Engineering and Computational Mechanics Proceedings of the Institution of Civil Engineers

http://dx.doi.org/10.1680/jencm.17.00002
Non-linear analysis of concrete structures: Paper 1700002
Received 17/02/2017 Accepted 13/05/2017
a practical design approach to Eurocode
Keywords: codes of practice & standards/offshore engineering/
Walker, Abhyankar and Gration stress analysis

ICE Publishing: All rights reserved

Non-linear analysis of concrete structures:

a practical design approach to Eurocode
George Walker MEng, CEng, MIMechE David Gration BEng, CEng, MIMechE
Senior Engineer, Arup, London, UK (corresponding author: Director, Arup, London, UK
[email protected])
Sonja Abhyankar MEng, CEng, MICE
Senior Engineer, Arup, London, UK

Reinforced concrete is an inherently non-linear material with non-linear compressive behaviour, concrete cracking and
reinforcement yielding altering the stiffness and causing redistribution of forces within statically indeterminate structures.
A practical design approach is described that has been developed to ensure robust computation of serviceability limit
state and ultimate limit state demand in line with Eurocode BS EN 1992, including explicit modelling of material and
geometric non-linearities and construction imperfections. The approach uses the MAT_CONCRETE_EC2 material model
within the LS-DYNA finite-element analysis package and has been extensively used during the detailed engineering
design of the concrete gravity substructure for the White Rose Extension Project off the coast of Newfoundland, Canada.
The material model and validation of its behaviour in accordance with the Eurocode standard are described, along with
other typical engineering applications where the approach may be of benefit. Specific benefits and implementation
challenges encountered during the design process are also highlighted with a non-linear buckling assessment case study.

1. Introduction solver and, specifically, the MAT_CONCRETE_EC2 material

Reinforced concrete is an inherently non-linear material. model (LSTC, 2013), which has been set up in accordance
Under design loading the concrete is expected to crack, alter- with Eurocode BS EN 1992 (BSI, 2004) (henceforth referred
ing the section’s stiffness. Despite this, the conventional to as Eurocode 2). The method also includes geometric non-
method for the design of large concrete structures does not linearity in accordance with Eurocode 2, providing a complete
consider this change in stiffness during structural analysis. The solution in line with code guidance.
conventional method involves linear finite-element (FE) analy-
sis followed by a non-linear calculation to establish the LS-DYNA is a powerful analysis tool which has an increasingly
response of the cross-section. This creates an inconsistency important role in the analysis of extreme load cases – for
between the stiffness distribution assumed in the structural example, seismic loading, blast response and impact assessment
analysis and that assumed in the sectional design. This sim- (Gibson, 2012; Gogus, 2016; Hoorelbeke et al., 2006; Izatt,
plification is considered conservative in most instances and 2009; NEHRP JV, 2014). Recently, it has been shown to have a
has been widely adopted to enable an efficient design process more general application: as a practical design tool aligned
(Nyhus, 2014). with Eurocode 2. This approach has been extensively used
during the detailed engineering design of a concrete gravity
More rigorous consideration of material non-linearity is often substructure (CGS) as part of the White Rose Extension
necessary for reinforced-concrete structures under extreme load Project, Canada. The approach enables more realistic consider-
cases (e.g. seismic loading). It is also necessary in more ation of material non-linearity without the need for a manual,
standard load cases where neglecting the change in stiffness iterative, linear process. It was used to validate linear methods,
is significantly unconservative (e.g. in the estimation of service- which relied on simplified assumptions to account for material
ability limit state (SLS) displacements) or significantly overcon- non-linearity. It also allowed analysis of extreme and thermal
servative (e.g. forces induced from thermal strains). The load cases, which would typically be conducted with a separate
current design practice typically involves modification of the non-linear analysis, using a single-analysis model.
stiffness of a linear structural analysis to account for material
non-linearity. This is an iterative approach, however, and This paper describes how the non-linear design approach pro-
usually requires significant simplification of the spatial pattern vides a robust assessment of SLS and ultimate limit state
of cracking across a structure. (ULS) demands in line with Eurocode 2 (Section 3) including
a worked example in the form of a buckling assessment from
This paper describes an approach which allows explicit consider- the White Rose Extension Project (Section 4). The worked
ation of reinforced-concrete material non-linearity as part of the example includes a description of implementation challenges
structural analysis. The method uses the LS-DYNA explicit FE for the benefit of other practitioners using non-linear analysis

