1

Crashworthiness of the Douro Waterway Lock Gates

And Potential Improvements

André S. G. T. Farinha

andre.t.farinha@tecnico.ulisboa.pt

Mechanical Engineering Department

Instituto Superior Técnico, Universidade de Lisboa, Portugal

November, 2016

ABSTRACT

The increasing traffic in the Douro waterway and the prospect of reopening the Moncorvo iron mines

have settled the conditions for starting the preliminary works on the modernisation of the waterway.

Since the construction of the Douro river lock gates, the mass of the ships navigating the waterway have

increased from a maximum of 2200 tons to 2900 tons, so the evaluation of the lock gates’

crashworthiness to ship impacts needs to be updated to the current navigation conditions. A methodology

to solve the impact problem is firstly defined in terms of simulation type, geometry definition, impact

location, material model and contact model. This methodology is then applied to the Carrapatelo

upstream lock gate and using the Abaqus/Explicit solver, finite element dynamics simulations were

performed to evaluate the response of the lock gate to impacts by two different ships at variable water

levels. The greatest risk associated with the impacts is pondered to be the possibility of the gate falling

into the lock chamber, or that its deformation is enough for a large mass of water to fall into the lock

chamber. In order to evaluate this, a series of criteria that quantify the damage delivered to the lock gate

and its supports are defined. The obtained results show that the damage in the lock gate is not sufficient

for the occurrence of a hazard with possible loss of life. However, for one of the impact cases, the

reaction force on the chains that connect the gate to the electromechanical actuators was equivalent to

3.6 times the expected load from the weight of the gate, which is believed to be dangerous for the

operation of the lock. Finally, the possibility of using an energy dissipating mechanism to reduce the

damage in the lock gate is evaluated, and the results show good improvement on that course.

Keywords: Crashworthiness, Metal structures, Lock gate, Finite elements, Fracture modelling

1 INTRODUCTION

The lock is an important infrastructure for inland water navigation. Due to high flow rates, without human

intervention, inland waterways (rivers, lakes, canals, et cetera) are at best seasonally navigable for

commercial purposes. By constructing dams the flow rate can be controlled, leading to safer navigation.

This creates yet another problem which is that ships need to navigate through the water head created by

the dam. In order to permit this, a chamber with changeable water depth, occupying a small portion of the

dam, raises and lowers the ship inside it (Figure 1.1). This procedure places the ship at the same water

level as the portion of the river it needs to travel to, allowing its safe passage.

In a simplistic form, a lock is composed by a chamber (where the water level can be varied), one

upstream and one downstream gate and two valves. In advance to a vessel approaching a lock, the water

level in the chamber should be the same as the water level where the vessel is. For a vessel traveling

downstream, the upstream gate is then opened and once the vessel is inside the chamber, it is closed. By

actuating a valve, the water in the chamber flows through gravity to the downstream part of the waterway,

up to the point that the water level in the chamber is the same as downstream. The downstream gate may

then be opened so that the vessel may proceed.

2

Figure 1.1: Longitudinal cut of simplistic navigation lock system with a ship travelling downstream.

After the water level in the chamber is equalised with the downstream level, the downstream gate is

pulled up and the ship passes bellow the concrete wall.

There are three main types of lock gates used in commercial navigation: the sliding gates, the vertical

gates and the mitre gates which appear to be the preferred option in most of the waterways. These gates

are quite well described by (Daniel, 2000), we are interested in particular in the vertical gates as the lock

gates in the Carrapatelo lock are of this type. The upstream gate of the vertical drop type (Figure 1.2), the

gate is lowered when being opened, while the downstream one is of the vertical lift type (Figure 1.3) and

so the gate is lifted when being opened.

Figure 1.2: Upstream lock gate in closed

position.

Figure 1.3: Downstream lock gate in open

position.

