Advanced SDOF Model For Steel Members Subject To Explosion Loading
Advanced SDOF Model For Steel Members Subject To Explosion Loading
Advanced SDOF Model For Steel Members Subject To Explosion Loading
TECHNICAL NOTE
Technical Note 10
• This document is a deliverable of the Fire and Blast Information Group (FABIG)
• FABIG would like to encourage comment and feedback from its membership. If you have any
comments on this Technical Note or any other FABIG activities please address them to the FABIG
Project Manager at The Steel Construction Institute
This publication is provided for use by FABIG members and shall not be lent, re-sold, hired out or
otherwise circulated without the prior written consent of the publishers.
Although care has been taken to ensure, to the best of our knowledge, that all data and information
contained herein are accurate to the extent that they relate to either matters of fact or accepted practice or
matters of opinion at the time of publication, the Steel Construction Institute, the authors and the reviewers
assume no responsibility for any errors in or misinterpretations of such data and/or information or any loss
or damage arising from or related to their use.
This publication is supplied to the members of the Fire and Blast Information Group
The Steel Construction Institute, Silwood Park, Ascot, Berkshire, SL5 7QN, United Kingdom
Tel: +44 (0) 1344 636525, Fax: +44 (0) 1344 6636570
Email: [email protected], Website: www.fabig.com
May 2007
An Advanced SDOF Model for Steel Members Subject to Explosion Loading: Material Rate Sensitivity
FOREWORD
This Technical Note has been prepared as one of the FABIG deliverables to FABIG members.
The single-degree-of-freedom (SDOF) model in this document extends the SDOF model previously
developed in FABIG Technical 7 to include strain rate effects and to provide more rational ductility
measures than those based on the rate-insensitive model.
This Technical Note was compiled by Dr Viken Chinien of the Steel Construction Institute. It is based on
the work that was carried out by Professor Bassam A Izzuddin, Professor of Computational Structural
Mechanics at Imperial College, under contract to the Steel Construction Institute.
Contents
FOREWORD v
GLOSSARY OF HUMAN FACTORS TERMS vii
1. INTRODUCTION 1
2. OBJECTIVE 3
3. BENEFITS FROM THE APPLICATION AND IMPLEMENTATION OF HUMAN FACTORS 5
3.1 General issues 5
3.1.1 Introduction 5
3.1.2 Compliance with standards and expectations 5
3.1.3 Benefits to safety & health 5
3.1.4 Benefits to operability 5
3.2 Cost-Benefit Analysis 6
4. APPLICATION OF HUMAN FACTORS STUDIES 7
5. TOOLS AND METHODOLOGIES 11
5.1 Currently available tools and methodologies 11
5.1.1 Allocation of Function 12
5.1.2 Hierarchical Task Analysis (HTA) 12
5.1.3 Human Error Analysis (HEA) 14
5.1.4 Human Reliability Analysis 15
5.1.5 Link Analysis 16
5.1.6 Distributed Cognition 17
5.2 Developing approaches 17
5.3 Integration of Human Factors tools and methodologies 18
CONTENTS
Page
FOREWORD iii
1. INTRODUCTION 1
2. PROBLEM CHARACTERISTICS AND ASSUMPTIONS 3
3. FORMULATION METHOD 5
3.1 Cross-sectional Response 5
3.2 Bending Stage 6
3.3 Catenary Stage 7
4. CROSS-SECTIONAL RESPONSE 9
4.1 Major-Axis Bending 9
4.2 Minor-Axis Bending 9
5. DYNAMIC STRENGTH 11
5.1 Bending Stage 11
5.2 Catenary Stage 11
5.3 Evaluation of Model Parameters 12
6. DUCTILITY MEASURES 13
7. MODEL SOLUTION PROCEDURE 15
8. EXAMPLES AND VERIFICATION 17
8.1 Cross-section Response 17
8.2 UDL Blast loading 17
9. CONCLUSION 23
REFERENCES 25
APPENDIX A NOTATION 27
APPENDIX B FIGURES 31
APPENDIX C TABLES 37
A simplified SDOF model was previously The model is now extended here to account for
formulated [1] for a steel beam under dynamic the strain-rate effect, where the following
loading, where the problem characteristics were additional assumptions are made:
as follows:
• Material strain-rate sensitivity is governed by
• The member has uniform cross-sectional the Cowper-Symonds model.
properties along its length. • The cross-sectional response under a constant
• The cross-sectional response is deformation rate is elastic/perfectly plastic,
elastic/perfectly plastic (i.e. no strain with an enhanced dynamic plastic strength
hardening). that is determined from rigid/plastic theory as
a function of the deformation rate.