1
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

in accordance with Eurocode 2. The paper also highlights 3. Design approach

other typical engineering situations where this design The design approach described in this paper was developed
approach could be of value for reinforced-concrete designers during the detailed engineering design of the White Rose
(Section 5). Extension Project, a 200 000 t CGS due to be installed off
the coast of Newfoundland, Canada (Figure 1). The approach
is summarised in Figure 2. The CGS was modelled with
2. Background: material non-linearity of shell elements and linear and non-linear FE analyses were
reinforced concrete conducted in parallel. The linear analysis was used for initial
Concrete cracking, non-linear compressive behaviour and sizing calculations, following the conventional method (i.e.
reinforcement yielding are the dominant non-linear material linear FE analysis followed by non-linear assessment of the
effects in reinforced concrete. Each act to reduce the stiffness response of the cross section). The non-linear analysis provided
of a reinforced-concrete section as it is loaded. validation of simplified linear methods and insight into the
structural behaviour to enable refinement of the design.
Even in simple concrete structures, material non-linearity
is often an important consideration; conservative estimation of The same FE mesh and loading were used for both the linear
displacements in slabs relies on considering the cracked concrete and non-linear models. An automated process for rapid gener-
stiffness, for example. In large, indeterminate concrete structures ation of non-linear models from their linear equivalents was
the influence of material non-linearity is more complicated. developed, which enabled the non-linear analysis to become an
Loads will follow the stiffest structural path. Capturing the integrated part of the design process. Concrete utilisations were
reduction in stiffness due to material non-linearity is therefore calculated based on sectional forces from both models and
important to assess load redistribution between adjacent struc- iteration of the reinforcement layout and/or geometry was
tural elements. The effect of redistribution can be beneficial; undertaken where required.
a reduction in stiffness can reduce demands on heavily loaded
sections if there are sufficient underutilised sections with which The non-linear analysis captured both material and geometric
to share the load. It can also be detrimental; redistributed non-linearities in accordance with Eurocode 2. The primary
demands from adjacent elements can cause a load increase. benefit was the explicit calculation of the redistribution of
Material softening also increases displacements, which can forces and moments. In general, this reduced the peak demand
amplify second-order geometric effects. in critical areas compared with the results from the linear FE
model. It should be noted that consideration of force and
Where significant these non-linear effects must be accounted moment redistribution to enable an efficient design is particu-
for in accordance with code guidance, even when relying larly necessary for CGS structures due to stringent weight
on linear analysis. Typically, this is undertaken by modifying requirements (to enable the structure to float during installa-
the stiffness of a linear analysis, with appropriate stiffness tion) in combination with extreme loading in the offshore
reduction based on codified methods or estimations from non- environment.
linear sectional solvers (e.g. Bentz, 2000; OL, 2014). As the
reduction in stiffness affects the load distribution in the struc-
tural element, this is an iterative process and significant simpli- 3.1 Modelling material non-linearity
fication is often required. For complicated cases, the use of The MAT_CONCRETE_EC2 material model captures con-
simplified linear methods in isolation can be challenging. crete cracking in tension, crushing in compression, tension stif-
fening and compression softening as well as yield, strain
Non-linear FE analysis, which forms the basis of the design hardening and failure of the reinforcement. The material prop-
approach described in this paper, directly accounts for the soft- erties vary with temperature, enabling analysis of extreme
ening of the reinforced concrete as part of the structural analy- thermal loading (e.g. fire).
sis. This provides a less cumbersome and more accurate
assessment of the effect of material non-linearity than simpli- The concrete stress–strain behaviours in MAT_CONCRETE_
fied, linear methods. EC2 are those recommended for non-linear analysis in
Eurocode 2 (Expression 3·14). Although strain hardening can
There are a range of typical engineering situations where con- be included, the reinforcement was modelled as elastic–per-
crete material non-linearity must be considered and the non- fectly plastic in the examples discussed in this paper. The
linear design approach could be beneficial. A specific example design concrete compression curve was set up in accordance
is described in Section 4 of this paper and a range of other with Eurocode 2, Section 5·8·6, which instructs the use of
applications suited to the non-linear method are outlined in design values to scale the non-linear stress–strain curve
Section 5. (Figures 3 and 4). This approach is consistent with the ‘partial

