Thesis topics
AI Supported Calibration of Discrete Element Models
title
AI Supported Calibration of Discrete Element Models
institution
doctoral_school
supervisor
co-supervisor
discipline
description
One of the most widely known and commonly used tools for modeling the mechanical behavior of granular assemblies is the Discrete Element Method (DEM). The essence of this approach is the Lagrangian description of the individual motion of solid bodies that compose the entire aggregate. Through a series of short time steps, the interaction force system and the changes in position and orientation of the elements are calculated using a simulation cycle. Determining the parameters of the contact model used to calculate interaction forces and moments is a key aspect of every modeling task. In the simplest approximation, at least six individual micromechanical parameters must be identified; however, most of them can only be measured indirectly. A widely used method for determining a set of micromechanical parameters is the so-called calibration. This involves determining the real and simulated macro behavior of a specific physical phenomenon tested and modeled with the given granular material, followed by iterative adjustments of micromechanical properties until the difference between the virtual and real-world results is reduced to an acceptable level.
The aim of the research: The calibration process often requires significantly more time and effort than modeling a practical problem. However, it is almost always necessary due to the lack of reliable material libraries and databases. In this field, the development of an artificial intelligence-based or machine learning-assisted calibration algorithm would be of great significance, as it could accelerate or even automate the identification of micromechanical parameters. Therefore, the aim of this research is to develop a procedure that enables the automatic determination of the micromechanical properties of a granular assembly in a given physical state.
Current devices: Direct shear tester, Former research results
Applied technologies: Discrete Element Method (DEM), Artificial Intelligence (AI), or machine learning based, or other parameter identification techniques
The aim of the research: The calibration process often requires significantly more time and effort than modeling a practical problem. However, it is almost always necessary due to the lack of reliable material libraries and databases. In this field, the development of an artificial intelligence-based or machine learning-assisted calibration algorithm would be of great significance, as it could accelerate or even automate the identification of micromechanical parameters. Therefore, the aim of this research is to develop a procedure that enables the automatic determination of the micromechanical properties of a granular assembly in a given physical state.
Current devices: Direct shear tester, Former research results
Applied technologies: Discrete Element Method (DEM), Artificial Intelligence (AI), or machine learning based, or other parameter identification techniques
student count limit
2
location
Szombathely, Budapest
deadline
2026-05-31