• The member has two end supports where
transverse displacements are restrained • The plastic curvature rate decreases rapidly
(Figure B.1). along the member away from the point of
maximum bending moment, justifying the
• Arbitrary elastic/perfectly plastic conditions
related assumption of lumped plastic-hinge
are considered for the two end supports for
rotation.
both rotational and axial deformations
(Figure B.1). • The dynamic strength of a plastic hinge
remains constant over the response duration,
• The strain-rate effect is ignored.
and is determined from a deformation rate
• The dynamic blast loading is UDL. associated with a plastic collapse mechanism
• The initial static loading is UDL. and an average displacement rate,
accounting, where appropriate, for the
• The mass is uniformly distributed along the plastic-hinge length.
member length.
• Where the static plastic capacity of a support
• Both bending and catenary actions are is less than that of the beam, the plastic hinge
considered. is completely in the support (i.e. outside the
The following assumptions were made to beam), and the corresponding dynamic
facilitate the formulation of a relatively strength is obtained directly in terms of the
uncomplicated model which should nevertheless support deformation rate.
capture the essential problem characteristics: • The shape of the plastic bending moment
diagram is governed by the blast load
• The member response under initial static configuration.
loading is elastic.
• The location of the internal plastic hinge is Full details of the notation employed throughout
governed by the blast load configuration. this work are provided in Appendix A.
• The interaction between the plastic bending
moment and axial force is linear (Figure B.2).
3. FORMULATION METHOD
This work extends the previous SDOF model [1] consideration of the plastic stress distributions
to deal with the influence of material rate under constant κ& p and ε& cp , respectively:
sensitivity on the blast response of steel
members. With the strain-rate effect defined on ⎛ ⎛ ± yκ& p ⎞
1/ n ⎞
the material level, according to the Cowper- ∫
M d = − σ d ydA = ∫ ± σ⎜1 + ⎜⎜ ⎟
⎟
⎟ydA
⎜ ⎟
⎝ ⎝ ⎠
D
Symonds model [2], its influence is first
(3): ⎠
considered on the cross-sectional response and ⎛ ⎛ κ& ⎞ ⎞⎟
1/ n
subsequently on the overall member response. = M p ⎜1 + ⎜⎜ ⎟⎟
⎜ ⎝ Dκ ⎠ ⎟⎠
⎝
3.1 Cross-sectional Response
⎛ ⎛ ε& cp ⎞
1/ n ⎞
The material response under a constant strain Fd = ∫ σ d dA = σ y ⎜1 + ⎜⎜
∫ ⎟
⎟
⎟dA
⎜ ⎟
rate ( ε& ) can be idealised, in the absence of strain ⎝ ⎝ D ⎠ ⎠
(4):
hardening, as elastic/perfectly plastic with an ⎛ ⎛ ε& cp ⎞
1/ n ⎞
enhanced dynamic yield strength (σd) = F p ⎜1 + ⎜⎜ ⎟
⎟
⎟
⎜ ⎟
(Figure B.3). According to the Cowper-Symonds ⎝ ⎝ D ⎠ ⎠
model:
where κ& p and ε& cp are considered in absolute
⎛ 1n ⎞ value, Mp and Fp are the static plastic bending
⎛ ε& ⎞
(1): σd = σ y ⎜ 1 + ⎜ ⎟ ⎟ moment and axial force capacities, respectively,
⎜ ⎝D⎠ ⎟
⎝ ⎠ and Dκ is a rate-sensitivity parameter for plastic
bending that depends on the cross-section shape,
where n and D are material rate-sensitivity
as determined in Section 4.
parameters, and σ y is the static yield strength.
