Dynamic analyses of ship impact to the new bridge over storstrømmen

Dynamic Analyses of Ship Impact to the New Bridge over
Storstrømmen
Journal: IABSE/Vancouver 2017
Manuscript ID YVR-0028-2017.R2
Theme: Performance Based Design
Date Submitted by the Author: n/a
Complete List of Authors: Egede Andersen, Jacob; COWI A/S, Bridges interrnational
Talic, Edita; COWI A/S, Bridges interrnational
Kock, Henrik; COWI A/S, Bridges interrnational
Iqbal , Muhammad ; COWI A/S, Bridges interrnational
Material and Equipment: Concrete
Type of Structure: Bridges
Other Aspects: Accidental Loads, Dynamic effects / vibrations

39th
IABSE Symposium – Engineering the Future
September 21-23 2017, Vancouver, Canada
1
Storstrømmen
Jacob Egede Andersen, Edita Talic, Henrik Bredahl Kock, Muhammad Rizwan Iqbal
COWI A/S, Copenhagen, Denmark
Contact: jca@cowi.com
Abstract
Ship impacts to bridges are relatively rare and therefore treated as accidental loads. Due to the
low probability of occurrence, it is logical to allow some degree of plastic behaviour of the
impinged structure, since the alternative, a completely elastic response, may lead to
disproportionally large material usage.
This paper presents the principle of, and results from, numerical analyses conducted for the
illustrative design of the new bridge over Storstrømmen in Denmark. This is an approximately 4km
long bridge consisting of 80m viaduct spans and two navigations spans of 160m in a single-pylon
cable-stayed configuration. The girder is a continuous, post-tensioned concrete box girder carrying
two railway tracks, two road lanes and a combined pedestrian/bicycle path.
Since ship impact is a transient event, the numerical analyses conducted consist of dynamic
analyses in the form of time-series that include relevant non-linearities of the ship, soil and bridge
bearings. Hereby a realistic picture of the bridge response during, and after, impact is obtained
allowing the comparison between pre-defined failure modes and the bridge response.
In addition, the time-series produced are used to calibrate a linear model for train
safety/runability calculations in conjunction with ship impact to define design criteria's for
maximum bridge accelerations levels at ship impact, in order to prevent trains from overturning.
The runability model itself have be tested against the Danish Great Belt West Bridge, a comparable
railway concrete girder bridge, in order to justify that the model gives correct acceleration levels
for the train/structure interaction and subsequently acceleration levels at ship impact.
Based upon the investigations made also risk analysis have been carried out, in order to show the
overall risk complies with railway authorities and Eurocode requirements.
Keywords: cable stayed bridge, ship impact, train/structure interaction, dynamic analysis, train
runability.
1 Introduction
The present paper presents the principle of, and
results from, the numerical analyses conducted
for the illustrative design of the new bridge over
Storstrømmen in Southern Denmark.
Since ship impact is a transient event, the
numerical analyses conducted consist of dynamic
analyses in the form of time-series that include
relevant non-linearities of the ship, soil and
bridge bearings.
The time-series produced are used to calibrate a
linear model for train/structure dynamic
interaction analysis in conjunction with ship
impact to define design criteria's for maximum
bridge accelerations levels at ship impact
preventing trains from overturning.
Page 1 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
2
The runability model itself have been tested on
the Danish Great Belt West Bridge, a comparable
railway concrete girder bridge with two tracks, in
order to justify that the model gives correct
acceleration levels for the train/structure
interaction and subsequently acceleration levels
at ship impact.
Based upon the investigations made risk analysis
have been carried out in order to show that the
overall risk complies with railway authorities, EU
requirements for train interoperability and
Eurocode requirements.
2 New Storstrøm Bridge
An illustrative design has as part of the tender
process for a new combined road and railway
bridge been carried out for the owners
Vejdirektoratet (Danish Road Directorate) and
Banedanmark (Rail Net Denmark).
2.1 Alignment
Storstrømsbroen shall in its entirety provide a:
• Two-lane single carriageway for road traffic.
• Double track railway line for passenger and
freight trains.
• Bidirectional pedestrian and cycle path.
The planned alignment of the 4km long crossing
can be seen in Figure 1.
Figure 1. The alignment of the new bridge. The
navigation spans are located between the red
dots.
The bridge spans are modelled as consisting of
80m viaduct spans and two navigations spans of
160m in a single-pylon cable-stayed
configuration. The girder is a continuous, post-
tensioned concrete box girder. Figure 2 shows a
typical cross section of the bridge.
Figure 2. Illustrative design, typical cross section
of the planned New Storstrøms Bridge.
2.2 Articulation
Railway expansion devices pertaining to EN
13232 [1] shall not be situated closer than 400 m
to the pylon.
The number of expansion joints/rail expansion
devices shall be four. Two at abutments and two
intermediate.
On all non-fixed pier locations there will be 2 x
bearings:
• 1 bearing free to move in all horizontal
directions
• 1 bearing restricting movement perpen-
dicular to the bridge alignment.
The pylon will have a monolithic connection to
the bridge girder.
The centre pylon constitutes a monolithic
connection. At expansion joints, each girder end
is supported by a set of two bearings.
