HB
H. Bijl
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
18 records found
1
Large eddy simulation with energy-conserving schemes and the smagorinsky model
A note on accuracy and computational efficiency
Despite advances in turbulence modelling, the Smagorinsky model remains a popular choice for large eddy simulation (LES) due to its simplicity and ease of use. The dissipation in turbulence energy that the model introduces, is proportional to the Smagorinsky constant, of which many different values have been proposed. These values have been derived for certain simulated test-cases while using a specific set of numerical schemes, to obtain the correct dissipation in energy simply because an incorrect value of the Smagorinsky constant would lead to an incorrect dissipation. However, it is important to bear in mind that numerical codes may suffer from numerical or artificial dissipation, which occurs spuriously through a combination of spatio-temporal and iterative errors. The latter can be controlled through more iterations, the former however, depends on the grid resolution and the time step. Recent research suggests that a complete energy-conserving (EC) spatio-temporal discretisation guarantees zero numerical dissipation for any grid resolution and time step. Therefore, using an EC scheme would ensure that dissipation occurs primarily through the Smagorinsky model (and errors in its implementation) than through the discretisation of the Navier-Stokes (NS) equations. To evaluate the efficacy of these schemes for engineering applications, the article first discusses the use of an EC temporal discretisation as regards to accuracy and computational effort, to ascertain whether EC time advancement is advantageous or not. It was noticed that a simple non-EC explicit method with a smaller time step not only reduces the numerical dissipation to an acceptable level but is computationally cheaper than an implicit-EC scheme for wide range of time steps. Secondly, in terms of spatial discretisation on uniform grids (popular in LES), a simple central-difference scheme is as accurate as an EC spatial discretisation. Finally, following the removal of numerical dissipation with any of the methods mentioned above, one is able to choose a Smagorinsky constant that is nearly independent of the grid resolution (within realistic bounds, for OpenFOAM and an in-house code). This article provides impetus to the efficient use of the Smagorinsky model for LES in fields such as wind farm aerodynamics and atmospheric simulations, instead of more comprehensive and computationally demanding turbulence models.
...
Despite advances in turbulence modelling, the Smagorinsky model remains a popular choice for large eddy simulation (LES) due to its simplicity and ease of use. The dissipation in turbulence energy that the model introduces, is proportional to the Smagorinsky constant, of which many different values have been proposed. These values have been derived for certain simulated test-cases while using a specific set of numerical schemes, to obtain the correct dissipation in energy simply because an incorrect value of the Smagorinsky constant would lead to an incorrect dissipation. However, it is important to bear in mind that numerical codes may suffer from numerical or artificial dissipation, which occurs spuriously through a combination of spatio-temporal and iterative errors. The latter can be controlled through more iterations, the former however, depends on the grid resolution and the time step. Recent research suggests that a complete energy-conserving (EC) spatio-temporal discretisation guarantees zero numerical dissipation for any grid resolution and time step. Therefore, using an EC scheme would ensure that dissipation occurs primarily through the Smagorinsky model (and errors in its implementation) than through the discretisation of the Navier-Stokes (NS) equations. To evaluate the efficacy of these schemes for engineering applications, the article first discusses the use of an EC temporal discretisation as regards to accuracy and computational effort, to ascertain whether EC time advancement is advantageous or not. It was noticed that a simple non-EC explicit method with a smaller time step not only reduces the numerical dissipation to an acceptable level but is computationally cheaper than an implicit-EC scheme for wide range of time steps. Secondly, in terms of spatial discretisation on uniform grids (popular in LES), a simple central-difference scheme is as accurate as an EC spatial discretisation. Finally, following the removal of numerical dissipation with any of the methods mentioned above, one is able to choose a Smagorinsky constant that is nearly independent of the grid resolution (within realistic bounds, for OpenFOAM and an in-house code). This article provides impetus to the efficient use of the Smagorinsky model for LES in fields such as wind farm aerodynamics and atmospheric simulations, instead of more comprehensive and computationally demanding turbulence models.
Vortex generators (VGs) are a widely used means of flow control, and predictions of their influence are vital for efficient designs. However, accurate CFD simulations of their effect on the flow field by means of a body fitted mesh are computationally expensive. Therefore the BAY and jBAY models, which represent the effect of VGs on the flow using source terms in the momentum equations, are popular in industry. In this contribution we examine the ability of the BAY and jBAY model to provide accurate flow field results by looking at boundary layer properties close behind VGs. The results are compared with both body fitted mesh and other source term model RANS simulations of 3D incompressible flows, over flat plate and airfoil geometries. We show the influence of mesh resolution and domain of application on the accuracy of the models and investigate the influence of the source term on the generated flow field. Our results demonstrate the grid dependence of the models and indicate the presence of model errors. Furthermore we find that the total applied force has a larger influence on both the intensity and shape of the created vortex than the distribution of the source term over the cells.
