In the field of geoscience, the full knowledge of atmospheric conditions is necessary in the fields of urban planning, disaster mitigation, emergency response and air pollution assessments. While we see the remarkable advantages of the meso-scale model of Weather Research and Forecast (WRF) in the simulation of the atmosphere, it shows significant deficiencies in describing the micro-scale effects of the complex underlying surface. For solving this problem, a new adapter is developed based on preCICE. Thanks to the open source nature of the WRF model, the developed adapter enables the WRF to be coupled with other widely used Computational Fluid Dynamics (CFD) software, such as OpenFOAM. In fact, the developed adapter aims to provide the WRF model the Fluid-Fluid coupling capacity, which could be used to conduct a trans-scale numerical simulation downscaling the results from a numeric weather prediction software to drive a micro-scale (~1m) flow simulation. Such a trans-scale simulation shows the detailed flow structures inside the urban boundary layer or above complex hilly terrain.
-Jingya Li, Delft University of Technology, The Netherlands
+ Simulation of coupled particle transport and FSI with application in the drilling industryPatrick Höhn (@hoehnp), Institute for Computer Science, University of Göttingen, Germany
Accurate simulation of fluid-structure interactions (FSI) remains a difficult task in computational mechanics, especially when dealing with complicated geometries and dynamic coupling between fluid and solid domains. This paper introduces novel benchmarks in the field of FSI that take advantage of isogeometric analysis (IGA) and the adaptibility of the preCICE coupling library. We offer a framework that combines the IGA-based solid mechanics library G+Smo, the computational fluid dynamics capabilities of openFOAM, and the Julia-based WaterLily.jl fluid solver, aimed at advancing hydrodynamic simulations. Central to our approach is the utilization of spline-based communication for IGA-based fluid-structure interaction simulations. We employ spline-based communication instead of quadrature points to minimize the amount of information exchanged. A comparison of accuracy and efficiency between spline-based communication and quadrature point-based communication will be presented. Several benchmarks will be discussed, ranging from the replication of established preCICE cases to direct comparisons with other solid mechanics libraries, to demonstrate the effectiveness of spline-based communication. Through these benchmarks, we conclude that spline-based communication is more efficient than quadrature point-based communication and yields the same level of accuracy. +
Drilling is essential for the recovery and storage of sub-surface energy, such as oil, gas and geothermal heat. It typically accounts for large parts of the project costs. For optimal drilling operations it is required to achieve an efficient transport of cuttings from the drill-bit to the surface. As drSilling often reaches several thousand meters below the surface, in-situ measurements of drilling parameters are very challenging. Therefore, limited field knowledge about the underlying phenomena exists and many investigations rely on simplified laboratory setups and detailed simulations. Besides technical challenges, drilling projects are always very costly, e.g. in case of deep geothermal wells the typically drilling costs account for 50% of the total project costs. Large shares of these costs are caused by non-productive time during the drilling process caused by damages to underground equipment. Particular importance in these fatigue processes are lateral vibrations of the drill string. The research problem studied by the author attempts to evaluate the influence of the cuttings transport on the damping of lateral vibrations, which requires a simulation consisting both of particle transport and fluid-structure interaction. One approach using OpenFOAM and the particle solver XDEM was already presented in previous work. Because the code of XDEM, is not publically available, the author decided to solely use publically available open source libaries for his own approach. OpenFOAM was kept as solid base for the development. A big challenge caused by the community is the limitation that code contributions are usually bound to the OpenFOAM version of the initial development with no adoptions to newer versions. Since the initial design the particle transport is realized using the CFDEM®coupling libarary and the particle solver LIGGGHTS. Both were modified to allow a deformable mesh in LIGGGHTS. The FSI aspect was more recently realized by the FSI-library solids4Foam, which has seen significant changes in version 2. Most significantly it is now compatible with the multi-physics framework preCICE. Inspired by this change, the author realized that preCICE cannot only solve the issue of coupling different codes, but also help to overcome compatibility issues between different OpenFOAM additions to be coupled. Implementing the solvers from CFDEM®coupling-PFM in preCICE would allow a much wider application with other simulation codes for the simulation of coupled particle transport simulations, e.g. solid models in solids4Foam could be easily coupled with all solvers from CFDEM®coupling.
