc, New York, NY 10001.
Note. Discussion open until December 1, 1988. To extend the closing date one
month, a written request must be filed with the ASCE Manager of Journals. The
manuscript for this paper was submitted for review and possible publication on
August 26, 1987. This paper is part of the Journal of Structural Engineering, Vol.
114, No. 7, July, 1988. ©ASCE, ISSN 0733-9445/88/0007-1658/$ 1.00 + $.15 per
page. Paper No. 22638.

1658

J. Struct. Eng., 1988, 114(7): 1658-1673

 Weidlinger 1983) that by decoupling the medium from the structure,
computations can be performed more efficiently without loss of accuracy.
This simplification makes FE analysis of underground structures possible
on microcomputers and PCs using existing computer codes (DAEDALUS
1986).
This decoupling concept is applied to SDOF models, to obtain both
elasto-plastic deformation response and rigid body response. This formu-
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

lation is also suitable for risk analysis by applying the methods outlined in
Benaroya and Rehak (1987). Numerical examples are given for both
applications.
The procedure is based on a number of simplifying assumptions. The
explosive is taken to be a point source, and does not directly include the
effect of charge shape and other details. AJso, the source is located at a
sufficient distance from the structure, so that cratering and penetration do
not significantly affect the response. The resistance of the model assumes
the usual elasto-plastic behavior in flexure as the principal mode of
response. Alternate failure modes are not considered. The effects of these
assumptions may have to be examined in final design computations. The
method presented is especially useful and efficient in the analysis and
evaluation of preliminary concepts and cost optimization studies.

PROBLEM FORMULATION

Fig. 1 shows, schematically, a box-shaped reinforced concrete structure

located below the ground surface. The explosive energy coupled into the
ground depends on the location of the center of gravity (e.g.) of the
explosion which may be above or below the surface depending on terminal
velocity, angle of impact (obliquity), fusing, and soil properties. Three
generic positions of the explosive source and the stand-off distance, Ra ,
are shown in Fig. 1.
The explosion produces a radially expanding compression wave that
propagates through the soil. This is called the free-field stress, P(R, t),
where R is the radial distance measured from the center of gravity of the
explosive charge and t is time measured from the time of arrival of the
stress wave.

FIG. 1. Underground Protective Structure

1659

J. Struct. Eng., 1988, 114(7): 1658-1673

 When the stress wave interacts with the structure, first the structural
components respond by deforming, then the entire box responds by
moving as rigid body. Consequently, the purpose of the SMI analysis is
two-fold. The first task is to obtain the deformation response of the
affected elements (roof, side-walls, base) to provide the moments and
forces needed for design. The second task is to evaluate the effect of the
ground shock on the contents and personnel, based on the rigid body
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

displacement (and/or rotation) response of the entire structure. In calcu-

lating the rigid body response of the structure it is assumed that the
structure has survived the explosion without significant damage.

FREE-FIELD
The simplest formulation of the free-field stress wave generated by
conventional weapons is the semi-empirical formula provided in Funda-
mentals of Protective Design for Conventional Weapons (1984: 111). It
gives the peak amplitude of the stress wave at the time of arrival as

P0 = Mpc)(Jm) (1)
where P0 = peak pressure;/ = coupling factor of the explosive energy; pc
= acoustic impedance of soil; n = attenuation coefficient; W = charge
weight, i.e., measure of energy content measured in TNT equivalents; R =
distance measured from the e.g. of the weapon; and (3 = 160 forf 0 in psi,
pcinpsi/fps, Win lbs, and/? in ft; or p = ^Vfor/VnMPa, pc inMPa-s/m,
Win Newtons, R in meters, and n = 2.75. The decay of the stress wave in
time is given by
P(R, t) = Pae~"<" (2)
where
R
(3)
'« = ? •-•••• •

is the time of arrival of the stress wave and c is the seismic velocity.
By neglecting the curvature of the wave front, we approximate the
particle velocities in the free field by the linear plane wave relation
P(R, t)
**•'> = - £ - W
Eqs. 1-4 define the free-field in a form adequate for preliminary
investigations. Values of parameters for various soils are given in Funda-
mentals of Protective Design for Conventional Weapons (1984).
The free field, along a plane surface at the box structure, (see Fig. 2), is
obtained by

v/(R, <f>, t) = ,j _' ,- [v sin2 4> + (1 - v) cos2 <J>] (5)

where Poison's ratio of the soil may be set at v = 0.3, if no other

information is available, and the angle, <j>, in terms of the standoff distance,

1660

J. Struct. Eng., 1988, 114(7): 1658-1673

Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

Surface Area
Exposed to Blast

FIG. 2. Free-Field Stress Components

1/2
tan 4> = (6)

