Lattice Modelling of Concrete-Ice Abrasion

Lattice Modelling of Concrete-Ice Abrasion

Ramos, N.

Hendriks, M.A.N. (mentor)
Schlangen, H.E.J.G. (mentor)
Houben, L.J.M. (mentor)

Civil Engineering and Geosciences

Structural Engineering

Structural Mechanics

2015-10-16

An increasing number of activities are moved into Arctic regions. As oil and gas companies are exploring the options in these regions, structures are required which can withstand the severe environmental conditions. Existing structures are subjected to heavy mechanical loads due to moving ice sheets. For concrete structures in particular ice abrasion damage has been observed on the surfaces. In extreme cases, ice abrasion has caused deterioration of the entire concrete cover on offshore structures. Most research on concrete-ice abrasion is quite empirical. Little is known about the material science of concrete-ice abrasion. As the theoretical frameworks supporting the concrete-ice abrasion process are quite limited, it is very difficult to predict service life deterioration of concrete structures and to define measures against ice abrasion damage. A numerical model in which the onset of wear in the concrete-ice abrasion process is simulated, is defined in order to increase the understanding of the mechanics of the abrasion process and to define measures against ice abrasion damage. Onset of wear is de?ned as the crack initiation and propagation in the concrete material due to ice loading. Hertz contact theory which predicts excessive tensile stresses on the concrete surface due to sliding of ice asperities is used as an analytical basis for the numerical model. In order to define an appropriate numerical model for the simulation of cracking in concrete, several modelling approaches can be employed. The simulations are performed on meso-scale, which means that concrete is modelled as a three-phase material in which paste, aggregates and the interface transition zone are distinguished. This is one of the main reasons to use a lattice model for the concrete. Contact between concrete and ice is decoupled. Decoupling of the two materials is justified as long as the materials exhibit linear elastic behaviour in the contact area. On meso-scale, ice crystals fail in a brittle compressive way under the loading conditions employed in this report. This means that decoupling is justified. The loads exerted on the concrete surface by the ice crystal follow from Hertz theory and are applied externally on the concrete surface. Two numerical models are defined, namely a model with a smooth concrete surface and a model with a rough concrete surface. A rough surface profile is generated in a simplified manner by using roughness parameters found in experiments performed in the NTNU lab. Even though the rough model does not have a flat surface, the assumptions made in Hertz theory with respect to the loads in the contact area are still valid. Roughness parameters found in the experiments are quite large compared to the dimensions of the contact. Several numerical simulations are performed on the smooth model. The difference between normal and sliding contact is simulated on a homogenous concrete to make validation possible with analytical solutions proposed by Hertz. Sliding contact is applied on both a homogeneous and a heterogenous concrete to see whether this influences the results. The influence of the coefficient of friction on the onset of wear in the concrete surface is discussed by simulating sliding contact for increasing coefficients of friction. The influence of the concrete compressive strength on the initiation of damage is also simulated by using different concrete qualities. The influence of surface roughness on the onset of wear is simulated by applying sliding contact on various concrete surface profiles. Four different surface profiles have been generated and their response to the sliding ice loads is compared to the response of a smooth concrete surface. It can be concluded that a numerical model which simulates sliding contact between ice and concrete on meso-scale can be used to simulate the onset of wear during the concrete-ice abrasion process. Such model is able to capture both surface and subsurface cracking in the concrete. However, simulations show that (sliding) contact between concrete and ice is not purely a Hertzian thing. The locations at which the cracks initiate on the surface do not coincide with the location of maximum tensile stresses predicted by Hertz theory. This might be due to the loss of accuracy in decoupling the contact between concrete and ice. Nodal equivalent forces in the contact area deviate from the exact contact pressure. Also, the size of the contact area is not updated in each load step during the sequence of linear analyses. It is recommended that coupled contact is implemented in future research to increase accuracy of the results. Simulations have shown that high concrete compressive strength increase resistance against cracking due to sliding contact between ice and concrete. Also, reducing friction between these two materials and keeping the surface roughness to a minimum, increase the resistance against the onset of wear in the ice abrasion process.

concrete
ice
abrasion
micro-mechanics
numerical modelling

http://resolver.tudelft.nl/uuid:cf63370e-da29-4909-a8cd-f6da3ca82a8d

Student theses

master thesis

(c) 2015 Ramos, N.