Arulnambi Palani, Helmut Schmidt University - University of the Federal Armed Forces Hamburg, Germany
+ Isogeometric Analysis Suitable Coupling Methods for Fluid-Structure Interactions with Solid-solver G+Smo Coupled via preCICEJingya Li, Delft University of Technology, The Netherlands
This contribution presents a fully coupled simulation methodology for modeling a fire in a building and the developing structural damage to the concrete walls. Simulation of a fire without considering structural damages is already a challenging task due to the need to accurately account for a variety of chemical and physical processes such as pyrolysis, combustion, turbulence and heat transfer by convection, conduction and radiation. To achieve a practical and computational efficient approach, it is crucial to consider the expected computing times when selecting the models. Currently, the simulation methodology is implemented using the open-source software Fire Dynamics Simulator (FDS), which is a finite-difference solver of the Navier-Stokes equations on Cartesian grids. FDS relies on the large-eddy simulation technique to account for the instantaneous turbulent flow. The complexity increases when the fire causes structural damage to the building. In this study, the concrete damage in the form of cracks, holes or spalling is computed using a phase-field method with a FEniCS-based solver. The thermal boundary conditions are provided by the fire simulation. Thus, both solvers are coupled using the open-source coupling framework preCICE, which transfers wall temperatures from the fire simulation to the structural solver. In return the structural solver sends the predicted spalling data to the fluid solver. Consequently, the computational domain of the fluid solver must be adapted to account for the generated holes in the wall structure, affecting the ongoing CFD simulation. These holes facilitate the leakage of smoke gases and radiative heat transfer through the concrete wall, thereby contributing to the spread of the fire. In this work, the newly developed two-way coupled approach for the fire-structure interaction is applied to sample cases of thermal spalling induced in a concrete wall structure and the resulting leakage of hot gases. Another challenge is the implementation on a high-performance computer. Similar to many other coupled problems, the computational effort is not equally distributed between the two disciplines involved. Simulating the turbulent fluid flow and heat transfer in an entire building typically requires much more CPU time than predicting the structural response. This imbalance is taken into account by assigning a larger number of nodes to the MPI-parallelized CFD simulation compared to the structural simulation. All three codes are implemented on the local HPC cluster HSUper, which consists of 571 compute nodes, each equipped with 2 Intel Icelake sockets (Xeon Platinum 8360Y, 36 cores) enabling fast and efficient simulations. +
Accurate simulation of fluid-structure interactions (FSI) remains a difficult task in computational mechanics, especially when dealing with complicated geometries and dynamic coupling between fluid and solid domains. This paper introduces novel benchmarks in the field of FSI that take advantage of isogeometric analysis (IGA) and the adaptibility of the preCICE coupling library. We offer a framework that combines the IGA-based solid mechanics library G+Smo, the computational fluid dynamics capabilities of openFOAM, and the Julia-based WaterLily.jl fluid solver, aimed at advancing hydrodynamic simulations. Central to our approach is the utilization of spline-based communication for IGA-based fluid-structure interaction simulations. We employ spline-based communication instead of quadrature points to minimize the amount of information exchanged. A comparison of accuracy and efficiency between spline-based communication and quadrature point-based communication will be presented. Several benchmarks will be discussed, ranging from the replication of established preCICE cases to direct comparisons with other solid mechanics libraries, to demonstrate the effectiveness of spline-based communication. Through these benchmarks, we conclude that spline-based communication is more efficient than quadrature point-based communication and yields the same level of accuracy.