2.3 Design speed
The bridge shall be designed for a speed limit of
200 km/h for passenger train traffic, 120 km/h for
freight train traffic and 80 km/h for road traffic.
Page 2 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
3
3 Ship Impact
The design philosophy applied to treat ship
impacts consists of a holistic approach, pivoting
around a ship risk analysis: a set of failure modes
are formulated and if the risk associated to
theses modes in conjunction with ship impacts is
acceptable, the design is acceptable.
The failure modes are formulated in terms of e.g.
displacements and accelerations with basis in the
desired behaviour of the bridge. As an example, a
relative bearing displacement limit can be
formulated that is based on a requirement that
the bearing should be able to carry vertical load
during, and after, ship impact. Similarly, if the
train is passing the bridge during, or after, a ship
impact, failure modes of the train in terms of
overturning and derailment can be defined.
Structurally, limits to the plastic behaviour are
needed to avoid global bridge failure, whilst local
damage can be accepted if the repair work is
limited in resources (time and economy).
4 Ship impact - Dynamic Analysis
The finite element model is produced in IBDAS
(Integrated Bridge Design and Analysis System),
COWI's in-house developed FE-software. The
bridge, including the substructure, is modelled by
beam elements. In Figure 3 below, an excerpt of
the finite element model of the approach spans is
shown.
The pier hit by ship impact is assumed cracked
and the Young's modulus of the concrete in this
pier is therefore reduced by a factor 2.
Figure 3. Excerpt of the FE-model used. Blue:
coordinate system of the ship impact.
4.1 Bearing Behaviour
For the dynamic analysis, a linear elastic
behaviour described above is substituted with a
non-linear behaviour of the bearings restricting
transverse movement, since these bearings are
prone to shear failure during ship impact.
Accordingly, the force-displacement curve in the
transverse direction is defined as shown in Figure
4.
Figure 4. Non-linear bearing behavior. Fshear,max is
the maximum shear force occurring in the bearing
at displacement sshear,max. The graph only shows
the force-displacement relationship for positive
values.
4.2 Foundation principle
As shown in Figure 5, the foundation rests on a
gravel pad and can therefore slide. This
horizontal sliding behaviour is modelled by an
elastic – perfectly plastic spring with cut-off
corresponding to the sliding capacity of the
foundation.
Figure 5. Foundation principle.
4.3 Ship Impact Force
The bridge substructure is subject to two types of
ship impact: Head-on Bow collisions (HOB) and
side (SW) collisions, where the vessel side drifts
into the structure sideways. Both types act on the
pier shaft. No direct impacts to the foundations
are expected, since they are fully, or partially,
Fbearing, shear
s
Fshear, max
sshear, max
Page 3 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
4
buried to avoid ship impact or located at depth
larger than the maximum ship draft.
The description of the ship impact force and its
interaction with the non-linear foundations
requires careful consideration. Attempting to
describe an explicit force-time-history for the
impact is intrinsically erroneous as it requires a
preconception of the bridge and ship response.
Instead it is chosen to apply an approach where,
to the largest extent possible, only known
parameters are described.
Figure 6. Numerical representation of the
impinging ship by means of two beam elements
and a spring (Master-slave connection) with a
non-linear force-displacement curve. Here the
force-indentation curve for HOB impacts is
shown.
The mass of the beam is determined on basis of
the mass of the impinging ship, whilst the
stiffness for head-on bow impacts is based on an
article by Pedersen [2], which proposes a quarter
sinusoidal shaped force-indentation curves based
numerical models of ships impinging on infinitely
rigid walls. The impact is produced by giving the
ship an initial velocity towards the pier.
4.4 Time Integration
The Newmark time integration applied, is defined
by a set of time integration parameters, i.e.
ߚ = 1/2, ߛ = 1/4, corresponding to the average
acceleration method. Damping is applied as
Rayleigh damping, with 2% damping of critical at
frequencies 0.1 Hz and 2 Hz. The typical time step
applied is 0.02 seconds.
4.5 Interface to train/structure interaction
analysis
The train/structure interaction analysis used for
the assessment of safe train operation is based
on the structural modes of the linear
representation of the bridge in frequency
domain. Since the numerical model for the ship
impacts of considerable magnitude is (highly)
non-linear, the two analyses cannot be conduct-
ed in the same model. Accordingly, the train
safety assessment is conducted in a separate
model and a relationship between the non-linear
and the linear model is established as illustrated
in Figure 7. The purpose of the interface is to
achieve a superstructure displacement and
rotation field experienced by the passing train
due to a ship impact similar in the linear model to
the one obtained in the non-linear model.
Figure 7. Establishment of interface at the impacted pier by decoupling the sub- and superstructure. Left:
Original configuration from the non-linear model. Middle: Completely equivalent non-linear model with
applied forces/moments corresponding to the degrees of freedom that have been decoupled.
Right: Linear decoupled system with applied forces corresponding to the degrees of freedom that have been