DECOUPLING

The decoupling method, based on Wong and Weidlinger (1983) is briefly

summarized in this section.
The loading experienced by a structure can be divided into two parts: oy
= the free-field pressure at the interface; and o,- = the interface pressure,
due to the velocity difference between the structure, it(t), and medium,
v(t).
The free-field stress, ay, is defined in Eq. 5, and the interface loading is
oy = pc(v — it) (7)
where (') indicates differentiation with respect to time. Consequently, the
total interaction load is
P-, = ay + o, = Uf + pc(v — it) (8)
By substituting Eq. 4 into Eq. 8 we eliminate the velocity term, v(t):
Pi = 2af- (pc)it (9)
Because the soil at the interface with the structure is not capable of
transmitting tension, Eq. 9 is set at
Pj = 0 when 2oy < (pc)« (10)
Eqs. 9 and 10 admit the following physical interpretations:
1661

J. Struct. Eng., 1988, 114(7): 1658-1673

 1. At the time of arrival of the stress wave, at / = 0, we have it = 0, and
consequently, the loading at this time is Pj = 2oy. This is the reflected
pressure due to the acoustic doubling of the free-field stress wave at a rigid
boundary.
2. When the structure responds and u > 0, the load is diminished due to
radiation damping, by the term (pc)u.
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

3. Radiation damping tends to transfer load to the neighborhood of the

support, where at all times u = 0, and it tends to unload the structure near
the center. This is sometimes called dynamic arching.
4. The structure is entirely unloaded, when the condition of Eq. 10 is
met. This indicates that there is a gap (cavitation) between the structure
and the soil.

Comparison between complete FE computations, in which both struc-

ture and soil are modeled, and decoupled FE computations (Wong and
Weidlinger 1983) shows good agreement, especially with respect to early
time peak responses when the effects of reflections from the surface or
from discontinuities within the soil are not significant.

DEFORMATION RESPONSE

The response of a buried structure (Fig. 3A) can be obtained in sufficient

detail by finite element computations in which the structure, the surround-
ing medium and the interface are modeled in the grid using appropriate
material models. This type of computation is used if a suitable source
algorithm is available that reproduces the free-field stresses and velocities
generated by the explosion in the soil. This type of complete model is
called the soil island model (Nelson 1978). Input to this model is provided
by the free-field stresses and velocities along the boundary of the FE grid
(Fig. 3b). Recent work (Wong and Weidlinger 1983) succeeded in simpli-
fying such large scale computations by decoupling the soil island from the
structure (Fig. 3c). Both of these methods of computations have been
successfully used in the prediction of displacement and strain response of
various tests and experimental investigations. In the computations, the
reinforcing bars in the concrete components can be explicitly modeled.
These computations are useful for detailed final design of structures.
For preliminary or optimization studies of components such as slabs and
beams, simplified mass-resistance SDOF models are more suitable and the
general methodology of developing such models is well established (Fig.
3d). A brief description of such a SDOF structural model is given here.
A lumped mass SDOF model is formulated so that its displacement
response, n{t), is identical (or close) to the relevant response (such as
mid-span displacement) of the actual structure. This displacement re-
sponse can be used to derive peak moments and forces needed for design.
The SDOF system has an elastic frequency that is identical to that of the
dominant modal response elicited by the actual dynamic load on the
structure. This is achieved by transformation of the actual mass, m,
resistance, r(u), and the averaged uniform loading, P,, into the equivalent
lumped parameters of the model, mc, i\,, and Pe.
The resulting equation of motion is
1662

J. Struct. Eng., 1988, 114(7): 1658-1673

 ta,1

Component to
be Modeled
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

(b)
(d)
Pe(t)