For mild steel (n = 5) and (D = 40 sec–1), though It is noted that (3) assumes that the position of
the extended SDOF model is formulated for the neutral axis is not influenced by material rate
generic n and D parameters, and is therefore also sensitivity, which is valid for bending about an
applicable to other materials provided (n ≥ 4). axis of symmetry. Although for asymmetric
cross-sections the position of the neutral axis can
The cross-sectional response under a constant vary with the plastic curvature rate, the influence
deformation rate can also be idealised as of such variation on dynamic magnification of
elastic/perfectly plastic. The enhanced dynamic the plastic moment capacity can be realistically
plastic strength can be obtained from first neglected, and hence (3) is also applicable to
relating the material strain rates to the moderately asymmetric cross-sections.
generalised cross-section plastic strain rates
(Figure B.4): Finally, the interaction between the dynamic
plastic moment and axial force is realistically
(2): ε& = ε& p = ε& cp − y κ& p assumed to have the same shape as the static
where is the centroidal axial plastic strain rate, interaction curve (Figure B.5), where the
κ& p is the plastic curvature rate, and y is the transition between full plastic bending and
plastic axial resistance is governed by:
distance of a material fibre from the reference
centroidal line. The material plastic strain rate ⎛ ε& cp < rp κ& p ⇒ M = M d
( ε& p ) is identical to ε& in the plastic range, since (5): ⎜
⎜ ε& cp = rp κ& p ⇒ M + rp F = M d
material strain hardening is ignored. ⎝
with,
The dynamic plastic moment and axial force Mp
capacities can then be determined from (6): rp =
Fp
where ( D 1Δ , n 1f ) and ( D rΔ , n rf ) represent axial response characteristics for the left and right
rate-sensitivity parameters for the left and right supports, and hence axial plasticity is equally
supports, respectively. shared between the two supports. Furthermore,
as for bending plasticity, the above expressions
With Δ& tp , θ& 1p , θ& p and θ& rp related to the reference assume that the localisation of axial plasticity is
related to the static rather than the dynamic
displacement rate ( v& m ), depending on the
relative axial capacities of the supports and the
plastic collapse mechanism, the dynamic plastic
beam. This is considered to be a realistic
axial capacity can be obtained from one of the
assumption, which avoids major complications
applicable equations in (15), as presented for the
without significant compromise to accuracy in
individual blast loading configurations in m
Sections 5 to 7. the determination of Fd .
4. CROSS-SECTIONAL RESPONSE
According to [3], the determination of the dynamic where de represents an effective distance from the
plastic bending moment capacity requires a rate- neutral axis.
sensitivity parameter (D κ) that depends on the
cross-sectional shape and the axis of bending. In Two cross-sectional configurations are
general, this can be related to the material rate- considered for the enhanced SDOF model, as
sensitivity parameter (D) by: depicted in Figure B.7, with bending about both
the major and minor axes considered for the
D
(16): D κ = I-section.
de
The same expressions (17) and (18) can also be used for the out-of-plane one-way bending of a
corrugated wall section (Figure B.7b).
For a symmetric I-section subject to major-axis bending, the above expression for de simplifies to:
n
d ⎡ w ⎤
(19): d e = w ⎢1 − (2 n + 1) (w + 2) ⎥ such a cross-section being characterised with ( δ = 0 ).
2 ⎣ ⎦
∏
2j
= =
j=1 2 j + 1 (2 n + 1)!
The new rate-sensitive SDOF model estimates according to Section {5, 6 or 7}, as
the dynamic strength based on a constant appropriate to the blast load configuration.
displacement rate, which can be taken as the iii. Apply the previously developed SDOF model
average or maximum value from dynamic [1] with the static strengths
analysis. For typical rate-sensitive materials ( M 1p , M p , M rp , F pm ) replaced by the dynamic
based on the Cowper-Symonds model [2], the values ( M 1d , M d , M rd , Fdm ). Determine all
order of the strain rate is more significant than its
parameters of the SDOF model using the new
exact value in determining the dynamic strength. dynamic strengths, except for the plastic
This has a similar implication on the member
interaction radii ( r p1 , r p , r pr ), which remain
level, where considering a steel beam with
( n = 5 ) subject to PT load, increasing the unchanged.
m iv. Perform dynamic analysis according to the
displacement rate ( v& ) by 50% increases the
solution procedure described previously [1].