...
Vortex generators (VGs) are a widely used means of flow control, and predictions of their influence are vital for efficient designs. However, accurate CFD simulations of their effect on the flow field by means of a body fitted mesh are computationally expensive. Therefore the BAY and jBAY models, which represent the effect of VGs on the flow using source terms in the momentum equations, are popular in industry. In this contribution we examine the ability of the BAY and jBAY model to provide accurate flow field results by looking at boundary layer properties close behind VGs. The results are compared with both body fitted mesh and other source term model RANS simulations of 3D incompressible flows, over flat plate and airfoil geometries. We show the influence of mesh resolution and domain of application on the accuracy of the models and investigate the influence of the source term on the generated flow field. Our results demonstrate the grid dependence of the models and indicate the presence of model errors. Furthermore we find that the total applied force has a larger influence on both the intensity and shape of the created vortex than the distribution of the source term over the cells.
Journal article
(2017)
-
Shuanghou Deng, Mustafa Perçin, Bas van Oudheusden, Hester Bijl, Bart Remes, T Xiao
Journal article
(2017)
-
Ye Zhang, Thijs Gillebaart, Alexander van Zuijlen, Gerard van Bussel, Hester Bijl
This paper presents the experimental and numerical study on MEXICO wind turbine blades. Previous work by other researchers shows that large deviations exist in the loads comparison between numerical predictions and experimental data for the rotating MEXICO wind turbine. To reduce complexities and uncertainties, a non-rotating experimental campaign has been carried out on MEXICO blades Delft University of Technology. In this new measurement, quasi-2D aerodynamic characteristics of MEXICO blades on three spanwise sections are measured at different inflow velocities and angles of attack. Additionally, RANS simulations are performed with OpenFOAM-2.1.1 to compare numerical results against measured data. The comparison and analysis of aerodynamic loads on the blade, where three different airfoil families and geometrical transition regions are used, show that for attached flow condition, RANS computation predicts excellent pressure distribution on the NACA airfoil section (r=R D 0.92) and good agreement is observed on the DU (r=R D 0.35) and RISØ (r=R D 0.60) airfoil sections. Unexpected aerodynamic characteristics are observed at the intermediate transition regions connecting the RISØ and DU airfoils, where sudden lift force drop is found at the radial position r=R D 0.55. Through numerical flow visualization, large-scale vortical structures are observed on the suction side of the blade near the mid-span. Moreover, counter-rotating vortices are generated behind the blade at locations where unexpected loads occurs. Consequently, the RISØ airfoil could not give expected 2D aerodynamic characteristics because of upwash/downwash effects induced by these counter-rotating vortices, which make 3D effects play an important role in numerical modeling when calculating the aerodynamic loads for MEXICO rotor.
...
This paper presents the experimental and numerical study on MEXICO wind turbine blades. Previous work by other researchers shows that large deviations exist in the loads comparison between numerical predictions and experimental data for the rotating MEXICO wind turbine. To reduce complexities and uncertainties, a non-rotating experimental campaign has been carried out on MEXICO blades Delft University of Technology. In this new measurement, quasi-2D aerodynamic characteristics of MEXICO blades on three spanwise sections are measured at different inflow velocities and angles of attack. Additionally, RANS simulations are performed with OpenFOAM-2.1.1 to compare numerical results against measured data. The comparison and analysis of aerodynamic loads on the blade, where three different airfoil families and geometrical transition regions are used, show that for attached flow condition, RANS computation predicts excellent pressure distribution on the NACA airfoil section (r=R D 0.92) and good agreement is observed on the DU (r=R D 0.35) and RISØ (r=R D 0.60) airfoil sections. Unexpected aerodynamic characteristics are observed at the intermediate transition regions connecting the RISØ and DU airfoils, where sudden lift force drop is found at the radial position r=R D 0.55. Through numerical flow visualization, large-scale vortical structures are observed on the suction side of the blade near the mid-span. Moreover, counter-rotating vortices are generated behind the blade at locations where unexpected loads occurs. Consequently, the RISØ airfoil could not give expected 2D aerodynamic characteristics because of upwash/downwash effects induced by these counter-rotating vortices, which make 3D effects play an important role in numerical modeling when calculating the aerodynamic loads for MEXICO rotor.