UJ —ur—

Za-.-pCU

FIG. 3. Modeling Techniques: (a) Actual Structure; (b) FE Soil Island; (c) Decou-
pled FE Model; (d) SDOF Model; (e) Decoupled SDOF Model

m +r l
<[w) ^ <)ll=Pe(t) 00
where the subscript e implies the transformed value of the parameters,
defined in terms of the factors, KL , and, KM, as follows:
me = KMm = equivalent mass (12)
re - KLr = equivalent resistance function (13)
Pe = KLPt = equivalent loading function (14)
It is customary to refer these quantities to unit area (i.e., mass/area,
force/area, etc.) and tables are given for the equivalence factors KM and KL
1663

J. Struct. Eng., 1988, 114(7): 1658-1673

 in Biggs (1964), Fundamentals of Protective Design for Conventional
Weapons (1984), and Structures to Resist the Effects of Accidental
Explosions (1969).
For buried structures, the modeling of the SMI load, P,, presents a
special problem, as was discussed in the introduction of this paper. The
semi-empirical modification of the load and/or mass that is currently used
{Fundamentals 1984) lacks a precise, quantifiable basis for the underlying
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

simplifications. In the next segment of this paper, the decoupling method-

ology that has been applied to FE models is now developed to obtain a
rigorous definition of the stress, P,, required in a SDOF formulation.

DECOUPLED SDOF
Structure-medium interaction effects are introduced into Eq. 11 by
modifying Eq. 9 so that it is compatible with the standard SDOF formu-
lation described in Biggs (1964), Fundamentals (1984), and Structures
(1969). This is accomplished by replacing the free-field stress, a//?, 4>, t),
which is variable over the surface, by an averaged, uniformly distributed
stress, 67(f), having a peak, 67(0), such that the resulting static displace-
ment
»,,[<r/(°)] = ul<*j{R, <t>, 0)] (15)
The time history of <5y can be approximated by assuming a linear decay
with a duration
td=2ta (16)
set to preserve the impulse of Eq. 2 at an appropriate location (usually at
the quarter-span) of the structure.
Replacing the free-field stress by a uniform average value is useful
because equivalence factors tables used in the literature assume such a
distribution.
Averaging over the span is also used for the radiation damping term. In
most cases, the assumption of sinusoidal distribution in space provides an
acceptable approximation, and we use

(p7)=-(pc) (17)
IT

Both average uniform free-field and radiation damping terms are multi-
plied by appropriate equivalence factors (i.e., KL , KM) obtain the param-
eters of the equation of motion (Fig. 3<?) in the form
meii + (pc)eii + re(u)u = 2Pe(t) (18)
with the initial conditions M(0) = w(0) = 0, and where
(pl)e = KLfa) (19)
Pe = KLuf (20)
In order to satisfy Eq. 10, Eq. 18 is used only when
2Pe = (pc)cti > 0 . , , (2!)
1664

J. Struct. Eng., 1988, 114(7): 1658-1673

 At times when this term becomes negative, it is replaced by:
m/i + re{u) = 0 (22)

RIGID BODY MOTION

The pressures and velocities in Fig. 4 are shown to act at the sides of the
structure perpendicular to the direction of the blast. Due to the rapid decay
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

of the stress wave, we ignore interface shear and friction on the other sides
of the structure.
The motion is characterized by
MU = CTJVW - M O + (pc)N(VN -t/) + (pc)F(VF -U) (23)
where U is the rigid body displacement of the structure and V is the
averaged free-field velocity. The subscripts N and F identify the near (N)
and far (F) surfaces with respect to the source.
The averaged pressure functions in Eq. 23, aN(t) and crF(t), are defined
such that the peak value is given by

< W 0 ) = -;— I CTJVJK/?, 4>,t)dA • (24)

AN F
' JAN,
where a(R, §, f) is the normal free-field stress defined in Eq. 5, and A is the
surface area exposed to the blast.
Using Eq. 4 and collecting terms, we have
{H?B)U+ 2( P c)i/ = 2<TA,M (25)
where the arithmetic average of the near and far side soil impedance is

(pc) = \ [(pc)N + (pc)F] (26)

and (HpB) is the total mass per unit area exposed to the blast that includes
the mass of the entire structure and contents not mounted on shock
absorbers.