over-strength by only 7%. Therefore, if the
m v. Estimate the new displacement rate ( v& m ),
dynamic strength at v& is, say, 1.25 times the
and repeat from step (ii) if it is significantly
static strength, the dynamic strength at 1.5 v& m is different from the previously assumed value.
changed to only 1.267 times that the static With regard to step (iv), it was previously
capacity, an overall change of only 1.4%. Given suggested that conservation of kinetic energy
this fact, it is only necessary to estimate the may be used to adjust the velocity at the
displacement rate, either from an expectation of transition between two adjacent resistance
the maximum displacement over the duration of stages. This normally leads to over-conservative
the blast, or from a prior rate-insensitive results, and should therefore be used only as a
analysis. means for bracketing the maximum displacement
response. The alternative approach is to consider
The following procedure describes the steps the reference velocity to be continuous over the
involved in the application of the proposed full range of the response, and is the one used in
rate-sensitive SDOF model: the verification examples of the following
section.
i. Estimate the displacement rate ( v& m ), either
from a prior rate-insensitive analysis or from
a rule of thumb (e.g. v& m = L/[20 × (blast
duration)])).
ii. Determine the dynamic bending strengths
( M 1d , M d , M rd , Fdm ) corresponding to v& m ,
The enhancement of the previous SDOF model 8.2 UDL Blast loading
[1] to account for material rate-sensitivity is
Two examples are provided here to verify the
verified here through comparisons with the
rate-sensitive dynamic response for UDL blast
nonlinear finite element analysis program
loading, where comparisons are made against
ADAPTIC [3]. With the details of determining
ADAPTIC [3]. The UB of Section 10.1 is used,
the SDOF model characteristics in terms of the
and the member length L is taken as 5 m. The
static bending and axial strengths discussed in
beam is assumed to have a uniformly distributed
previous Technical Note [1], focus is given here
mass with a total value mL of 104 kg, and is
to the determination of the dynamic bending and
subjected to a triangular loading pulse with a
axial strengths, which then simply replace the
rise time and duration of 10 msec and 100 msec,
corresponding static strengths in the original
respectively, where the peak value of the total
SDOF models. Several dynamic analysis
load Pmax is varied for the particular problem.
examples are provided for a beam under
Two sets of support conditions, used previously
different boundary conditions and subject to
for verifying the original rate-insensitive model
UDL blast loading.
[1], are considered hereafter. It is noted that in
8.1 Cross-section Response all cases, no adjustment is made to the reference
velocity at the transition between different
In the following examples, a Grade 50 resistance stages (Section 9).
UB 356 × 171 × 57 is used for which the
following properties apply: Support conditions: set (3)
This case ignores the catenary stage. The
Flanges :172.1 × 13mm 2 , Web :332.6 × 8 mm 2 support boundary conditions are as follows:
EI = 3.3223 × 107 N.m 2 , EA = 1.4984 × 109 N
K lm = 7.9734 × 107 N.m , K m
r
= 1.9934 × 107 N.m
M p = 3.5303 × 10 N.m,
5
Fp = 2.5331 × 10 N
6
M lp = 1.7652 × 105 N.m, M pr = 3.5303 × 105 N.m
−1
D = 40sec , n=5
leading to the intermediate parameters:
Considering bending about the major axis, and
k lm = 12 , k rm = 3
taking dw as the centre-to-centre distance
between the flanges (dw = 03456 m), de and Dκ Considering first the rate-insensitive response,
are determined from (19) and (16) as: generic bending case B2 is applicable, where the
resulting response characteristics are shown in
d e = 0.1555 m, D κ = 257.2 m−1s ec−1 Table 8.1.