2
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

~150 m
~120 m

Figure 1. Indicative render of the White Rose Extension Project CGS and topside

Shared inputs
Linear FE model
• Geometry
• Loading

Sectional forces
and moments
Utilisation calculation
Additional inputs

• Non-linear material
properties
Non-linear FE model
• Reinforcement
layout
• Geometric
imperfections

Figure 2. Non-linear concrete analysis design approach

safety factor’ method for non-linear analysis of concrete The non-linear LS-DYNA model was composed of multilayer
described in the fib Model Code 2010 (Walraven, 2013), in (also referred to as ‘composite’) shell elements. Each layer con-
which further guidance can be found. tained a single integration point defined as either concrete or

3
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

steel as appropriate, with the steel modelled as a layer of

smeared thickness as illustrated in Figure 5. This set-up cap-
tured the redistribution of forces through the section; as concrete
fctd cracked or reinforcement yielded, force was transferred to the
adjacent layers. It also captured redistribution across the struc-
tural elements (e.g. slabs and walls); as elements softened, forces
Tensile stress

and moments redistributed into the neighbouring elements.

3.2 Modelling geometric non-linearity

The non-linear analysis included all the second-order geo-
metric effects required by Eurocode 2, Section 5·8·6.
Crack opening strain εtmax Specifically, the explicit LS-DYNA solver simulates geometric
non-linearity (i.e. P-delta effects) and so enables buckling or
collapse behaviour to be modelled. The model was also set up
Figure 3. Tensile behaviour of concrete in MAT_CONCRETE_EC2
to include initial imperfections, scaled to match specified con-
struction tolerances or design imperfections in Eurocode 2.
All structural components were modelled using non-linear
material properties and so the analysis inherently included the
fcd softening of adjacent members, another Eurocode requirement.

3.3 Validation of the non-linear method

Compressive stress

MAT_CONCRETE_EC2 has been used on a range of projects

and extensive studies have been completed to demonstrate that
the material model simulates experimental tests and/or theoreti-
cal solutions. On the White Rose Extension Project, the per-
formance of the MAT_CONCRETE_EC2 shell elements was
compared with theoretical solutions calculated using other
accepted methods for typical thickness and reinforcement
configurations used in offshore applications. Figure 6 con-
εc1 εcu1 tains an example plot from this investigation, which shows
Compressive strain
close agreement between the capacities predicted using a
MAT_CONCRETE_EC2 non-linear shell element and a con-
Figure 4. Compressive behaviour of concrete in crete section solver. This comparison was also undertaken for
MAT_CONCRETE_EC2
in-plane shear, twisting moment and combination cases and
showed acceptable agreement.
Slab thickness

Concrete
X-direction
reinforcement
Y-direction
reinforcement
Example two-way spanning slab layout Corresponding composite shell element

Figure 5. Example representation of a two-way spanning slab as a composite MAT_CONCRETE_EC2 shell

