HPC for Material Transport in Water: From Microscopic Dynamics to Global Environmental Safety

Ravi Ayyala Somayajula_FAU - Material Transport in Water_ Simulating Temperature-Driven River Dynamics Using HPC

Sediment and material transport are fundamental drivers of river ecosystem health, coastal stability, and the structural integrity of hydraulic infrastructure. Yet, simulating these movements with high precision, especially when factoring in rapid environmental changes, has historically been a monumental computational challenge. As part of the HiDALGO2 project, the Material Transport in Water (MTW) pilot is utilising High-Performance Computing (HPC) to move beyond traditional fluid dynamics, focusing on the complex interplay between water, solid particles, and thermal energy.

Led by PhD researcher Ravi Ayyala Somayajula at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), this pilot addresses a critical gap in current modelling: the integration of scalar transport (such as temperature) into sediment simulations. It aims to support the protection of livelihoods, infrastructure, and biodiversity.

The following video, recorded at the HiDALGO2 Clustering Event in Stuttgart, details the technical framework and the transition toward exascale-ready simulations, including some impressive visualisations:

HiDALGO2 Pilot on Material Transport in Water - Ravi Ayyala Somayajula
 

Why Microscopic Understanding Matters

Traditional models often focus on the broad flow of water. However, the MTW pilot looks at the "microscopic" level to understand how individual particles interact. As Ravi explains, this level of detail is essential for several critical environmental and engineering domains that tackle global challenges:

  • Cryosphere Dynamics: In Arctic regions,  rising temperatures cause permafrost thawing, which causes the release of trapped organic matter and sediments. This can be particularly dangerous when it affects built areas. Precise simulations are required to predict how these processes destabilise soil and impact carbon cycling and support Climate Change Adaptation.
  • Infrastructure Longevity & Safety: Sediments are primary agents of erosion. By modelling how particles interact with the surfaces of dams and hydraulic structures, engineers can more accurately predict maintenance cycles and extend the lifespan of dams and hydraulic structures.
  • Protecting Ecosystems: The transport of heat and nutrients directly influences fish habitats and overall river health. Understanding how sediments move by simulating fluid-particle-scalar integration allows for a more comprehensive assessment of water quality.

A dramatic photograph of coastal permafrost erosion in an Arctic tundra landscape, showing a large block of frozen soil and ice breaking off the cliff face into the sea. A high-angle photograph showing water runoff cutting through a sandy beach slope, creating small rippled channels as it empties into a larger body of clear blue water. An underwater photograph of a crystal-clear riverbed filled with smooth, sunlit stones and rocks leading toward deeper blue water in the background.

  1. Permafrost thawing
  2. Application of AntiDunes, one of the typical examples of a free-surface simulation
  3. A typical river-bed

A key highlight of the HiDALGO2 work is the integration of temperature into sediment simulations. By studying how heat impacts water quality through "fluid-particle-scalar" integration, researchers can now predict how thermal changes—driven by industry or climate change—will affect aquatic life and water safety.

A four-panel scientific diagram illustrating computational fluid dynamics methods, featuring particle bed simulations, fluid splashes, grid boundary force vectors, and adaptive mesh refinement. 

The model is capable of simulating free surface flows, particulate flows, face field flows and also turbulent flows, as well with capabilities of doing measure refinements.

 

Working towards Exascale Computing

To achieve the necessary resolution, the team utilises a sophisticated software stack designed for the Exascale era:

  • waLBerla: A high-performance C)+ multiphysics framework capable of simulating complex particulate and turbulent flows.
  • lbmpy: An automatic code generation tool for Lattice Boltzmann kernels that ensures the software runs at peak efficiency, on both CPUs and GPUs, on any hardware.
  • Massive Scaling: The framework has already demonstrated its power by scaling across thousands of GPUs (more than 2,048 GPUs ($2^{11}$) on top-tier systems like the LUMI supercomputer.

A 3D simulation rendering of a rectangular wireframe container filled with multi-colored spherical particles of varying sizes.       A computational simulation plot displaying a uniform dark blue rectangular domain representing temperature distribution alongside a color bar legend.

  1. The riverbed simulation on which this use case was based  2. Extension towards temperature transport simulations

 

For a deep dive into the technical framework of this pilot, visit our webpage dedicated to the Material Transport in Water Pilot

 

Industrial Applications and Next Steps

The beauty of the waLBerla framework lies in its versatility. The methods from the MTW pilot can also be used in heavy industry, applying the same technology used for riverbeds.:

  1. Renewable Energy: Modelling the fluidisation of granules around suction buckets—foundations used for offshore wind turbines—to prevent structural failure due to seabed erosion.
An engineering diagram illustrating the multi-scale numerical modeling setup for a offshore wind turbine suction bucket foundation, zooming in from real-world installation to fluid-particle interaction mesh resolution.
  1. Advanced Manufacturing: Advanced Manufacturing: Predicting  laser beam welding pools in 3D, providing data that physical experiments simply cannot capture in three dimensions.

A two-panel comparative simulation diagram comparing melt pool dynamics and keyhole shapes in laser powder bed fusion under 500 W and 1,000 W laser powers, showing streamlines and boundaries modeled by KiSSAM and waLBerla..          A 3D simulation visualization showing internal fluid velocity streamlines and thermal flow patterns around a vertical laser keyhole and melt pool boundary set against a blue-to-orange gradient background.

Laser Beam Welding Application – Predictions of Melt Pool and Keyhole depths

The success of a technical pilot is measured not just by code, but by its impact on society and industry. The MTW pilot is currently focused on the Exploitation phase, which involves translating these computational capabilities into accessible tools for external stakeholders.

The MTW pilot is just one example of how HiDALGO2 is tackling global challenges. Stay updated on our progress:

About the Contributors

The core team of behild this research includes Harald Köstler and Shubham Kavane, from the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). The article is edited by Georgia Nikolakopoulou and Kyriaki Daskaloudi, from Future Needs, leading Dissemination and Exploitation in the HiDALGO2 project. The presentation video was recorded in the HiDALGO2 Clustering Event in November 2025, in Stuttgart, organised by Future Needs.