A variety of methods are employed to investigate frozen PFS, including both numerical and experimental approaches. This project aims to enhance the understanding and simulation of frozen PFS by utilizing Discrete Element Method (DEM) simulations, specifically through the application of Bonded Particle Models (BPM). One key phenomenon in frozen structures is creep behavior, and this study will leverage DEM-BPM to account for the creep behavior and breakage criteria of frozen particles.

To gain detailed insights into the internal structure and mechanical behavior of frozen PFS, it is essential to calibrate the simulation parameters and validate the results. To streamline the calibration process, Artificial Neural Networks (ANN) will be employed, facilitating faster and more accurate simulations (Figure 1). A comprehensive database is required to train the ANN effectively.

The credibility of BPM simulations is typically validated by comparing the mechanical responses observed in experiments with those predicted by simulations. Additionally, micro-CT measurements will be used to observe the internal structure of frozen PFS under various conditions, such as compression and shear. This contributes to a deeper understanding of how these materials behave under mechanical stress. Given that frozen PFS can undergo structural changes due to creep behavior, real-time experimentation is crucial. In-situ micro-CT allows for the capture of time-dependent data, providing valuable insights into the dynamic nature of these materials.