4
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

–25 000 The CGS comprises three main sections, as shown in Figure 7.
Axial force: kN/m
The central shaft supports the topsides and contains the main
–20 000 mechanical outfitting. The central shaft and the closed caisson
cells are initially empty and are gradually flooded to lower the
–15 000
CGS onto the seabed at its final location. The open perimeter
cells provide stability during float-out and are filled with solid
ballast once the CGS is in place, to resist overturning moments
–10 000
due to wave loads. The caisson and perimeter cells are divided
into compartments, separated by radial walls. The buckling
–5000
assessment described here was undertaken for the caisson
–2000 –1500 –1000 –500 0 500 1000 1500 2000 radial walls.
0
Moment: kN/m
Linear analysis identified the governing loading condition for
5000 the caisson radial walls to be just prior to touchdown of the
Concrete section solver MAT_CONCRETE_EC2
CGS onto the seabed. At this point in the installation
sequence, the CGS has not yet been fully flooded, and so there
Figure 6. Comparison of force–moment interaction diagram exists a large differential hydrostatic pressure on the external
produced using a MAT_CONCRETE_EC2 shell element and walls of the caisson (Figure 8). The load case also contains
a non-linear concrete section solver
differential water levels between compartments as a result of
potential variation in the rate of flooding (Figure 9). The
differential compartment water levels lead to out-of-plane
deflections, which combined with the in-plane compression
4. Design example – non-linear buckling generate additional moments and necessitate a buckling assess-
assessment ment of the caisson radial walls.
The benefits and challenges associated with this design
approach are illustrated through a non-linear buckling assess- The design of the caisson radial walls was initially based on a
ment undertaken on the White Rose Extension Project. linear buckling analysis. This method relied on underestimating

Caisson Central Perimeter

compartments shaft compartments

Caisson
radial wall

Perimeter
radial wall

Figure 7. CGS structural elements relating to non-linear buckling assessment

5
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

the linear buckling method were conservative and acceptable

External water level
~120 m for the design.

The non-linear model was set up using MAT_CONCRETE_

EC2 composite shell elements, as previously described. The
model also included initial geometric imperfections. For struc-
tural elements prone to buckling, such as the caisson radial
wall, a constructed geometry similar to the shape of the
primary buckling mode was used. This imperfect geometry
was scaled to have a maximum out-of-plane deflection equal
to the specified construction tolerances.

ULS and SLS material parameters and load factors were speci-
fied for separate strength and crack width assessment. The
results quoted here concern the ULS assessment. The load
Shaft
Caisson case previously described, including differential water levels,
water level
water level was applied to both the linear and non-linear models. The
~20 m
~17 m
explicit LS-DYNA analysis was run statically; the loading was
applied as a linear ramp and held at the required load factor.

Figure 8. Elevation view of CGS water levels at touchdown

4.1 In-plane behaviour
Material non-linearity was an essential consideration for the
design of the caisson radial walls as softening of the adjacent
Caisson radial structural elements increased the in-plane compressive load
wall considered
during the touchdown load case. The non-linear model was
initially run without initial imperfections or differential water
levels between the caisson compartments to investigate the in-
plane structural behaviour in isolation.

~17 m During the touchdown load case, the perimeter radial walls
were primarily loaded in tension and thus when non-linear
~20 m material properties were considered, the perimeter radial wall
elements cracked and softened. As the tensile stiffness of the
~20 m perimeter radial walls decreased, the in-plane compression in
the caisson radial wall increased to take the hydrostatic load.
This effect was investigated in the linear analysis through alter-
ing the stiffness of the perimeter radial walls to represent
cracking. The non-linear analysis provided a more accurate
assessment of this effect, shown in Figures 10–12.

The MAT_CONCRETE_EC2 material model also explicitly

calculates the cracking pattern, shown in Figure 13 for this
load case. Cracking was seen at the location of the tensile tie
in the linear model and at positions of maximum bending in
Figure 9. Plan view of CGS water levels at touchdown the caisson radial wall. The non-linear analysis complimented
the understanding of the structural behaviour gained from the
equivalent linear analysis, and provided a more accurate esti-
mation of the governing forces.
the stiffness and restraint conditions to obtain a conservative
buckling factor. A non-linear analysis explicitly modelling
these effects was completed to (a) better understand the struc- 4.2 Non-linear buckling analysis
tural behaviour and redistribution of internal forces as the con- The non-linear buckling assessment included initial imperfec-
crete strains and (b) confirm that the assumptions used within tions and out-of-plane loading caused by differential water

6
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

Perimeter radial wall

Caisson radial wall

Linear analysis Non-linear analysis

Compression Tension Hydrostatic load

Figure 10. Elevation view representation of the in-plane behaviour of the caisson and perimeter radial walls during touchdown