decoupled. Ship force, Fship, removed.
Fship,HOB
Fbow
smax
Pedersen's formula
Vy
Ms
Fship Fship
Ms
Vy
Vy
Ms
Fship = 0
Non – linear model Completely equivalent
non – linear model
Linear model representation
Page 4 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
5
5 Train/structure interaction due to
ship impact
Dynamic analysis of the train/structure
interaction has been carried out for real trains
specified in one or two tracks using FE-software
IBDAS. IBDAS allows for direct dynamic
train/bridge interaction analysis, i.e. the dynamic
trains represented by suspended and damped
masses (14 DOF per coach) are interacting
directly with the global FE model utilizing a
frequency domain approach, see Figure 8. The
behaviour of the structure is represented by a
system of modal solutions specified by a free
vibration analysis.
The dynamic analysis for derailment and
overturning of the trains due to ship impact is
verified by the traffic safety requirements in
accordance with EN-14363 [3] and EN14067-6 [4]
respectively, which are referred to in BDK1 [5].
Verification has been carried out by application
of the dynamic ship impact to different section of
the bridge including several ships with different
sizes. Selection of the ship classes for the traffic
safety verification is based on the ship sizes not
causing bearing failure.
Parameters for the real trains are used without
any partial coefficients. Similarly, the max line
speed of 200 km/h for passenger trains and 120
km/h for freight trains have been used instead of
the design speed, thus without safety factor.
The outcome has then been used as basis for ship
impact risk assessment and subsequently to
determine design basis requirement with regards
to maximum deck acceleration.
5.1 Model of dynamic (real) train
The trains are modelled by means of a 14 DOF
per coach containing two separate masses: bogie
mass and coach mass. The bogie mass contains 4
DOF: vertical and horizontal lateral translation as
well as rolling and pitching. The coach mass
contains the same two translational DOF in
addition to rolling. The contact interface between
wheels and rails is modelled such that the wheels
can move independently of the structure (track
level) in cases where contact pressure between
the wheel and rail is in ´tension´. Between the
wheel-rail contact points and the bogie mass the
primary suspension is modelled by means of a
vertical as well as lateral spring/damper system.
An identical composition exists for the secondary
suspension between the bogie and the coach
mass. All springs and dampers in the model are
characterized by constant coefficients, i.e. there
is a full linear relation between displacement
/velocity and force. The figure below provides a
diagram overview of the dynamic model for one
coach. In between the individual bogies, belong-
ing to the same coach, rolling effect between the
two masses shown is shared.
Figure 8. Definition of dynamic train vehicle in the model
The model is not capable of modelling the
twisting effect from an overturning coach on
neighbouring coach and for this reason an
overturning coach cannot benefit from stabilising
forces transferred through the connection to the
Page 5 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
6
neighbouring coach. The approach is obviously
conservative.
5.2 Real train types
The required train properties for 5 different types
of preliminary real trains assessed close to the
trains most likely to operate on the bridge are
used for the analysis;
Passenger trains:
• ICx: Intercity passenger train (Siemens ICx)
• Coradia: Alstrom Coradia (EMU – Jacobs
bogie)
• DD: Bombardier Twindexx, double decker
passenger train
Freight trains:
• FTH: Heavily loaded (EuroSprinter - ES64
locomotive with Shimmns coaches)
• FTE: Empty (EuroSprinter - ES64 locomotive
with Shimmns coaches)
All listed trains are of conventional types with
two bogies per vehicle, except for Coradia with
normal bogies at the ends and intermediate
Jacobs’s bogies. Supplementary parameter
studies of the bogie suspension systems for
freight trains has been carried out, where the
influence of the secondary as well as the primary
suspension on the train derailment and
overturning has been investigated. The dynamic
train/structure interaction verification for the
ship impact is composed of more than 1000
different cases carried out by complete and
individual computer runs.
5.3 Ship impact scenarios
Several scenarios with regards to the timing of
the passing train and the ship impacts have been
investigated, i.e. ship impacting the substructure
simultaneously as the first bogie of the train has
approached the corresponding location on the
deck level. Additionally, scenarios concerning
delaying the train relative to the ship impact as
well as delaying the ship impact relative to the
passage of the train have been investigated.
However, the first mentioned scenario has
proven to give the highest risk of the derailment
and overturning.
The analysis has been carried out for cases where
ships are colliding transversally, longitudinally, at
an angle as well as deck house collision at
superstructure. The results for safe train
operation show that highest risk of train
overturning or derailment is obtained for ships
colliding transversally. For this reason only results
for HOB: Head-on Bow will be presented.
5.4 Train overturning and derailment
As an example of overturning and derailment
results for all investigated train types for ship
class 11 impacting the pylon at stationing 1840m
is shown in Figure 9 and Figure 10, respectively.
The results are shown for the case where the
impact occurs simultaneously with the first bogie
