PSI - Issue 61

Igor Gribanov et al. / Procedia Structural Integrity 61 (2024) 89–97

90

2

I.Gribanov et al. / Structural Integrity Procedia 00 (2024) 000–000

Fig. 1. (a) ice floes on St.Lawrence river; (b) sea ice floes; (c) sea ice breakup near Villinki, Finland.

• the computational cost is reduced compared to using solid elements • the thickness is uniform or can be assumed to be constant throughout the domain

In solid ice sheets, common failure modes are crushing, spalling, bending, and splitting (Sanderson (1988)). Frac ture processes are influenced by the composition of the material, temperature, and contact areas, among other things. The scope of this work is limited to the investigation of bending and splitting modes, and assumes that the material is homogeneous and isotropic. Such assumptions are reasonable when looking at large-scale natural processes in ice covers, such as wave-induced breakup. In nature, the process of ice breakup is somewhat random and chaotic, making impossible to predict the exact path of every crack. However, knowing the distribution of stress in the material, one may predict how a crack propagates once initiated. The proposed modeling technique is based on FEM with plate elements and simulates the breakup of solid ice covers. The simplifying assumption of the approach is that cracks develop perpendicularly to the plane of the ice sheet. FEM plate approach is found in recent work on large-scale modeling of ice sheets (Seroussi et al. (2020)) and wave-ice interaction (Williams et al. (2017)). Numerical simulations of ice can be distinguished by the scale of the ice features in question, which typically vary from several centimeters to hundreds of kilometers. Larger scales address the motion of mountain glaciers and ice streams (Hruby et al. (2020)), whereas smaller scales are found in laboratory experiments and field measurements (Boroojerdi et al. (2018); Gribanov et al. (2018); Kolari (2017); Murdza et al. (2016)). This review will focus on the scales on the order of several meters, which is comparable to the sizes of ice floes. Fracture simulation is a key component of an ice model but is often the least accurate. For example, some particle based simulations produce fractures at angles that are multiples of 30 degrees due to hexagonal arrangement of parti cles. Non-realistic fracture patterns are produced due to the di ffi culty of predicting cracks’ initiation and propagation. Naturally occurring ice has random inclusions that a ff ect its strength and crack paths. In addition, a positive feedback loop exists between the location of the crack and the dynamic redistribution of stresses, which results in irregular fragment shapes. Various ways of addressing fracture modeling are discussed in the review of numerical methods in ship-ice interaction (Xue et al. (2020)). 2.1. Fracture models 2. Related work

2.2. Floe size distribution

Some research e ff orts are dedicated to the analysis of crack formation and floe size distribution. The distinguishable dominant failure modes for the sloped structure interaction are local bending failures and global splitting failures. These modes correspond to di ff erent scales of fracture – local cracks develop on a small scale in the vicinity of the load, whereas the global splits can fracture an entire floe or develop larger than the loading structure (Bhat et al. (1991); Lu et al. (2016)).