Frazil ice crystal collisions

Thesis

Frazil ice crystals are small, often disk-shaped, millimetre sized crystals that form in supercooled turbulent water. Frazil ice crystals are found in rivers, lakes and the sea. Despite their small size, frazil ice crystals have significant impacts on infrastructure and the climate, in particular through the role frazil ice formation plays in the water transformation processes in coastal polynyas. The processes governing the evolution of a frazil ice crystal population are complex. One of these processes is collisions between frazil ice crystals. It is hypothesised that this process can result in crystal fracture, and is responsible for the rapid increase in the number of frazil ice crystals that has been observed in experiments. Despite its significance, many uncertainties remain about how to parametrise crystal collisions and many key physical processes are unaccounted for. In this thesis we include some of these processes into a model for frazil ice crystal collisions.

Typical models assume that the crystal collision rate can be determined by calculating the number of crystals each frazil ice crystal meets per second as it moves through a fluid. The other crystals are assumed to be stationary and point-like. We develop a model for crystal collisions that accounts for the velocity and finite size of both crystals and better accounts for the cylindrical shape of the crystals. We derive a new method to distribute fractured particles into the crystal population which better accounts for the size difference between the fractured crystal and the original crystal. We also use the lubrication approximation to calculate the hydrodynamic forces on 2D plates and 3D crystals as they approach and find a critical velocity for collision when the crystals approach close to parallel. We also obtain the conditions for fracture of 2D plates by elastic bending.

Including these non-hydrodynamic processes leads to a large increase in the collision rate and a decrease in the mean radius of the crystal population when compared to a typical frazil ice model. Hydrodynamic effects act to reduce the number of collisions. However, there are still discrepancies between our model and experimental results. We hypothesise a main cause of these differences is from neglecting flocculation and assuming that collisions always introduce crystals of the smallest size class. This suggests that future models could be improved by focussing on understanding these processes.

Authors: Rhee, D

• DOI: 10.5287/ora-exerv1n60
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: frazil ice; particle collisions; lubrication approximation; secondary nucleation
• Subjects: Ice crystals; Fluid dynamics