We have witnessed in recent years a steady increase in tourism related traffic in the Douro waterway and

a renewed interest in reactivating the Moncorvo mines. This has created the necessary conditions so that

the first steps for the modernisation of the Douro waterway may be taken. As part of this modernisation

project, impact safety in the lock gates present along the waterway is being reviewed, since previous

studies regarding impact safety on the Douro lock gates date from 1970s. In these studies the maximum

mass of the considered ships was 2200 tons (Mota, 2012), but current directions impose that the vessels

navigating in the Douro waterway should not exceed a mass of 2500 tons (Administração dos Portos do

Douro, 2016), limit which is expected to be risen to 2900 tons in the near future. This escalation on the

mass of the ships navigating the Douro river urges that new studies should be conducted with the purpose

3

of certificating the lock gates for the new impact conditions. Besides that, current computational methods

can be useful in performing more detailed studies than the ones conducted in the 1970s.

The aim of this work is therefore to evaluate the performance of the existing lock gates on the Douro river

when impacted by recreational ships that currently navigate the river and by the ships expected to

transport the ore from the Moncorvo mines. In order to achieve this, a methodology to assess the

crashworthiness of the lock gates is developed, which is then employed, using the software

Abaqus/Explicit solver to evaluate the response of the Carrapatelo upstream lock gate to the impact of two

different ships. Besides that, some guidelines towards the improvement of the crashworthiness of these

structures are suggested.

1.1 Current Methods to Solve Ship Impacts on Lock Gates

Ship collision on lock gates is addressed in design codes such as the Engineering Manual EM 1110-2-

2703 that suggests empirical approaches where the dynamic load is converted to equivalent hydrostatic

loads and maximum allowable stresses are defined for different load cases. The German standards DIN

19704 and 19703 besides advising similar approaches, proposes the usage of impact energy dissipating

mechanisms with a minimum energy dissipation capacity of 1 MJ.

One possible option to simulate ship impacts on lock gates is to use an analytical method, which relies on

analytical solutions for the deformation of beams and shells to calculate the deformation of the structure.

These analytical models were created based on similar models used to simulate ship to ship impacts, in

fact, there is an extensive literature, when it comes to simulating ship to ship collisions, for example (Paik

and Pedersen, 1996), (Zhang, 1999), (Hong and Amdahl, 2008), (Tabri et al., 2009), (Liu and Amdahl,

2010), (Pedersen, 2010). These models where first applied to impacts on lock gates by (Le Sourne et al.,

2002) and have since been further developed: (Buldgen et al., 2012) and (Buldgen, 2014). Their work

reveal that it is a reliable and computationally fast tool, very useful during the pre-design stage of a gate,

but for more advanced design stages, other tools that can produce results with greater detail should be

used.

The finite element method is a very flexible tool when calculating the deformation of structures and can

be used to simulate ship impacts on lock gates, being the most reliable options quasi static-analysis and

dynamic analysis since the effect of the ship geometry is considered in both. Both simulation methods can

also take into account material nonlinearity and nonlinearities related to the large deformations on the

structure. The first is a non-linear static analysis where the ship acts as an indenter, the deformation

energy of the structure can be used to calculate the initial kinetic energy of the ship. This option is used

for example by (Gernay and Rigo, 2010) to enhance the crashworthiness of a lock gate design and by

(AbuBakar and Dow, 2013) for the crashworthiness assessment of different ship hull structures under

grounding damage. A dynamic analysis takes into account the dynamic effects of the impact and delivers

more detailed results such as the time evolution of the impact force, stresses and strains, displacements

and energy in the system. This simulation type is used by (Le Sourne et al., 2002) for the validation of the

developed analytical method and by (Hogström and Ringsberg, 2013) for the crashworthiness assessment

of different ship hull structures under different impact scenarios.

2 METHODS

This methodology will follow the general guidelines in (Ehlers et al., 2014), first a simulation method is

chosen to evaluate the ship impacts, then the geometry of the gate and impacting ships is defined,

followed by more specific properties of the model and the definition of the impact characteristics, at last

the criteria for the allowable level of damage are defined. Using the data from the simulations, the

damage on the lock gates is then evaluated in face of the defined criteria and simulations are created for

the chosen solutions.