U
-® K.4

"*% «r«wt**3»8s*

FIG. 4. Rigid Body Motion

1665

J. Struct. Eng., 1988, 114(7): 1658-1673

 By Eqs. 2 and 4 we define
vN(t) = (Pc)VQe->«» (27)
where V0 is the averaged peak normal particle velocity at the near side and
the average arrival time on surface AN is given by:
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

- m m •• (28)

where

-^U: uN(R, 4>,t)dtdA

Calling the characteristic transit time across the structure:

(29)

--(g)? <*>
we arrive at the differential equation:
T 0 £ / + U= V0e~'"° (31)
with initial rest conditions, U(0) = E/(0) = 0.
Other effects, such as rotation may also be included in Eq. 31 by adding
the appropriate terms.
We note that in this formulation actual i.e., not equivalent parameters
are used. Also, because of our interest in the frequency content of the
response, an exponential decay of the pressure (Eq. 2) is used.
The solution of Eq. 31 is

U(t) = r - ^ T [ ^ ' ^ - e(-"h)] (32)

(To ~ h)
ONE-WAY SLAB RESPONSE—EXAMPLE

A one-way, symmetrically reinforced, concrete roof slab of an under-

ground structure is considered. The dimensions and reinforcing percent-
age, p4., are shown in Fig. 5.
The parameters used in this example are listed below.
Free-Field (Eq. 1)
/ = 0.4 (33a)
pc = 22 psi/fps (0.5 MPa - s/m) (33/?)
c = 1,000 fps (304.8 m/s) (33c)
n = 2.75 ; (33d)
W= 1,014 lbs (4,510.7 N) (33e)
R0 = 21 ft (6.4 m) . , . (33/)
1666

J. Struct. Eng., 1988, 114(7): 1658-1673

Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

FIG. 5. One-Way Slab Example

Model Parameters (Eqs. 12-14)

ATi = 0.52 . (34a)
KM = 0.37 (34b)
Material Properties
f[. = 6,125 psi (42.2 MPa) (35A)

fy = 72,000 psi (496.8 MPa) (35b)

where the 28-day compressive strength of concrete f'c, and the yield
strength of steel,/,,, are enhanced to account for the effects of high strain
rate induced by the explosive loading and, in the case of concrete, the
effects of aging.
The mass is
m = 1.19 x 10~ 2 lb - s2/cu in (3.23 x 10~ 2 kg/mm2) (36)
The resistance function, r(u), is obtained by constructing a uniform-
load/displacement graph for a fixed-fixed slab, consisting of three linear
segments (Fig. 6) defined by the following three points:
1. Point A—linear elastic limit.
2. Point B—plastic hinge formation at fixed ends.
3. Point C—plastic hinge formation at mid-span.
Based on Structures to Resist the Effects of Accidental Explosions
(1969), the resistance function of the SDOF model is simplified to an ideal,
1667

J. Struct. Eng., 1988, 114(7): 1658-1673

 c
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

-• u
FIG. 6. Elasto-Plastic Resistance Function

elasto-plastic response (solid line on Fig. 6) that conserves the strain

energy of the three segment response beyond point C, and is given by
r(u) = ku for ku < ru (37)
where, for the dimensions and materials of this example, we obtain the
elastic constant
k = 153 lb/cu in (0.0415 N/mm3) (38)
and
ru = 45 psi (0.31 MPa) for /<»>/-„ (39)
From these values, the system parameters of Eq. 18 are determined.
Solving Eq. 18 by standard numerical techniques, we obtain the peak
displacement response
"max = 2.9 in. (73.7 mm) (40)
corresponding to a ductility of
'ku„
V> = 9.8 (41)

These quantities define the performance of the slab when it is subjected to

the explosion.
Discussion of Results
The displacement response in Fig. 7 (shown by a solid line) is compared
with the results of a decoupled FE computation (shown by a dashed line).
The technique used in the FE computation is described in detail in Wong
and Weidlinger (1983).
1668

J. Struct. Eng., 1988, 114(7): 1658-1673

 SDOF COMPUTATION
- FE COMPUTATION (MID-SPAN)
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

u (in.)