MAX_PRINC_STRESS
mid-surface:
MPa
–20·00

–15·00

–10·00

–5·00

5·00

Figure 11. Mid-surface maximum principal stress plot from linear (left) and non-linear (right) analysis

levels between caisson compartments. Both the linear and non- capture material non-linearity explicitly. The non-linear analy-
linear models were subject to the same loading. Force and sis therefore included the effects accounted for in the linear
moment resultants were compared between the models to vali- buckling assessment and could be used to validate its assump-
date the assumptions used in the linear method. tions. It should be noted that as non-linear material properties
were used throughout the CGS in the non-linear analysis, it
The linear buckling model was set up with initial imperfections explicitly captured the stiffness reduction in adjacent structural
and captured the geometric non-linearity (i.e. P-delta effects). elements, which were conservatively assumed to be fully
The stiffnesses of the caisson radial wall and adjacent elements cracked in the linear buckling analysis.
were also modified to account for material non-linearity,
in accordance with the recommendations in Eurocode 2, An illustrative set of sectional resultants are presented below to
Section 5·8. highlight the benefit of the non-linear analysis. The result of
non-linearity (material and geometric) can be seen by compar-
The non-linear model included geometric non-linearity, initial ing Figures 14 and 15, which shows a significant increase in
imperfections and used the MAT_CONCRETE_EC2 to moment resultants, confirming the sensitivity of this structural

7
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

MIN_PRINC_STRESS
mid-surface:
MPa
–20·00
–15·00
–10·00
–5·00
0
5·00

Figure 12. Mid-surface minimum principal stress plot from linear (left) and non-linear (right) analysis

Uncracked
concrete Horizontal
Cracked moment resultant:
concrete kNm/m

2600
1940
1280
620
–40
–700
–1360
–2200
–2680
–3340
–4000
Figure 13. Concrete cracking pattern

Figure 14. Horizontal moment resultants from a linear model

element to these effects. Figure 16 shows the results from the (i.e. not including of geometric or material non-linearity)
non-linear analysis. The non-linear analysis predicted a smaller
increase in mid-span and edge moments due to the second-
order effects compared with the linear buckling model. This 4.3 Implementation challenges
validated the assumptions used in the linear method and This section shares some important lessons learned during the
enabled more accurate understanding of the behaviour of this implementation of the non-linear design approach used on the
structural element. White Rose Extension Project.

8
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

A key benefit of the method is that the redistribution between

adjacent structural elements is automatically captured in the
non-linear analysis model. However, due to the sequential
nature of the design process, there were instances where only a
preliminary reinforcement design was available for adjacent
elements which was insufficient to cope with redistributed
demands. The implication of automatically considering mul-
tiple structural elements is that the designer must consider the
2D moment, Mx design of more than one element at a time.
output axis global
2600 kNm/m
2200 kNm/m
A similar issue occurred for the analysis of elements where
1800 kNm/m second-order effects are important (e.g. the Caisson Radial
1300 kNm/m Wall previously described). In accordance with ISO 19903-07
870 kNm/m
430 kNm/m (ISO, 2013), elements where second-order effects were impor-
–12 kNm/m tant should be designed to a higher load factor. Hence, the
–450 kNm/m
–900 kNm/m situation can arise where the non-linear analysis predicts
–1300 kNm/m failure in parts of the structure away from the region of interest
–1800 kNm/m
–2200 kNm/m
as they have been designed for a lower load factor. This can be
–2700 kNm/m overcome by applying mean material properties to structural
–3100 kNm/m elements away from the region of interest. This prevents them
–3600 kNm/m
–4000 kNm/m from early failure, while still allowing a reasonable represen-
tation of the holistic redistribution to be captured. Distorted
failure modes such as this are a recognised problem associated
Figure 15. Horizontal moment resultants from the linear model
set-up in accordance with the linear buckling method
with the partial safety factor method for non-linear analysis,
which has been adopted here. Further guidance on the partial
safety factor method and its limitations are outlined in the fib
Model Code (Walraven, 2013; Cervenka, 2013).

