06 juli 2023 09:00 t/m 17:00 - Locatie: D@TA, gebouw 36 (EWI), Mekelweg 4, Delft | Zet in mijn agenda
On 6 July, 2023 there will be again an OpenFOAM Event at the TU Delft organized by the Dutch OpenFOAM user group. 4 speakers will give a talk on how OpenFOAM can be used for various applications.
Please register before 28 June, 2023.
Registration is free.
Talk by S. Salehi
Title:Implementation of Deep Reinforcement Learning in OpenFOAM
Recent advancements in artificial intelligence and machine learning have enabled tackling high-dimensional controlling and decision-making problems. Deep Reinforcement Learning (DRL), as a combination of deep learning and reinforcement learning, can perform immensely complicated cognitive tasks at a superhuman level. In fluid dynamics, DRL can be used to solve sophisticated optimization problems or perform complex active flow control strategies. In the present work, a coupled DRL-CFD framework was developed within OpenFOAM, as opposed to previous attempts in the literature in which the CFD solver was treated as a black box. Here, the DRL agent is implemented as a boundary condition that is able to sense the environment state, perform some action, and record the corresponding reward. To test and verify the performance of the developed DRL-CFD software, the simple test case of vortex shedding behind a 2D cylinder is investigated. The actuator is a pair of synthetic jets on top and bottom of the cylinder. Thereby, the DRL agent (which is a deep neural network) learns to minimize the drag and lift coefficients by applying the optimum jet flow at each time step.
Talk by J.C. Goeree
Title: Numerical Simulation of Hopper discharging using OpenFOAM
The TSHD is designed to operate in shallow or deep waters and has a large hopper capacity for storing the dredged material. For instance when being used in shallow waters, the clearance between the sea-bottom and the bottom of the vessel is important. Typically these kind of vessels are employed (very) near shore for instance when replenishing beaches. The clearence between bottom of the ship and sea-bottom is in this case (very) small. The bottom-doors, which are installed in the hopper, are opened discharging the sand at the designated location. It is important, given the small distance between the ship and the sea-bottom, that the extension of the opened bottom-doors below the ship’s hull is minimized while maintaining
a decent discharge rate.
Two design configurations for the bottom doors are available: a single bottom-door configuration and a double bottom-door configuration. CFD (OpenFOAM) is used to determine (qualitatively) the discharge rate of a hopper by comparing the design configurations. A CFD calculation was setup and finished in a short amount of time (approximately 2 weeks). A drift-FluxFoam calculation was employed in 2D to decrease the calculation time when compared to a 3D case. It is assumed that the sand in the hopper is fully fluidized and has liquid-like properties. Furthermore, various typical scenarios were choosen in order to assess the performance of the discharge rate under different conditions. By utilizing these CFD calculations, qualitative understanding was obtained regarding the hydrodynamic behavior during hopper discharge. Moreover, the calculations certainly shortened the time needed for choosing an optimal design which fits the purpose of the vessel.
Talk by H. Ottens & C. Estourgie (Dynaflow Research Group)
Title: Validating and extending the use of BOSfluids using OpenFOAM
At Dynaflow we developed and use BOSfluids, a powerful 1D single phase flow solver to obtain flow parameters of large piping systems. An important aspect is the flow distribution in branches. OpenFOAM is used to validate different flow configurations of a branch. Since BOSfluids is valid for single phase, OpenFOAM is used to extend the BOSfluids results towards multi-phase flows in piping systems as well.
Talk by R.Pasolari
Title: Efficient Coupling of a Modified pimpleFoam Solver with a Vortex Particle Method in OpenFOAM
The field of external aerodynamics is a broad field of engineering and its computational study poses significant challenges, particularly when dealing with intricate phenomena such as strong body-vortex interactions. It is crucial to develop efficient and accurate tools to tackle such cases successfully. This presentation introduces a hybrid Eulerian-Lagrangian solver that focuses on the realm of external aerodynamics. This coupled solver leverages the benefits offered by both the Eulerian
and Lagrangian approaches. Specifically, the Eulerian solver is employed to resolve the vicinity of solid boundaries, while the Lagrangian solver efficiently evolves the downstream wake, minimizing artificial diffusion.
OpenFOAM is utilized for the Eulerian component, while a Vortex Particle Method is employed for the Lagrangian aspect. The presentation focuses on two main topics. Firstly, it explores the coupling strategy employed in the hybrid code and presents the validation process of this approach. Secondly, it discusses the modifications made to the pimpleFoam solver of OpenFOAM to achieve an efficient coupling with the in-house Vortex Particle Method code. A class called EulerianPimpleFoam replaces the pimpleFoam module, facilitating the incorporation of coupling steps between different Eulerian time-steps. The advantages and challenges associated with this approximation will be examined and discussed upon.