5 10 15 20 25 30 35 40 45 50
TIME (ms)
FIG. 7. Displacement Response of Slab

SDOF COMPUTATION
FE COMPUTATION

>- c (f ps)
1000 2000 3000 4000 5000
FIG. 8. Peak Displacement versus Seismic Velocity of Fill

1669

J. Struct. Eng., 1988, 114(7): 1658-1673

 The FE computation was executed on a PRIME 2250 computer by
means of the nonlinear FE code, FLEX (Vaughn 1985) which permits
explicit modeling of reinforcing bars and uses a CAP model (DiMaggio and
Sandler 1971) for the concrete. The model is axisymmetric, having five grid
elements through the thickness and an element aspect ratio of 2:1. It is of
interest to note that execution time required for the FE model was an order
of magnitude greater than for the SDOF model.
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

To determine sensitivity of the displacement response, « m a x , to the type

of fill material used, a series of computations were performed in which the
seismic velocity, c, and the attenuation coefficient, n, (which is dependent
on the value of c) were varied.
The results of this parametric study are shown in Fig. 8 (solid line). It is
observed that significant reduction of the response, in this case, may be
achieved only by a material that has a velocity c < 1,000 fps. The response
beyond that value is not very sensitive to back-fill specifications.
Comparison of the results using the decoupled FE method, shown by a
dashed line in Fig. 8, with the SDOF decoupled method shows excellent
agreement over the full range of seismic velocities.

RIGID BODY RESPONSE—EXAMPLE

Now consider the in-structure shock acting on the buried box structure
caused by a side-on burst. The geometry and threat are shown schemat-
ically in Fig. 4.
The free-field properties are the same as in the previous example, with
the following exceptions:
/ = 1.0 (42a)
R0 = 30.0 ft (9.1 m) (42b)
The relevant geometric and material parameters are:
AN= 10,115.5 sq ft (939.7 m2) (43A)
H = 210 ft (64.0 m) (436)

9B
— = 0.25 (43c)
P
These parameters are used to solve for the horizontal rigid body
velocity, U, of Eq. 31. Then, using the methods of Timoshenko et al.
(1974), this velocity is applied as a forcing function on a linear SDOF
oscillator of variable frequency. The peak velocity response of this system,
expressed as a function of frequency, defines the internal shock environ-
ment for the structure.
Discussion of Results
The results are shown in Fig. 9 in the form of a response spectrum. This
spectrum is used to determine the shock isolation required for sensitive
equipment within the structure, such as computers or generators, by
comparing Fig. 9 with the fragility curves of the equipment obtained from
the suppliers. For preliminary studies, the generic fragility curves provided
1670

J. Struct. Eng., 1988, 114(7): 1658-1673

 100
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

1000

FREQUENCY (Hz)

FIG. 9. Shock Spectrum

in chapter 12 of Fundamentals of Protective Design for Conventional

Weapons (1984) may be used.
At the present time, there is only limited experimental verification of
shock spectra generated by conventional weapons. Until more data
become available, considerable engineering judgment is required in using
this method. In particular, the writers note that spectral values are
sensitive to the value of the p^/p ratio used.

CONCLUSION

The decoupled SDOF formulation provides an efficient method for

determining the response of buried structures subject to explosions for
optimization studies, concept evaluations and preliminary designs. Elasto-
plastic response of this model compares well with more elaborate decou-
pled FE calculations using more sophisticated material models. Appli-
cation of this method is extended to the rigid body motion of the structure
to yield response spectra used in determining shock isolation requirements
for vital equipment.

APPENDIX I. REFERENCES
Beriaroya, H., and Rehak, M. (1987). "Parametric random excitation. I: Exponen-
tially correlated parameters." J. Engrg. Mech., ASCE, 113(6), 861-879.
Biggs, J. M. (1964). Introduction to structural dynamics. McGraw-Hill Book Co.,
New York, N.Y.
DAEDALUS, database for engineers and architects to locate and utilize software.
(1986). Structural, Vol. 2, No. 2, ACEC Publications Dept., Washington, D.C.
Design of structures to resist nuclear weapons effects. (1985). ASCE Manuals and
Reports on Engineering Practice—No. 42, Revised Ed., ASCE, New York, N.Y.
Design of structures to resist nuclear weapons effects. (1964). ASCE Manuals of
Engineering Practice—No. 42, ASCE, New York, N.Y.
1671

