See how advanced simulation enhances nuclear safety

Learn how the UNESCO Chair in Numerical Methods spearheads frontier innovation in the Global South

Large Scale Multiphysics Computations

Kratos Multiphysics

Principal Investigator
Riccardo Rossi
Back
Overview
Research
Staff
Projects
Publications

This research group develops the Kratos framework—an open-source, general-purpose code for complex multiphysics simulations—leveraging High Performance Computing to integrate diverse physical models for realistic engineering applications and digital twins.

The Kratos Multiphysics Research Group at CIMNE focuses on developing a general-purpose research code that integrates state-of-the-art capabilities across multiple fields, explicitly designed to simulate complex multiphysics problems. This framework represents a unified computational environment where diverse physical phenomena can interact and be analysed simultaneously, providing researchers with powerful tools for tackling challenging engineering applications.

A primary focus of the group is the exploitation of High Performance Computing (HPC) capabilities for simulating realistic engineering problems. This goal is pursued through two complementary approaches: developing new solution technologies optimized for parallel architectures, and exploring the integration of models from different scientific domains, making the research inherently transversal and interdisciplinary.

The group has achieved significant advances in developing CFD models and FEM technologies, including level set-based computational fluid dynamics and fluid-structure interaction problems featuring two-phase flows. A particularly innovative area of research is the development of Simulation Based Digital Twins, where the group combines Hyperreduced-order models (HROMs) with HPC to significantly enhance the effectiveness and deployment of digital twin technology.

 

The group actively develops HPC workflows for constructing reduced-order models (ROMs), addressing the computationally intensive offline processes through parallel computing strategies. Their workflow, implemented using Kratos Multiphysics and the COMPSs parallelization framework from the Barcelona Supercomputing Center, encompasses model creation, high-fidelity simulation, reduction techniques, and feedback refinement.

Committed to open science principles, the research group fosters open-source developments and collaboration with groups across different locations and disciplines. Their work contributes to integrating diverse technologies within a single, unified workflow, continuously expanding the solution capabilities of the Kratos framework and making advanced simulation tools accessible to the broader scientific community.

Research areas

Development of CFD models and other FEM technologies, including model order Reduction

Development of new solver capabilities within Kratos, and as a tools for the development of projects. This includes in particular the improvement of the existing capabilities for the solution of “embedded” CFD problems and the development of new real-time interactive solvers based on ROM.

Uncertainty Quantification (UQ) and Optimization Under Uncertainties (OUU)

Uncertainty Quantification studies the characterization and the reduction of uncertainties in problems where some variables of the system are not exactly known. Optimization Under Uncertainties aims at solving optimization problems by considering the aforementioned uncertainties in the objective function, constraints or parameters of the problem.

Finished projects
Search

Related news

More News