Radial Basis Function (RBF) mesh deformation is one of the most robust mesh deformation methods available. Using the greedy (data reduction) method in combination with an explicit boundary correction, results in an efficient method as shown in literature. However, to ensure the method remains robust, two issues are addressed: 1) how to ensure that the set of control points remains an accurate representation of the geometry in time and 2) how to use/automate the explicit boundary correction, while ensuring a high mesh quality. In this paper, we propose an adaptive RBF mesh deformation method, which ensures the set of control points always represents the geometry/displacement up to a certain (user-specified) criteria, by keeping track of the boundary error throughout the simulation and re-selecting when needed. Opposed to the unit displacement and prescribed displacement selection methods, the adaptive method is more robust, user-independent and efficient, for the cases considered. Secondly, the analysis of a single high aspect ratio cell is used to formulate an equation for the correction radius needed, depending on the characteristics of the correction function used, maximum aspect ratio, minimum first cell height and boundary error. Based on the analysis two new radial basis correction functions are derived and proposed. This proposed automated procedure is verified while varying the correction function, Reynolds number (and thus first cell height and aspect ratio) and boundary error. Finally, the parallel efficiency is studied for the two adaptive methods, unit displacement and prescribed displacement for both the CPU as well as the memory formulation with a 2D oscillating and translating airfoil with oscillating flap, a 3D flexible locally deforming tube and deforming wind turbine blade. Generally, the memory formulation requires less work (due to the large amount of work required for evaluating RBF's), but the parallel efficiency reduces due to the limited bandwidth available between CPU and memory. In terms of parallel efficiency/scaling the different studied methods perform similarly, with the greedy algorithm being the bottleneck. In terms of absolute computational work the adaptive methods are better for the cases studied due to their more efficient selection of the control points. By automating most of the RBF mesh deformation, a robust, efficient and almost user-independent mesh deformation method is presented.
...
Radial Basis Function (RBF) mesh deformation is one of the most robust mesh deformation methods available. Using the greedy (data reduction) method in combination with an explicit boundary correction, results in an efficient method as shown in literature. However, to ensure the method remains robust, two issues are addressed: 1) how to ensure that the set of control points remains an accurate representation of the geometry in time and 2) how to use/automate the explicit boundary correction, while ensuring a high mesh quality. In this paper, we propose an adaptive RBF mesh deformation method, which ensures the set of control points always represents the geometry/displacement up to a certain (user-specified) criteria, by keeping track of the boundary error throughout the simulation and re-selecting when needed. Opposed to the unit displacement and prescribed displacement selection methods, the adaptive method is more robust, user-independent and efficient, for the cases considered. Secondly, the analysis of a single high aspect ratio cell is used to formulate an equation for the correction radius needed, depending on the characteristics of the correction function used, maximum aspect ratio, minimum first cell height and boundary error. Based on the analysis two new radial basis correction functions are derived and proposed. This proposed automated procedure is verified while varying the correction function, Reynolds number (and thus first cell height and aspect ratio) and boundary error. Finally, the parallel efficiency is studied for the two adaptive methods, unit displacement and prescribed displacement for both the CPU as well as the memory formulation with a 2D oscillating and translating airfoil with oscillating flap, a 3D flexible locally deforming tube and deforming wind turbine blade. Generally, the memory formulation requires less work (due to the large amount of work required for evaluating RBF's), but the parallel efficiency reduces due to the limited bandwidth available between CPU and memory. In terms of parallel efficiency/scaling the different studied methods perform similarly, with the greedy algorithm being the bottleneck. In terms of absolute computational work the adaptive methods are better for the cases studied due to their more efficient selection of the control points. By automating most of the RBF mesh deformation, a robust, efficient and almost user-independent mesh deformation method is presented.
A robust and efficient dynamic grid strategy based on an overset grid coupled with mesh deformation technique is proposed for simulating unsteady flow of flapping wings undergoing large geometrical displacement. The dynamic grid method was implemented using a hierarchical unstructured overset grid locally coupled with a fast radial basis function (RBF)-based mapping approach. The hierarchically organized overset grid allows transferring the grid resolution for multiple blocks and overlapping/embedding the meshes. The RBF-based mapping approach is particularly highlighted in this paper in view of its considerable computational efficiency compared with conventional RBF evaluation. The performance of the proposed dynamic mesh strategy is demonstrated by three typical unsteady cases, including a rotating rectangular block in a fixed domain, a relative movement between self-propelled fishes and the X-wing type flapping-wing micro air vehicle DelFly, which displays the clap-and-fling wing-interaction phenomenon on both sides of the fuselage. Results show that the proposed method can be applied to the simulation of flapping wings with satisfactory efficiency and robustness.
