Double-scale FEM-DEM modeling of sea ice floes at m-km scales: Model Initialization and Development
Abstract
Our objective is to understand the interaction between sea ice dynamics and mechanics, and to develop models for predicting stress-strain field on an ice floe. To achieve this goal, our work aims to integrate double-scale FEM-DEM (FDEM) models and data from in-situ and satellite observations to correlate sea ice heterogeneity with its mechanical behavior and improve sea ice modeling. Our primary motivation is to develop models that can simulate the behavior of sea ice at scales of meters to kilometers across a realistic, heterogeneous ice cover. At these scales, ice structure, strength heterogeneity, and stresses imposed by wind and ocean currents at the surface and boundaries are responsible for stress propagation and crack formation through the ice cover. Due to the lack of in-situ observing methods, prior research mainly used phenomenological constitutive relations to approximate the mechanical behavior of sea ice. A limitation of phenomenological approaches is that they use tuningparameters that cannot be directly related to physical properties of ice. An alternative to the phenomenological approach is to directly simulate the material structure for input to the macro-scale constitutive models. Computational limitations generally prohibit incorporating the full microstructure computation in allbut the simplest of simulations or the smallest of scales. To address the shortcomings of previous research, our proposal aims to integrate various areas of expertise to develop more comprehensive numerical models. Specifically, we will leverage recent advances in discrete element modeling, remote sensing, sea ice dynamics modeling and observation, and machine learning and image processing tosimulate the complex interaction between sea ice structure and its macro-mechanical behavior. Moreover, our team plans to utilize observational data to improve the initialization and parameterization of sea ice models. This approach has been shown to be effectivein previous research, but its application for model parametrization remains limited. We will employ an FDEM approach for modeling to capture processes that result in ice fragmentation and the impact of having different sea ice floe sizes. Our efforts to systematically tackle this objective will rely on the following intermediate goals, which highlight the novel elements of our work: 1) Use In-situ Observations to Initialize Sea Ice Models. We will use observations of ice floe breakup recorded during the ONR-funded Sea Ice Dynamics Experiment (SIDEx) campaign to characterize statistical properties of the sea ice structure for model initialization. SIDEx collected dense stress-strain measurements on landfast and free-floating ice in the Beaufort Sea allowing mapping of stress fields over an ice floe and relating these to forcing applied on the floe; 2) Develop a Double-Scale Model for the Sea Ice Floe in an FDEM Framework. We hypothesize that if we could set a model with a realistic ice type, thickness, and strength parameterization map, and properly define boundary forces, we would be able to accurately predict stress localization and locations of failure across the ice cover; and 3) Use In-situ Observations to Assess FDEM Simulations. We will use observations of variables such as stresses, strains, and crack location, with broad spatial coverage to ascertain the likelihood of a simulated state, and its compatibility with observations. We will use the ice stress and strain obtained from the SIDEx campaign for this purpose.We anticipate the work can be extended to developing multiscale modeling approaches covering sea ice centimeters to larger scales represented in earth system models. Amulti-scale model has the advantage that macroscopic material behavior is derived directly from microstructural effects, without the need for phenomenological macroscale relations. Its proper initialization will require a comprehensive set of lab-scale experiments under different loading conditions
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 09, 2024
- Source ID
- N000142412691
Entities
People
- Ali Khosravi
Organizations
- Auburn University
- Office of Naval Research
- United States Navy