For the purpose of obtaining more detailed results from the simulations, dynamic finite element analyses

are used. The software Abaqus/Explicit is used to create the models and the geometry is discretised using

the 2 dimensional S4R elements with 5 integration points through the thickness. The gate itself has a span

of 13 meters and 7 meters of height, divided into 5 elements (Figure 2.1) connected by a non-structural

sealant at the skin plate and by pin plates that transfer the vertical load between the diaphragms. The

mesh used consisted in quadrangular shell elements with the dimensions of the edges ranging between 50

mm and 30 mm.

4

Figure 2.1: Upstream lock gate front view.

As shown in Figure 2.2 each element is composed of a skin plate, horizontal girders, vertical girders and

diaphragms; the weight of the gate is supported by 2 chains connected to the top element diaphragms and

the hydrostatic load is transferred to the lock’s concrete structure through a series of rollers, also

connected to the diaphragms.

Figure 2.2: Structural elements that compose the Carrapatelo upstream lock gate.

The lowest 2 elements were not modelled in the finite element analysis to save computation time, the

diagonals where also excluded as their purpose is mostly to resist torsion and the rollers were introduced

as boundary conditions in the holes where they are connected. A full symmetry boundary condition was

introduced at the centre as it would not allow the folding of the central vertical girder.

Two ship hull geometries were used for the impact analysis, one is a cruise ship that currently navigates

the river and the other a pre-design geometry of a ship that will possibly transport the iron ore from

Moncorvo. As the reinforcing structure of the hulls was not available, the ships are modelled as rigid;

besides that, all the impacts considered occur at the centre of the gate because the width of the ships is

almost as great as the lock chamber’s. By aligning the normal water level in the gate with the ships’

draught line, the two impact cases shown in Figure 2.3 assume the normal water retention level in the

dam. The first case consists in the iron Ore cargo ship impacting at the Normal water retention Level

(ONL case) while the second consists in the Cruise ship impacting at the Normal water retention Level

(CNL case).

Figure 2.3: Impact cases with the normal water retention level in the dam for the iron ore cargo ship (a)

and cruise ship (b).

5

Impacts with higher retention level were considered because the water level may vary over the year, as

shown in Figure 2.4 two impacts are considered for the cruise ship and another for the cargo ship. The

iron Ore cargo ship impacts at the Highest water retention Level (OHL case), next the Cruise ship impacts

at the Highest water retention Level (CHL case) and finally the Cruise ship impacts at an Intermediate

water retention Level (CIL case).

Figure 2.4: Impact cases with higher than normal water retention level for the iron ore cargo ship (a) and

cruise ship (b) and (c), being (c) the intermediate water retention level.

The material used for the construction of the Carrapatelo lock gate observes the S355 – JR standard,

which is discriminated in the standard EN 10025-2:2004, European standard for hot-rolled structural

steel. The steel was modelled with an initial elastic range, an isotropic strain hardening range and a

damage evolution range, as shown in Figure 2.5. The ductile damage initiation criteria was used, with no

dependence on the strain rate or stress triaxiality, the damage evolution law is of the linear type and an

element for which the damage evolution function reaches one is eroded. The numerical values for Figure

2.5 are presented in Table 2.1.

Figure 2.5: True stress – true strain curve used to describe the construction steel for the Carrapatelo lock

gate.

Table 2.1: Material properties used to model the steel in the lock gate.

𝐸𝑌 𝜎𝑦 𝐸𝑇 𝜎𝑢 𝜀�̅� �̅�𝑓𝑝𝑙

L

Young

Modulus

Yield

Stress

Tangent

Modulus

Ultimate

Stress

Strain at

damage

initiation

Plastic

displacement

at failure

Reference

length

200 GPa 345 MPa 1.25 GPa 470 MPa 12 % 5.5 mm 30 mm

3 RESULTS

The main concern with the ship impacts on the lock gate is that the gate may be knocked out of its

supports and fall, along with all the water mass, into the lock chamber. This may happen by excessive

deformation of the gate, excessive damage to the rollers, or excessive damage to the hoisting chains. In

order to evaluate this, the reaction forces on the rollers and the hoisting chains are evaluated. Another

concern is that a considerably large breach may be opened in the skin plate, leading to a big mass of water

falling down into the lock chamber. This will be evaluated by tracking the occurrence of fracture and

plasticity in the skin plate.