Table 8.1 Rate-insensitive response parameters for set (3) support conditions
Elastic Elasto-plastic(1) Elasto-plastic(2) Plastic
k ( N / m) 5.1030 × 10 7
2.9160 × 10 7
4.4374 × 10 6 0
R max ( N) 4.2364 × 10 5 8.9435 × 105 9.8849 × 10 5 9.8849 × 10 5
vm
max ( m) 8.3017 × 10 −3 2.4444 × 10 −2 4.5660 × 10 −2 -
Under a peak total load Pmax of 2 × 106 N, the Re-applying the SDOF model with
rate-insensitive response is depicted in ( M 1p , M p , M rp ) replaced by ( M 1d , M d , M rd ),
Figure B.8, where very good comparison is generic bending case B2 is still applicable, and
observed against the results of ADAPTIC. The the rate-sensitive response characteristics
average displacement rate up to the peak obtained are shown in Table 8.2.
deflection is around 4 m/sec, but allowing for a
reduction in the response due to high strain-rate, Using the modified SDOF model characteristics,
a reduced value ( v& m = 2 m/sec) is assumed next the predicted rate-sensitive response is shown in
for the rate-sensitive response. Figure B.8, where excellent comparison is
obtained against the results of ADAPTIC, with
With ( M 1p < M p ), the left plastic hinge is in the rate sensitivity shown to lead to over 50%
support, and hence the corresponding reduction in the maximum achieved
rate-sensitivity parameters are assumed to be displacement. Favourable comparison is also
given as: demonstrated for the reactions in Figure B.9, with
the small discrepancies attributed to high
Dlθ = 53.58sec−1 , n lm = 6 frequency components that are not typically
reflected by a SDOF model.
Considering Table C.3 in Appendix C for the
remaining plastic hinges, the following The plastic strains are evaluated for the plastic
parameters are obtained: hinges within the member, at midspan and at the
right end, according to Table C.7 in Appendix C,
α = 0.3694 where the following intermediate parameters are
m lp = 0.5, m pr = 1.0 obtained:
Dθ = 360.0sec −1 , Dθr = 28.91sec−1
M y = 3.1323 × 105 N.m, y = 0.1793m
n m = 5.5, r
nm =6
m ld = 0.8432, md = 1.5481
which in combination with Table C.4 and
m dr = 1.7469, mdt = 2.8070
Table C.1 in Appendix C, leads to the dynamic
bending strengths: In the range of displacements relevant to this
Table 8.2 Rate-sensitive response parameters for set (3) support conditions
Elastic Elasto-plastic(1) Elasto-plastic(2) Plastic
k ( N / m) 5.1030 × 107
2.9160 × 10 7
4.4374 × 10 6 0
R max ( N) 6.3386 × 105 1.2457 × 106 1.4249 × 106 1.4249 × 106
vm
max ( m) 1.2421 × 10−2 3.3405 × 10−2 7.3776 × 10−2 -
The following table compares the maximum leads to lower strain concentrations. It should
plastic strains from the new model with those also be noted that much lower plastic strains
predicted by ADAPTIC at a specific would be achieved if a small amount of material
displacement (vm = 0.1768 m): strain hardening is included, though this was
outside the scope of the current SDOF model
Internal Right plastic development. Considering a bi-linear stress-
plastic hinge hinge
Plastic strain strain model for steel with a post-yield strain-
(SDOF model) hardening slope of 2% of the elastic modulus
0.03436 0.20187
εp , εpr (E), much reduced plastic strains are achieved
Plastic strain
with ADAPTIC at a maximum displacement
(ADAPTIC) (vm = 0.1483 m); these strains compare as
0.03146 0.14466
εp , εpr follows with the predictions of the SDOF model:
Δ = D Δ = 50 m / sec, n f = n f = 5
Dm
Clearly, therefore, a more realistic assessment of l m l
plastic strains with the SDOF model requires its
further extension to account for material strain Noting that the bending-rate parameters for the
hardening. support plastic hinges are not required, since
their static bending strengths are zero, the
Support conditions: set (4) parameters associated with the internal plastic
This case accounts for the catenary response, hinge are obtained from Table C.3 in
where the following rotational and axial support Appendix C as:
stiffnesses are assumed:
α = 0.3694
K lm = K rm = 0 N.m , M lp = M rp = 0 N.m m lp = 0, m rp = 0
K lf = 1.4984 × 108 N / m , K fr = 1.4984 × 108 N / m D θ = 475.0sec −1 , n m = 5.5
Fpl = 1.2666 × 106 N , Fpr = 2.5331 × 106 N which in combination with Table C.4, C.1 and
C.2 (Appendix C) leads to the following
dynamic strengths:
leading to the following intermediate
parameters:
θ& p = 3. 2 sec −1 M1d = M rd = 0 N. m
k lm = k m
r
=0 , rpl = rpr = 0 m Δ& tp = 1. 1815m / sec M d = 4. 9526 × 10 5 N. m
rp = rpt = 0.13937 m , K fe = 5.9937 × 107 N / m Δ& mp = 0. 7355m / sec Fdm = 1. 8112 × 10 6 N
catenary stages with the nonlinear catenary The resulting SDOF response characteristics are
model, are obtained as shown in Table 8.4. shown in Table 8.5.