J. Struct. Eng., 1988, 114(7): 1658-1673

 DiMaggio, F. L., and Sandler, I. S. (1971). "Material model for granular soils." J.
Engrg. Mech. Div., ASCE, 97(EM3), 935-950.
Fundamentals of protective design for conventional weapons. (1984). TM5-855-1,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, Miss., Jul.
Hinman, E. E., and Weidlinger, P. (1987). "Single degree of freedom solution of
structure-medium interaction." Proc. Int. Symp. on the Interaction of Conven-
tional Weapons with Structures, Vol 1., Mannheim West Germany, March 9-13,
Federal Minister of Defense, 5300 Bonn 1, West Germany, 44-59.
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

Nelson, I. (1978). "Numerical solution of problems involving explosive loading."

Proc. Dynamic Methods in Soil and Rock Mechanics, Karlsruhe, 5-16 Sept.,
1977, Vol. 2, A. A. Balkema Rotterdam, Holland, 239-297.
Structures to resist the effects of accidental explosions. (1969). TM5-1300, Dept. of
the Army, the Navy, and the Air Force, Washington, D.C., Jun.
Timoshenko, S., Young, D. H., and Weaver, W. Jr. (1974). Vibration problems in
engineering. 4th Ed., John Wiley & Sons, New York, N.Y.
Vaughan, D. K. (1985). FLEX user's guide, Revised Ed., Weidlinger Assoc, Palo
Alto, Calif., Mar.
Wolf, J. P. (1985). Dynamic soil structure interaction. Prentice-Hall, Inc., Engle-
wood Cliffs, N.J.
i Wong, F. S., and Weidlinger, P. (1983). "Design of underground protective
! structures." J. Struct. Engrg., ASCE, 109(8), 1972-1979.

APPENDIX II. NOTATION

The following symbols are used in this paper:

A — surface area;
c = seismic velocity;
f = coupling factor of the explosive energy;
fc = enhanced compressive strength of concrete (28 day);
fy = enhanced yield strength of steel;
H = length of box structure parallel to direction of blast;
I = total impulse acting on surface A;
KL = load equivalence factor;
KM = mass equivalence factor;
k = linear elastic stiffness;
M = mass per unit area of box structure;
m = mass per unit area;
me = equivalent mass;
n = attenuation coefficient;
P = radial free-field pressure;
Pe = equivalent uniform pressure;
Pe = equivalent uniform pressure for decoupled system;
Pi = interaction pressure;
P, = averaged uniform interaction pressure;
P0 _ peak radial free-field pressure;
R = radial distance from the e.g. of weapon;
R0 = stand-off distance;
r = resistance function;
re = equivalent resistance function;
r„ = ultimate resistance;
t = time measured from time of arrival of free-field stress;
>a = time of arrival of the stress wave;
= positive duration of blast;
u
1672

J. Struct. Eng., 1988, 114(7): 1658-1673

 'o = average time of arrival on surface A;
U = rigid body displacement of box structure;
ll — displacement response;
"max maximum displacement;
«« = static displacement;
v = averaged free-field velocity;
v0 .= peak averaged free-field velocity;
Downloaded from ascelibrary.org by Ryerson University on 05/04/17. Copyright ASCE. For personal use only; all rights reserved.

v = particle velocity of medium;

w = charge weight of weapon;
0 = constant in free-field stress equation;
M- = ductility;
v = Poison's ratio;
P = mass density of soil;
PB = mass density of box structure;
(pc) = acoustic impedance of soil;
(pc) = average acoustic impedance;
(PC), = equivalent acoustic impedance;
P.y = percentage of steel reinforcement;
= normal free-field stress;
°/
<T f = averaged normal free-field stress;
d = averaged normal free-field stress at structure interface;
°"; = interface stress;
T0 = characteristic time constant; and
4> = angle of incidence of free-field stress wave.

Subscripts
N = side of structure nearest to the blast; and
F = side of structure farthest from the blast.

1673

J. Struct. Eng., 1988, 114(7): 1658-1673