...
A robust and efficient dynamic grid strategy based on an overset grid coupled with mesh deformation technique is proposed for simulating unsteady flow of flapping wings undergoing large geometrical displacement. The dynamic grid method was implemented using a hierarchical unstructured overset grid locally coupled with a fast radial basis function (RBF)-based mapping approach. The hierarchically organized overset grid allows transferring the grid resolution for multiple blocks and overlapping/embedding the meshes. The RBF-based mapping approach is particularly highlighted in this paper in view of its considerable computational efficiency compared with conventional RBF evaluation. The performance of the proposed dynamic mesh strategy is demonstrated by three typical unsteady cases, including a rotating rectangular block in a fixed domain, a relative movement between self-propelled fishes and the X-wing type flapping-wing micro air vehicle DelFly, which displays the clap-and-fling wing-interaction phenomenon on both sides of the fuselage. Results show that the proposed method can be applied to the simulation of flapping wings with satisfactory efficiency and robustness.
Journal article
(2016)
-
David S. Blom, Philipp Birken, Hester Bijl, Fleur Kessels, Andreas Meister, Alexander H. van Zuijlen
In this article, we endeavour to find a fast solver for finite volume discretizations for compressible unsteady viscous flows. Thereby, we concentrate on comparing the efficiency of important classes of time integration schemes, namely time adaptive Rosenbrock, singly diagonally implicit (SDIRK) and explicit first stage singly diagonally implicit Runge-Kutta (ESDIRK) methods. To make the comparison fair, efficient equation system solvers need to be chosen and a smart choice of tolerances is needed. This is determined from the tolerance TOL that steers time adaptivity. For implicit Runge-Kutta methods, the solver is given by preconditioned inexact Jacobian-free Newton-Krylov (JFNK) and for Rosenbrock, it is preconditioned Jacobian-free GMRES. To specify the tolerances in there, we suggest a simple strategy of using TOL/100 that is a good compromise between stability and computational effort. Numerical experiments for different test cases show that the fourth order Rosenbrock method RODASP and the fourth order ESDIRK method ESDIRK4 are best for fine tolerances, with RODASP being the most robust scheme.
...
In this article, we endeavour to find a fast solver for finite volume discretizations for compressible unsteady viscous flows. Thereby, we concentrate on comparing the efficiency of important classes of time integration schemes, namely time adaptive Rosenbrock, singly diagonally implicit (SDIRK) and explicit first stage singly diagonally implicit Runge-Kutta (ESDIRK) methods. To make the comparison fair, efficient equation system solvers need to be chosen and a smart choice of tolerances is needed. This is determined from the tolerance TOL that steers time adaptivity. For implicit Runge-Kutta methods, the solver is given by preconditioned inexact Jacobian-free Newton-Krylov (JFNK) and for Rosenbrock, it is preconditioned Jacobian-free GMRES. To specify the tolerances in there, we suggest a simple strategy of using TOL/100 that is a good compromise between stability and computational effort. Numerical experiments for different test cases show that the fourth order Rosenbrock method RODASP and the fourth order ESDIRK method ESDIRK4 are best for fine tolerances, with RODASP being the most robust scheme.
High computational cost associated with the high fidelity wake models such as RANS or LES serves as a primary bottleneck to perform a direct high fidelity wind farm layout optimization (WFLO) using accurate CFD based wake models. Therefore, a surrogate based multi-fidelity WFLO methodology (SWFLO) is proposed. The surrogate model is built using an SBO method referred as manifold mapping (MM). As a verification, optimization of spacing between two staggered wind turbines was performed using the proposed surrogate based methodology and the performance was compared with that of direct optimization using high fidelity model. Significant reduction in computational cost was achieved using MM: a maximum computational cost reduction of 65%, while arriving at the same optima as that of direct high fidelity optimization. The similarity between the response of models, the number of mapping points and its position, highly influences the computational efficiency of the proposed method. As a proof of concept, realistic WFLO of a small 7-turbine wind farm is performed using the proposed surrogate based methodology. Two variants of Jensen wake model with different decay coefficients were used as the fine and coarse model. The proposed SWFLO method arrived at the same optima as that of the fine model with very less number of fine model simulations.