The beam is analysed with the SDOF model, for Using the modified SDOF model characteristics,
a peak load Pmax of 2 × 106 N, where the the predicted rate-sensitive response is shown in
predicted rate-insensitive response is compared Figure B.10, where good comparison is obtained
favourably to the ADAPTIC results in against the results of ADAPTIC, again
Figure B.10. The average displacement rate up demonstrating the significance of the strain-rate
to the peak deflection is around 6 m / sec, but effect. Favourable comparison is also shown for
allowing for a reduction in the response due to the reactions in Figure B.11, where the small
high strain-rate, a reduced value ( v& m = 4 m/sec) discrepancies are again attributed to high
is assumed for the rate-sensitive response. frequency components that are not typically
reflected by a SDOF model.
With ( Fpl < Fpr = Fp ), full axial plasticity is
concentrated in the left support, for which the
rate-sensitivity parameters are assumed to be
given, leading to the following axial parameters
according to Table C.2 in Appendix C:
Table 8.4 Rate-insensitive response parameters for set (4) support conditions
Table 8.5 Rate sensitive response parameters for set (4) support conditions
Elastic Plastic Catenary (1) Catenary (2)
k (N / m)
2. 0412 × 10 7 0 3.8360 × 10 7 ( v m − 0.13937 ) 2 2.8979 × 106
k s (N / m)
R max ( N ) 7.9242 × 105 7.9242 × 105 1.5889 × 106 -
vm
max ( m ) 3.8822 × 10−2 0.13937 0.41423 -
9. CONCLUSION
This Technical Note presents the second-stage minor axis, and for a corrugated wall section.
extension of a recently developed SDOF model Subsequently, the implications of material rate
[1] for steel members subject to explosion sensitivity are addressed on a member level,
loading, which was shown to provide significant where the dynamic strengths are established for
improvements over other existing models. The both the bending and catenary stages in terms of
proposed model [1] considered i) general the reference displacement rate.
support conditions both in terms of flexibility
and strength, and ii) catenary action resulting The application of the extended SDOF model is
from axial restraint at the supports, under illustrated through an example, where
uniformly distributed (UDL), blast loading. comparison is made against the predictions of
However, the model neglected the effect of the nonlinear finite element analysis program
material rate-sensitivity, which is very important ADAPTIC [3]. In general, it is shown that the
for steel members subject to blast loading. The proposed model provides very good accuracy in
aim of this second-stage of model extension is to comparison with ADAPTIC, and demonstrates
address this shortcoming, and to provide more that the strain-rate effect for steel members
rational ductility measures than provided with subject to blast can reduce the maximum
the previous rate-insensitive model. deflection by well over 50%. Importantly, the
model enhancement and accuracy are achieved
This Technical Note presents an overview of the through a reasonably simple formulation, which
problem characteristics, and outlines the is very well suited for practical application.
formulation method employed in developing the
advanced SDOF model for dealing with material Finally, through the major enhancements
rate sensitivity. Firstly, the strain-rate effect is incorporated in the proposed SDOF model, a
dealt with on a cross-sectional level, where more realistic assessment of steel members
relevant parameters are established for a general subject to explosion loading is now possible
I-section, subject to bending about the major or using simplified analysis, and accounting for the
important effect of the strain-rate.
REFERENCES
[2] Interim Guidance Notes for the design and protection of topside structures against explosion and
fire, SCI, 1992.
[3] B.A. Izzuddin, 1991, ‘Nonlinear Dynamic Analysis of Framed Structures’, PhD Thesis,
Department of Civil Engineering, Imperial College, University of London.
APPENDIX A NOTATION
A fb : Bottom flange area for I-section. Δt : Total extension over the member
and supports.
A ft : Top flange area for I-section. Elastic axial rigidity of beam.