being on the corresponding location at the deck
level.
Figure 9. Overturning ratio for passenger and
freight trains due to applied ship class 11 (79
MN). Requirement for overturning ratio of 0.9 is
shown with dashed black line.
Figure 10. Derailment ratio for passenger and
freight trains due to applied ship class 11 (79
MN). Requirement for derailment ratio of 0.8 is
shown with dashed black line.
Page 6 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
7
In general, the results for investigated passenger
trains reveal that the calculated ratios for
overturning and derailment for all ship load cases
are in compliance with the requirements.
For the investigated freight train types it is
computed that both freight train types are in
non-compliance with the requirements at
expansion joints and at main bridge for the two
largest ships investigated.
The above shown results have been used as
bases for the ship impact risk assessment
described in section 6.
5.5 Maximum Deck Acceleration
The maximum deck acceleration due to ship
impact has also been computed. An example of
the results for maximum vertical deck
accelerations for train ICx subjected to ship
classes of interest at the pylon are shown in
Figure 11. The results are shown as RMS Slow (1
sec.) values. Figure 11 shows that the calculated
maximum deck accelerations in vertical direction
exceed the requirement of 1 m/s2 according to
Danish Railway code BN1-59-4 [5]. Violation of
the requirements in vertical and transverse
direction follows similar patterns for the
remaining investigated sections of the crossing.
Figure 11. Peak deck acceleration levels RMS Slow
(1 sec.) for ICx running at speed 200 km/h in one
track (North) due to applied ship impact at the
pylon. The results are valid for HOB collision for
all ship classes.
There will be applied for a dispensation allowing
higher bridge deck accelerations during ship
impact than the current code requirement, but
still assuring safe train operation.
6 Ship collision risk analysis
The objective of the ship collision risk analysis is
to establish design requirements for ship impact
for the bridge that comply with the normative
requirements in Denmark for bridges exposed to
ship impact, i.e. Eurocode incl. Danish National
Annexes. The purpose of the risk analysis is to
identify and evaluate the risk to the bridge
structure as well as to the bridge users due to
ship-bridge impact and keep the risk within the
acceptable level by complying with the defined
risk acceptance criteria.
The approach used in the ship collision risk
analysis can be explained using the methodology
shown in Figure 12 where it can be seen that the
input data is required to perform ship collision
risk analysis which includes the following
information but not limited to:
• structure type and dimensions,
• ship traffic information including ship types,
sizes, characteristics, traffic pattern, traffic
prognosis etc.,
• type of waterway, bathymetry flood
conditions etc. and
• structure capacities (derived from the
dynamic analysis discussed in section 4)
Figure 12. Methodology- ship collision risk
analysis
This information is entered into COWI's in-house
ship collision risk model and results are derived.
Results from ship collision risk analysis are
combined with the train safety analysis
(derailment/overturning) discussed in sections 5.
Page 7 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
8
The risk level is then compared with the risk
acceptance criteria derived using the norms and
standards defined for this project and performed
the following steps:
• At locations where the risk level is lower
than the lower limit, the risk is accepted.
• At locations where the risk level is higher
than the upper limit, introduction of risk
reducing measures until the risk is within
acceptable level.
• At locations where the risk level is within
ALARP (as low as reasonably practicable)
zone, the risk is justified using the cost-
benefit assessment. At locations where the
cost to implement the risk reducing measure
is less than the achieved benefits, measure is
implemented otherwise risk is accepted.
6.1 Risk results
The final level of risk at different pier locations is
shown in Figure 13.
Figure 13. Risk results
7 Conclusions
The dynamic analyses conducted for the
illustrative design of the new bridge over
Storstrømmen have provided a basis for the
evaluation of the consequences of ship impact
that takes into account the dynamic nature of
ship impacts and the non-linear behaviour of the
ship and bridge. Hence the level of the
computations and hence detailing and utilisation
of material consumption in the complex of bridge
components is increased compared to quasi-
static analyses.
In addition, through the introduction of a force-
transfer interface, train safety in conjunction with
ship impact can be assessed in a linear model
whilst still taking relevant non-linearities into
account.
8 References
[1] EN 13232-7:2006+A1:2011 Railway applica-
tions. Track. Switches and crossings.
Crossings with moveable parts
[2] P. Pedersen, S. Valsgård, D. Olsen and S.
Spangenberg, “Ship Impacts: Bow
Collisions,” Int. J. Impact Engineering, vol.
13, no. 2, pp. 163-187, 1993.
[3] EN-14363:2016 Railway applications.
Testing and Simulation for the acceptance
of running characteristics of railway
vehicles. Running Behaviour and stationary
tests.
[4] EN14067-6:2010 Railway applications.
Aerodynamics. Requirements and test
procedures for cross wind assessment
[5] BDK1 Belastnings- og beregningsforskrift
for sporbærende broer og jord-
konstruktioner - BN1-59-4
Page 8 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