For the purpose of evaluating the crashworthiness of the gate, the ratio between the energy dissipated

through irreversible deformations (plasticity and fracture) and the total energy dissipated in the impact is

evaluated. The ship intrusion depth and impact force evolution during the impact are also assessed and

6

correlated with the occurrence of local instabilities in the structure. Furthermore, the energy dissipated by

different elements of the lock gate structure is evaluated in order to get a feeling of the distribution of

loading through the entire structure.

A shown in Table 3.1 the impact cases that resulted in higher damage to the lock gate (ONL and CNL

cases) also lead to higher indentation, while the ones where the damage to the lock gate was less, lead to

higher impact forces, higher damage to the supports and to the skin plate.

Table 3.1: Values of maximum indentation in the lock gate, impact force, fracture energy, permanent

damage ratio, load transmitted to the rollers and damaged skin plate area obtained for the 5 considered

impact cases, where FE amount of energy dissipated as fracture during the impact and DR% is the

percentage of the ship’s initial kinetic energy dissipated through permanent deformation.

Impact Case ONL OHL CIL CHL CNL

FE 34.16 0.12 13.36 21.37 46.69

DR% 61.8 % 28.6 % 43.8 % 50.9 % 62.3 %

Maximum Indentation 276 mm 60 mm 81 mm 105 mm 255 mm

Impact Force 2.468 MN 10.54 MN 5.189 MN 4.247 MN 3.479 MN

Rollers Load 0.53 MN 1.37 MN 0.86 MN 0.79 MN 0.16 MN

Damaged Plate Area 0.44 m2 0.75 m

2 0.58 m

2 0.78 m

2 0.64 m

2

As shown in Figure 3.1 the behaviour of the structure is highly non-linear for the ONL and CNL impact

cases but more predictable for the other cases. For the ONL case it is visible how the folding of the plate

in the top triangular girder lead to a stagnation of the impact resisting force while the indentation

increases, similarly, for the CNL case the occurrence of fractures in the vertical girders leads to a

decrease in force.

Figure 3.1: Impact force as function of the indentation on the gate for the 5 impact cases. Details

showing local effects for the ONL and CNL impact cases.

The reaction force on the hoisting chains revealed to be significant only for the CNL impact case, the

value obtained was 1.25 MN, which corresponds to 3.6 times the load obtained from the static weight of

the lock gate.

7

In order to reduce the damage to the lock gate, one possible solution can be increasing its stiffness,

however the stiffened structure will lead to higher loadings on the supports, whose failure will lead to

disastrous consequences. Another option that can improve the gate’s response to an impact is to use an

energy dissipater, the ONL impact case was used to test the performance of such an option by attaching

such a device to its skin plate. This device was modelled by a cylinder built with tailored material

properties so that the behaviour of the cylinder would be equivalent to the energy dissipater studied by

(Xudong et al., 2012).

As shown in Figure 3.2 and Figure 3.3 the permanent damage is reduced drastically, this is achieved

because the damper dissipated 150 kJ, which corresponds to 42 % of the ship’s initial kinetic energy. This

is however much less than the 1 MJ that this damper was designed to dissipate, which is to be expected as

the test conditions do not take into account a deformable gate in series with the damper.

Figure 3.2: Deformed configuration for the

ONL impact case with a damper; in red the

areas where plasticity occurs.

Figure 3.3: In red the areas where plasticity

occurs for the ONL impact case.

4 CONCLUSIONS AND FUTURE WORK

The objective of this work was to evaluate the crashworthiness of the lock gates present along the Douro

river, taking into account its current navigation characteristics and near future developments.

The results showed that the damage in the lock gate is not sufficient for the occurrence of a hazard with

possible loss of life. However, for the case of the CNL impact, the reaction force on the hoisting chains

was equivalent to 3.6 times the expected load from the weight of the gate and may thus lead to its

fracture, obviously information on the chain itself will be necessary to actually evaluate if fracture occurs.