EA :
Aw : Web area for I-section. EI : Elastic bending rigidity of beam.
dm
p :
Catenary extension entity. ε rp : Extreme fibre plastic strain at right
support.
dw : Web depth (centre-to-centre of F: Axial force.
flanges).
Fd : Dynamic plastic axial force
D: Material rate parameter. capacity of beam.
Dκ : Cross-sectional bending rate Overall dynamic plastic axial
parameter. Fdm :
capacity.
DlΔ : Axial rate parameter for left Fp : Static plastic axial force capacity
support plastic hinge. of beam.
Dm Overall member axial rate Static plastic axial force capacity
Δ : Fpl :
parameter. of left support (≤ Fp).
D rΔ : Axial rate parameter for right Overall static plastic axial capacity
support plastic hinge. Fpm :
(minimum of Fpl , Fp and Fpr ).
Dθ : Bending member rate parameter
for internal plastic hinge. Fpr : Static plastic axial force capacity
Bending rate parameter for left of right support (≤ Fp ).
Dlθ :
plastic hinge. Internal plastic hinge length.
hp :
Dθr : Bending rate parameter for right
plastic hinge. h lp : Plastic hinge length at left support.
Δp : Plastic extension at internal hinge.
Plastic hinge length at right
h rp :
support.
Δlp : Plastic extension at left hinge.
k: Stiffness of piecewise linear
Δ& m
p : Local axial extension rate. response segment.
M 0r : Right support bending moment due θ& rp : Plastic rotation rate at right hinge.
to initial static loading.
APPENDIX B FIGURES
K lf , Fpl K fr , Fpr
K lm , M lp K rm , M rp
EA, Fp
EI, M p
L/2 L/2
M lp , M p , M rp
Fpl , Fp , Fpr F
Figure B.2 Plastic interaction between axial force and bending moment
σ σd
dynamic yield strength
σd overstress
( σd − σ y )
σy σy
static yield strength
ε ε
(a) Stress/strain relationship (b) Overstress/strain-rate relationship
σd σ y
κp
y
ε cp
σ y σd
M
εcp < rp κp
Md
Mp εcp = rp κp
Mp
rp =
Fp
Fp F
Figure B.5 Rate-sensitive interaction between plastic bending moment and axial force
θlp θ rp
θp
(a) Plastic-hinge rotation rates
Md
Mp
M rp < M p
h lp
hp
κ lp
b ft
A ft
Aw
A ft Aw
dw dw A fb
A fb
b bf
0.50
ADAPTIC: no strain-rate SDOF model: no strain-rate
0.30
0.20
0.10
0.00
0.00 0.05 0.10 0.15 0.20
Time (sec)
Figure B.8 Dynamic response for UDL blast loading: set (3) support conditions
1000
ADAPTIC: left SDOF model: left
800 ADAPTIC: right SDOF model: right
600
Reaction (kN)
400
200
0
0.00 0.05 0.10 0.15 0.20
-200
Time (sec)
-400
-600
Figure B.9 Rate-sensitive reactions for UDL blast loading: set (3) support conditions
0.50
0.40
0.30
0.20
0.10
0.00
0.00 0.05 0.10 0.15 0.20
Time (sec)
Figure B.10 Dynamic response for UDL blast loading: set (4) support conditions