39th
1
Storstrømmen
Jacob Egede Andersen, Edita Talic, Henrik Bredahl Kock, Muhammad Rizwan Iqbal
COWI A/S, Copenhagen, Denmark
Contact: jca@cowi.com
Abstract
Ship impacts to bridges are relatively rare and therefore treated as accidental loads. Due to the
low probability of occurrence, it is logical to allow some degree of plastic behaviour of the
impinged structure, since the alternative, a completely elastic response, may lead to
disproportionally large material usage.
This paper presents the principle of, and results from, numerical analyses conducted for the
illustrative design of the new bridge over Storstrømmen in Denmark. This is an approximately 4km
long bridge consisting of 80m viaduct spans and two navigations spans of 160m in a single-pylon
cable-stayed configuration. The girder is a continuous, post-tensioned concrete box girder carrying
two railway tracks, two road lanes and a combined pedestrian/bicycle path.
Since ship impact is a transient event, the numerical analyses conducted consist of dynamic
analyses in the form of time-series that include relevant non-linearities of the ship, soil and bridge
bearings. Hereby a realistic picture of the bridge response during, and after, impact is obtained
allowing the comparison between pre-defined failure modes and the bridge response.
In addition, the time-series produced are used to calibrate a linear model for train
safety/runability calculations in conjunction with ship impact to define design criteria's for
maximum bridge accelerations levels at ship impact, in order to prevent trains from overturning.
The runability model itself have be tested against the Danish Great Belt West Bridge, a comparable
railway concrete girder bridge, in order to justify that the model gives correct acceleration levels
for the train/structure interaction and subsequently acceleration levels at ship impact.
Based upon the investigations made also risk analysis have been carried out, in order to show the
overall risk complies with railway authorities and Eurocode requirements.
Keywords: cable stayed bridge, ship impact, train/structure interaction, dynamic analysis, train
runability.
Page 9 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Dynamic analyses of ship impact to the new bridge over storstrømmen

Recommended

Recommended

More Related Content

What's hot

What's hot (6)

Similar to Dynamic analyses of ship impact to the new bridge over storstrømmen

Similar to Dynamic analyses of ship impact to the new bridge over storstrømmen (20)

Recently uploaded

Recently uploaded (20)

Dynamic analyses of ship impact to the new bridge over storstrømmen