...
High computational cost associated with the high fidelity wake models such as RANS or LES serves as a primary bottleneck to perform a direct high fidelity wind farm layout optimization (WFLO) using accurate CFD based wake models. Therefore, a surrogate based multi-fidelity WFLO methodology (SWFLO) is proposed. The surrogate model is built using an SBO method referred as manifold mapping (MM). As a verification, optimization of spacing between two staggered wind turbines was performed using the proposed surrogate based methodology and the performance was compared with that of direct optimization using high fidelity model. Significant reduction in computational cost was achieved using MM: a maximum computational cost reduction of 65%, while arriving at the same optima as that of direct high fidelity optimization. The similarity between the response of models, the number of mapping points and its position, highly influences the computational efficiency of the proposed method. As a proof of concept, realistic WFLO of a small 7-turbine wind farm is performed using the proposed surrogate based methodology. Two variants of Jensen wake model with different decay coefficients were used as the fine and coarse model. The proposed SWFLO method arrived at the same optima as that of the fine model with very less number of fine model simulations.
A numerical investigation is performed to address the flexing effect on the propulsion performance of flapping wing particularly on the counter-flapping wings of the biplane configuration. A Reynolds number of 10,000 is considered
in the present study which corresponds to the flight regime of most existing flapping wing micro air vehicles. The computation involves solving the compressible unsteady Reynoldsaveraged Native-Stokes equation using an inhouse developed code. The flapping motion is incorporated by an efficient deforming overset grid technique which allows multiple flexible bodies to be embedded into the flow field. Results show that the biplane wing with counterflapping configuration has a better propulsive performance in comparison to a single flapping wing. A low-pressure regime between the two wings during the outstroke produces more thrust, while the counter-flapping motion can also generate a surfeit momentum rushing in to the wake. The more flexible wing can produce more thrust while less power is required thus owning a better
propulsive performance.
...
A numerical investigation is performed to address the flexing effect on the propulsion performance of flapping wing particularly on the counter-flapping wings of the biplane configuration. A Reynolds number of 10,000 is considered
in the present study which corresponds to the flight regime of most existing flapping wing micro air vehicles. The computation involves solving the compressible unsteady Reynoldsaveraged Native-Stokes equation using an inhouse developed code. The flapping motion is incorporated by an efficient deforming overset grid technique which allows multiple flexible bodies to be embedded into the flow field. Results show that the biplane wing with counterflapping configuration has a better propulsive performance in comparison to a single flapping wing. A low-pressure regime between the two wings during the outstroke produces more thrust, while the counter-flapping motion can also generate a surfeit momentum rushing in to the wake. The more flexible wing can produce more thrust while less power is required thus owning a better
propulsive performance.
Particle image velocimetry measurements and simultaneous force measurements have been performed on the DelFly II flapping-wing MAV, to investigate the flow-field behavior and the aerodynamic forces generated. For flapping wing motion it is expected that both the clap and peel mechanism and the occurrence of a leading edge vortex during the translational phase play an important role in unsteady lift generation. Furthermore, the flexibility of the wing foil is also considered of primary relevance. The PIV analysis shows a strong influx between the wings during the peel but no downward expelling jet during the clap. The force measurements reveal that the peel, oppositely to the clap, contributes significantly to the lift. The PIV visualization suggests the occurrence of a leading edge vortex during the first half of the in- and outstroke, which is supported by a simultaneous augmentation in lift. The early generation of a leading edge vortex during the flex cannot be assessed from the PIV images due to optical obstruction, but is likely to appear since the wing flexing is accompanied with a large increase in lift.
...
Particle image velocimetry measurements and simultaneous force measurements have been performed on the DelFly II flapping-wing MAV, to investigate the flow-field behavior and the aerodynamic forces generated. For flapping wing motion it is expected that both the clap and peel mechanism and the occurrence of a leading edge vortex during the translational phase play an important role in unsteady lift generation. Furthermore, the flexibility of the wing foil is also considered of primary relevance. The PIV analysis shows a strong influx between the wings during the peel but no downward expelling jet during the clap. The force measurements reveal that the peel, oppositely to the clap, contributes significantly to the lift. The PIV visualization suggests the occurrence of a leading edge vortex during the first half of the in- and outstroke, which is supported by a simultaneous augmentation in lift. The early generation of a leading edge vortex during the flex cannot be assessed from the PIV images due to optical obstruction, but is likely to appear since the wing flexing is accompanied with a large increase in lift.