800
600
400
Reaction (kN)
200
Time (sec)
0
0.00 0.05 0.10 0.15 0.20
-200
ADAPTIC: left/right
-400
SDOF model: left/right
-600
-800
Figure B.11 Rate-sensitive reactions for UDL blast loading: set (4) support conditions
APPENDIX C TABLES
⎛ ⎛ &l 1/ n lm
⎞ ⎞ ⎛ ⎛ θ& 1/ n m ⎞ ⎛ ⎛ &r ⎞
1/ n rm ⎞
⎜ ⎜ θp ⎟ ⎞ ⎜ ⎜ θp ⎟
⎟ ⎜
Mp 1 + ⎜
p
⎟ ⎟ ⎟
M lp ⎜1 + ⎜ l ⎟ ⎟ ⎜ ⎜D ⎟ ⎟
M rp ⎜1 + ⎜ r ⎟ ⎟
⎜ ⎝ Dθ ⎠ ⎟ ⎜ ⎝ θ ⎠ ⎟ ⎜ ⎝ Dθ ⎠ ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠
Rate parameter
DlΔ 2 DlΔ D rΔ DL
Dm
Δ
Power parameter
n lf n lf n fr n
n fm
⎛ ⎛ &m 1/ n fm
⎞ ⎞
Axial strength ⎜ Δp ⎟
Fpm ⎜ 1 + ⎜ m ⎟ ⎟
⎜ ⎜⎝ D Δ ⎟
Fdm
⎠ ⎟
⎝ ⎠
Rate parameter Dκ L 2 α Dκ L Dκ L
Dlθ , Dθ , Dθr (n + 1) (
1.172 m rp + 6.828 ) m lp + m pr + 2 + 2 (m lp + 1) (m pr + 1) (
(n + 1) 1.172 mlp + 6.828 )
Power parameter 1
n +1 n+ n +1
n lm , n m , n rm 2
Left hinge rotation Internal hinge rotation Right hinge rotation Axial deformation
θ& lp θ& p θ& rp Δ& pt
⎛ d m ⎞ 4 v& m
2 v& m 4 v& m 2 v& m ⎜ rpt + p ⎟
L L L ⎜ 2 ⎟ L
⎝ ⎠
Table C.5 Effect of initial static loads on end moments and reaction forces
Entity Initial bending moment Initial bending moment Initial reaction Initial reaction
Static load M l0 M 0r V0l V0r
P0 L ⎡ k lm (k rm + 6) ⎤ P0 L ⎡ k rm (k lm + 6) ⎤ P0 ⎡ k lm k rm + 5 k lm + 3 k rm + 12 ⎤ P0 ⎡ k lm k rm + 3 k lm + 5 k rm + 12 ⎤
UDL ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
12 ⎢⎣ k lm k rm + 4 k lm + 4 k rm + 12 ⎥⎦ 12 ⎢⎣ k lm k rm + 4 k lm + 4 k rm + 12 ⎥⎦ 2 ⎢⎣ k lm k rm + 4 k lm + 4 k rm + 12 ⎥⎦ 2 ⎢⎣ k lm k rm + 4 k lm + 4 k rm + 12 ⎥⎦
Table C.6 Effect of initial static loads on midspan bending moment and reference displacement (UDL)
P0 L ⎡ k lm k rm + 6 k lm + 6 k rm + 36 ⎤ P0 L3 ⎡ k lm k rm + 8 k lm + 8 k rm + 60 ⎤
UDL ⎢ ⎥ ⎢ ⎥
24 ⎢⎣ k lm k rm + 4 k lm + 4 k rm + 12 ⎥⎦ 384 EI ⎢⎣ k lm k rm + 4 k m + 4 k rm + 12 ⎥⎦
l
Range v m Left plastic hinge Internal plastic hinge Right plastic hinge
⎧⎪ ⎫⎪ 2 v m ⎧⎪ 7 k rm + 48 r ⎫⎪ 2 v m
Plastic-hinge rotation
θlp , θp , θpr [v m ,ep( 2)
max →[
L
⎨6 M ps −
7 k lm + 48 l
M ps + M rps ⎬ +
L
{− 10 M ps + M lps + M rps + } 4 vm L
⎨6 M ps + M ps −
l
M ps ⎬ +
48 EI ⎪⎩ k lm ⎪⎭ L 24 EI L 48 EI ⎪⎩ k rm ⎪⎭ L
Static condition [v m ,ep( 2)
max →[ M lp = M p - M rp = M p
rp ( v m + v 0m ) 2 2 rp ( v m + v 0m ) 2 rp ( v m + v 0m ) 2
m ,ep ( 2)
[ v max → rpt − v 0m ]
Plastic-hinge extension rpt L rpt L rpt L
Δ lp , Δ p , Δ pr
rp (2 v m + 2v 0m − rpt ) 2 rp (2 v m + 2 v 0m − rpt ) rp (2 v m + 2v 0m − rpt )
[rpt − v 0m → [
L L L
Static condition [v m ,ep( 2)
max →[ M lp = M p & Fpl > Fpm Fp > Fpm M rp = M p & Fpr > Fpm