Eurocode 2 does not contain detailed guidance on the use of

numerical methods. The White Rose Extension Project team
combined those experts in numerical analysis with experienced
designers to enable the approach described in this paper to be
Horizontal implemented appropriately and effectively. The approach relied
moment resultant:
on the integration of large linear and non-linear FE models
kNm/m
with a reinforcement database and a post-processor to assess
2600 utilisations and significant effort was put into automating this
1940 integration. This was a critical part of the approach becoming
a successful design tool and will enable efficient application of
1280
the method on future projects.
620
–40 5. Further use of the approach
–700 The specific demands of the White Rose Extension Project
necessitated the development of a practical design tool based
–1360
on non-linear concrete analysis. As outlined in Section 2 of
–2200 this paper, there are a range of other engineering situations
–2680 where concrete material non-linearity must be considered and
the non-linear design approach could be of benefit. Two
–3340
examples are described below in detail to explain the potential
–4000 benefit. A wider list of examples includes

& storage tanks where thermal loading is significant – for

Figure 16. Horizontal moment resultants from the non-linear
LS-DYNA model example, liquefied natural gas
& nuclear containment structures

9
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

& integral bridges specifically MAT_CONCRETE_EC2 is being increasingly

& buildings with architecturally complex geometry used (e.g. Law, 2015).
& impact-resistant slabs or decks
& bridge supports susceptible to impact loads Concrete bridge construction is another area where weight
& blast-resistant structures. control is of significance and non-linear analysis would be of
benefit. A range of load cases exist where consideration of the
Analysis of thermal load cases in general lend themselves to non-linear behaviour of concrete is beneficial, for example,
non-linear analysis. Thermal load induces strains which can thermal shrinkage and certain construction load cases (e.g.
cause cracking, relieving the original strain-induced load. Rombach, 2007). More generally, indeterminacy within con-
Consideration of linear material properties therefore may result crete bridge structures present opportunities for using non-
in significantly overconservative design. A specific example linear concrete analysis to better understand the structural be-
is found in liquefied natural gas tank design. The loading com- haviour. Web design of multi-span concrete box girders is one
prises extreme cold on the concrete outer wall following a leak example, where the non-linear design approach could provide
in the primary inner containment, as shown in Figure 17. An further structural insight or validate the design of critical sec-
analysis of the combined effect of dead load, hydrostatic tions. Although post-tensioning ensures tight control of crack-
pressure and thermal strain is therefore required to justify ing in the longitudinal direction, the transverse direction is
structural performance. The current method typically involves often allowed to crack and non-linear concrete modelling
using a non-linear section solver to calculate the stiffness could be used to assess the effect of this on torsional stiffness
reduction for a given thermally induced curvature and then or displacements, without the need for iteration.
modifying the stiffness of a linear FE model to enable assess-
ment of the additional structural loads. The use of non-linear
concrete modelling would remove the need for two separate 6. Conclusions
tools allowing more accurate calculation of the combined The benefits of non-linear FE analysis of reinforced concrete
effect over the entire structure and better understanding of the are increasingly recognised for structures under extreme
structural behaviour. Another widespread thermal example is loading. This paper presented a practical design approach
fire-induced loading in buildings, where LS-DYNA and based on non-linear FE analysis for ULS and SLS assessment
Ambient external conditions

Liquefied natural gas

Liquefied natural ~–160 C
gas leak

Figure 17. Thermal load case for liquefied natural gas tank

10
Engineering and Computational Mechanics Non-linear analysis of concrete structures:
a practical design approach to Eurocode
Walker, Abhyankar and Gration