The choice of impact cases revealed to be quite varied and thus providing a set of different results that

provided some insight on how the impact location and ship geometry influence some of the damage

parameters. Among the parameters that vary with the impact cases, the impact location and ship hull

geometry are the ones that have greater influence on the results. This is visible in the CNL impact case

that although having considerably less kinetic energy than the ONL case and different ship hull geometry,

achieved similar values for maximum impact force and indentation.

The usage of an energy dissipating mechanism attached to the lock gate revealed to be quite effective in

reducing the damage in the lock gate consecutive to a ship impact and a better choice when compared to

the option of increasing the structure rigidity.

The methodology employed for these simulations should be altered if large scale simulation will be

carried out for all the lock gates in the river, quasi-static analysis is a much less time consuming approach

that will take into account important factors such as ship hull geometry and impact location.

Perhaps the most urgent work to be carried out on this subject is that information on the hoisting chains

and supporting rollers should be collected in order to evaluate if the forces transferred to it are dangerous

for the operation of the lock gate.

8

REFERENCES

ABUBAKAR, A. & DOW, R. S. 2013. Simulation of ship grounding damage using the finite element

method. International Journal of Solids and Structures, 50, 623-636.

ADMINISTRAÇÃO DOS PORTOS DO DOURO, L. E. V. D. C., SA. 2016. Navegação Comercial

[Online]. Available: http://douro.apdl.pt/index.php/pagina-exemplo/navegacao-comercial/

[Accessed 2nd July 2016].

BULDGEN, L. 2014. Simplified analytical methods for the crashworthiness and the seismic design of

lock gates. PhD, University of Liège.

BULDGEN, L., LE SOURNE & H. RIGO, P. 2012. Simplified Analytical Method for Estimating the

Resistance of Lock Gates to Ship Impacts. Journal of Applied Mathematics, 2012, 39.

DANIEL, R. A. 2000. Mitre gates in some recent lock projects in the Netherlands.

EHLERS, S., LE SOURNE, H., BULDGEN, L., OLLERO, J., ROBERTSON, C., BELA, A. & RIGO, P.

2014. Ship Collision On Lock Gates – Technical Solutions and Simulation Approaches. The 7th

International Conference on Thin-Walled Structures. Busan, Korea.

GERNAY, T. & RIGO, P. 2010. Analysis of ship impact on lock gates - Seine-Escaut Est Waterway.

Mathematical Modelling in Civil Engineering, 6.

HOGSTRÖM, P. & RINGSBERG, J. W. 2013. Assessment of the crashworthiness of a selection of

innovative ship structures. Ocean Engineering, 59, 58-72.

HONG, L. & AMDAHL, J. 2008. Crushing resistance of web girders in ship collision and grounding.

Marine Structures, 21, 374-401.

LE SOURNE, H., RODET, J. C. & CLANET, C. 2002. Crashworthiness analysis of a lock gate impacted

by two different river ships. International Journal of Crashworthiness, 7, 371-396.

LIU, Z. & AMDAHL, J. 2010. A new formulation of the impact mechanics of ship collisions and its

application to a ship–iceberg collision. Marine Structures, 23, 360-384.

MOTA, O. 2012. NAVEGAÇÃO NO RIO DOURO E TRANSPORTE FLUVIAL DO MINÉRIO DE

MONCORVO. Fórum Empresarial da Economia do Mar.

PAIK, J. K. & PEDERSEN, P. T. 1996. Modelling of the internal mechanics in ship collisions. Ocean

Engineering, 23, 107-142.

PEDERSEN, P. T. 2010. Review and application of ship collision and grounding analysis procedures.

Marine Structures, 23, 241-262.

TABRI, K., BROEKHUIJSEN, J., MATUSIAK, J. & VARSTA, P. 2009. Analytical modelling of ship

collision based on full-scale experiments. Marine Structures, 22, 42-61.

XUDONG, X., BASANT, K. P., Z., A. K. & RNORM, D. 2012. Impact analysis of an innovative shock

energy absorber and its applications in improving railroad safety 12th

International LS-DYNA

Users Conference. Washington.

ZHANG, S. 1999. The mechanics of ship collisions, Institute of Naval Architecture and Offshore

Engineering.