On ice mechanics in ice-induced vibrations

Abstract

The imminence of anthropogenic climate change has motivated a global energy transition towards sustainable power generation. Offshore wind—an important contributor to the energy transition—is expanding, not only in turbine size and number of installations, but also into regions with harsher environmental conditions. One of those conditions in places such as the Baltic Sea is drift ice. Offshore wind turbine support structures, with vertical sides at the waterline, must be designed to survive dynamic ice-structure interaction when ice fails in crushing against the structure. For a safe and efficient design of the support structure, dynamic ice-structure interaction resulting in ice-induced vibrations must be considered. Therefore, both an understanding of the problem and accurate modeling for the prediction of the development of ice-induced vibrations are required.

Significant progress has been made in recent years on the topic of ice-induced vibrations, and a numerical model for prediction of ice-induced vibrations has been developed based on the principles of velocity-dependent deformation and failure behavior of ice, and contact area variation between ice and structure during interaction. However, uncertainty remains regarding physical mechanisms within the ice which govern ice-induced vibrations. The ice mechanics involved in the development of ice-induced vibrations is therefore the main topic of this thesis.

The main objective was to investigate and identify the ice mechanics involved in the development of ice-induced vibrations, especially in the regime of frequency lock-in as historically defined. It was hypothesized that dynamic recrystallization played a relevant role in the ice mechanics involved in ice-induced vibrations. To test the hypothesis, ice mechanics experiments were performed at the ice laboratory specifically developed at Delft University of Technology for this purpose.

To identify grain-scale mechanisms in ice, such as dynamic recrystallization, a method was devised to elucidate ice thin section textures and (quarter) fabrics by means of crossed-polarized transmitted light and interference coloration of ice. An attempt was made to apply the method to the laboratory experiments which applied compressive loading to the edge of a thin freshwater columnar-grained ice plate, laterally confined by glass plates. Crossed-polarized transmitted light was shone through the glass plates to observe the grain structure of the ice during cyclic compression with a haversine velocity waveform. The loading and confinement scenario was intended to reproduce a vertical section of the ice edge during frequency lock-in vibrations. The experimental design demonstrated that the grain-scale mechanics of dynamic recrystallization did not obviously contribute to the peak load-velocity relation associated with frequency lock-in vibrations. As expected, fracture initiated on the grain scale was responsible for load drops. But, more interestingly, stress relaxation during periods of low relative velocity between ice and structure occurred rapidly. Following the stress relaxation, when velocity increased, the peak load was higher than previous brittle peak loads. The results indicated that the mechanisms involved in the stress relaxation were occurring on a scale smaller than the grain size. A loading path dependency was also observed with respect to the peak load-velocity relation.

Ice penetration experiments at the Aalto Ice and Wave Tank in ethanol-doped cold model ice were performed with a rigid structure, controlled oscillation, and a single-degree-of-freedom structure, and comparison of results showed that the peak global ice loads depended on the amount of time spent at low relative velocities where an ice strengthening effect developed. This has implications for the so-called velocity effect and compliance effect in design of structures subject to dynamic ice-structure interaction.

Overall, the load signals from the ice mechanics experiments on freshwater ice resembled the load signals obtained from the controlled-oscillation experiments from the model-scale ice tank tests. The qualitatively similar velocity and resulting load patterns give confidence in the idea that the mechanisms involved in both types of experiments were similar, even for different ice types and loading scenarios.

These similar results demonstrate a link in the ice mechanics across different ice types and loading scenarios, which may be explained with further research on path-dependent constitutive ice behavior, and with scrutiny regarding ice dislocation and grain boundary mechanics. Suggestions for future research are proposed, including the testing of strain rate-varying uniaxial compression of ice and ice penetration experiments with haversine velocity waveforms.