In multi-phase flows the interactions between phases further increase the already stringent requirements on methods and software capabilities for flow simulations.
The computational elements have to resolve a multitude of scales, handle the physics imposed by the different phases and capture their interfaces.
By their adaptivity, accuracy and robustness, particle methods provide desirable methods to address these issues.
In this project we will develop multi-resolution semi-largrangian particle methods and software that can harness the capabilities of supercomputing architectures to simulate multi-phase flows. We will also consider hybrid strategies where different computational blocks and algorithms are distributed to heterogeneous hardware (CPU/GPU) to optimize the parallel scalability of the software.