Offprint provided courtesy of www.icevirtuallibrary.com

Author copy for personal use, not for distribution

of more standard load cases, in accordance with Eurocode 2. Gibson R (2012) Performance based seismic design of an offshore
The approach uses the explicit LS-DYNA solver and has been platform in the Caspian Sea. The Structural Engineer 90(11):
14–22.
extensively used during the detailed design of the White Rose
Gogus ABE (2016) Seattle civic square tower-engineering a high rise
CGS. The approach enables a less iterative and more realistic for increasing performance goals and seismic hazard in the
consideration of material and geometric non-linearity than Pacific Northwest. Los Angeles Tall Buildings Structural Design
many existing techniques, all within one structural model. On Council Proceedings of the 2016 Annual Meeting, Los Angeles,
the White Rose Extension Project, it has been used to validate CA, USA.
Hoorelbeke P, Izatt C, Bakke JR, Renoult J and Brewerton RW (2006)
linear structural assessments and provided insight into the
Vapor cloud explosion analysis of onshore petrochemical
structural behaviour to inform a better design. The approach facilities. Proceedings of the American Society of Safety
complements simpler techniques based on linear analysis and Engineers Middle East Chapter, 7th Professional Development
offers excellent opportunities to be used more widely as a prac- Conference & Exhibition, Kingdom of Bahrain, paper
tical design tool aligned with Eurocodes. ASSE-MEC-0306-38.
ISO (International Organization for Standardization) (2013) 19903-07:
Petroleum and natural gas industries – fixed concrete offshore
Acknowledgements structures. ISO, Geneva, Switzerland.
The authors acknowledge Nicholas Christie, Maya Shinozaki, Izatt C (2009) Assessing perforation limits of steel section impacts
Mary Ferguson, Babak Pourazarm, Chris Ariyaratana and on reinforced concrete slabs. Proceedings of the Institution of
Civil Engineers – Structures and Buildings 162(1): 37–44,
Alasdair Parkes who have contributed to the development of the
https://doi.org/10.1680/stbu.2009.162.1.37.
techniques and results presented in this paper. They also give Law AK (2015) Engineering geometrically bi-linear columns to deliver
credit to Arnstein Godejord and Richard Sturt who have con- fire resistance: standard heating. Engineering Structures 100:
tributed greatly to the development of this technique. 590–598, https://doi.org/10.1016/j.engstruct.2015.06.046.
The authors are also grateful to Ian Feltham, Gopal Srinivasan LSTC (Livermore Software Technology Corp.) (2013) LS-DYNA Keyword
User’s Manual, Version 971 R7.0. LSTC, CA, USA.
and Mohammed Ezzat, for their assistance with the examples
NEHRP JV (National Earthquake Hazards Reduction Program Consultants
section. All of those previously mentioned are from Arup. Joint Venture) (2014) NIST GCR 14-917-25: Recommendations for
seismic design of reinforced concrete wall buildings based on
REFERENCES studies of the 2010 Maule, Chile Earthquake. NEHRP JV,
Bentz E (2000) Sectional Analysis of Reinforced Concrete Members. Gaithersburg, MD, USA.
PhD thesis, Department of Civil Engineering, University of Nyhus BS (2014) Consistent practical design of concrete structures.
Toronto, Toronto, Canada. Structural Concrete 15(3): 305–316.
BSI (2004) BS EN 1992-1-1:2004, B: Eurocode 2: Design of concrete OL (Oasys Ltd) (2014) Oasys ADSEC User Manual, Version 8.3. OL,
structures part 1-1 general rules and rules for buildings. BSI, Newcastle, UK.
London, UK. Rombach G (2007) Bow-shaped segments in precast segmental bridges.
Cervenka V (2013) Global safety formats in fib Model Code 2010 for Engineering Structures 30(6): 1711–1719.
design of concrete structures. Proceedings of the 11th International Walraven J (2013) fib Model Code for Concrete Structures 2010.
Probabilistic Workshop, Brno, Czech Republic, pp. 31–40. Ernst & Sohn, Berlin, Germany.

How can you contribute?

To discuss this paper, please email up to 500 words to the
editor at [email protected]. Your contribution will be
forwarded to the author(s) for a reply and, if considered
appropriate by the editorial board, it will be published as
discussion in a future issue of the journal.
Proceedings journals rely entirely on contributions from the
civil engineering profession (and allied disciplines).
Information about how to submit your paper online
is available at www.icevirtuallibrary.com/page/authors,
where you will also find detailed author guidelines.

11