Patrick Höhn (@hoehnp), Institute for Computer Science, University of Göttingen, Germany
+ Application of preCICE for fire-structure interaction predicting the damage of concrete walls under fire loadArulnambi Palani, Helmut Schmidt University - University of the Federal Armed Forces Hamburg, Germany
Drilling is essential for the recovery and storage of sub-surface energy, such as oil, gas and geothermal heat. It typically accounts for large parts of the project costs. For optimal drilling operations it is required to achieve an efficient transport of cuttings from the drill-bit to the surface. As drSilling often reaches several thousand meters below the surface, in-situ measurements of drilling parameters are very challenging. Therefore, limited field knowledge about the underlying phenomena exists and many investigations rely on simplified laboratory setups and detailed simulations. Besides technical challenges, drilling projects are always very costly, e.g. in case of deep geothermal wells the typically drilling costs account for 50% of the total project costs. Large shares of these costs are caused by non-productive time during the drilling process caused by damages to underground equipment. Particular importance in these fatigue processes are lateral vibrations of the drill string. The research problem studied by the author attempts to evaluate the influence of the cuttings transport on the damping of lateral vibrations, which requires a simulation consisting both of particle transport and fluid-structure interaction. One approach using OpenFOAM and the particle solver XDEM was already presented in previous work. Because the code of XDEM, is not publically available, the author decided to solely use publically available open source libaries for his own approach. OpenFOAM was kept as solid base for the development. A big challenge caused by the community is the limitation that code contributions are usually bound to the OpenFOAM version of the initial development with no adoptions to newer versions. Since the initial design the particle transport is realized using the CFDEM®coupling libarary and the particle solver LIGGGHTS. Both were modified to allow a deformable mesh in LIGGGHTS. The FSI aspect was more recently realized by the FSI-library solids4Foam, which has seen significant changes in version 2. Most significantly it is now compatible with the multi-physics framework preCICE. Inspired by this change, the author realized that preCICE cannot only solve the issue of coupling different codes, but also help to overcome compatibility issues between different OpenFOAM additions to be coupled. Implementing the solvers from CFDEM®coupling-PFM in preCICE would allow a much wider application with other simulation codes for the simulation of coupled particle transport simulations, e.g. solid models in solids4Foam could be easily coupled with all solvers from CFDEM®coupling. +
This contribution presents a fully coupled simulation methodology for modeling a fire in a building and the developing structural damage to the concrete walls. Simulation of a fire without considering structural damages is already a challenging task due to the need to accurately account for a variety of chemical and physical processes such as pyrolysis, combustion, turbulence and heat transfer by convection, conduction and radiation. To achieve a practical and computational efficient approach, it is crucial to consider the expected computing times when selecting the models. Currently, the simulation methodology is implemented using the open-source software Fire Dynamics Simulator (FDS), which is a finite-difference solver of the Navier-Stokes equations on Cartesian grids. FDS relies on the large-eddy simulation technique to account for the instantaneous turbulent flow. The complexity increases when the fire causes structural damage to the building. In this study, the concrete damage in the form of cracks, holes or spalling is computed using a phase-field method with a FEniCS-based solver. The thermal boundary conditions are provided by the fire simulation. Thus, both solvers are coupled using the open-source coupling framework preCICE, which transfers wall temperatures from the fire simulation to the structural solver. In return the structural solver sends the predicted spalling data to the fluid solver. Consequently, the computational domain of the fluid solver must be adapted to account for the generated holes in the wall structure, affecting the ongoing CFD simulation. These holes facilitate the leakage of smoke gases and radiative heat transfer through the concrete wall, thereby contributing to the spread of the fire. In this work, the newly developed two-way coupled approach for the fire-structure interaction is applied to sample cases of thermal spalling induced in a concrete wall structure and the resulting leakage of hot gases. Another challenge is the implementation on a high-performance computer. Similar to many other coupled problems, the computational effort is not equally distributed between the two disciplines involved. Simulating the turbulent fluid flow and heat transfer in an entire building typically requires much more CPU time than predicting the structural response. This imbalance is taken into account by assigning a larger number of nodes to the MPI-parallelized CFD simulation compared to the structural simulation. All three codes are implemented on the local HPC cluster HSUper, which consists of 571 compute nodes, each equipped with 2 Intel Icelake sockets (Xeon Platinum 8360Y, 36 cores) enabling fast and efficient simulations.