| 1. |
- Aryal, Pradip, et al.
(författare)
-
Comparative study of the main electromagnetic models applied to melt pool prediction with gas metal arc : Effect on flow, ripples from drop impact, and geometry
- 2022
-
Ingår i: International Journal of Heat and Mass Transfer. - : Elsevier. - 0017-9310 .- 1879-2189. ; 194
-
Tidskriftsartikel (refereegranskat)abstract
- The present work concerns the electromagnetic force models in computational fluid dynamics simulations of melt pools produced with electric arcs. These are commonly applied to gas metal arcs with metal transfer, in welding and additive manufacturing. Metal drop impact on the melt pool is thus included in this study. The electromagnetic force models applied in literature use either numerical solutions of Poisson equations or one of the two analytical models developed by Kou and Sun, or Tsao and Wu. These models rely on assumptions for which the effect on the melt pool predictions remains to be understood. The present work thoroughly investigates those assumptions and their effects. It has been supported by dedicated experimental tests that did provide estimates of unknown model parameters and validation data. The obtained results show that the assumptions that fundamentally distinguish these three models change the electromagnetic force, including the relation between its components. These changes, which can also be spatially non-uniform, are large. As a result, these models lead to significantly different recirculation flow pattern, thermal convection, melt pool morphology, bead dimensions, and free surface response to the metal transfer. We conclude by proposing conditions in which each of these models is suited or questionable.
|
|
| 2. |
- Aryal, Pradip, et al.
(författare)
-
Effect of Substrate Orientation on Melt Pool during Multi-Layer Deposition in V-Groove with Gas Metal Arc
- 2021
-
Ingår i: Proceedings of the World Congress on Mechanical, Chemical, and Material Engineering. - : Avestia Publishing. - 2369-8136. - 9781927877616
-
Konferensbidrag (refereegranskat)abstract
- Thermo-fluid dynamic and experimental approaches are used to investigate the influence of 20° uphill, downhill and sideway substrate orientation during metal deposition over a previously deposited bead in a V-groove. The computational fluid dynamic model with free surface deformation and metal transfer gives insight into the melt pool flow and causes of defect formation observed on the solidified beads. The experimental metallographs, high-speed images and computational results show good agreement. It is found that the deposition of a second layer on a smooth first layer cooled down to room temperature leads to large changes in melt pool flow pattern at 20° substrate inclination compared to flat condition. It results in undercut and humps with the uphill orientation and undercut with the side inclination. Therefore, lower angle range is necessary for multilayer gas metal arc deposition for these two last configurations.
|
|
| 3. |
- Buntic-Ogor, Ivana, et al.
(författare)
-
An Adaptive Turbulence Model for Swirling Flow
- 2006
-
Ingår i: Conference on Turbulence and Interactions, TI2006, May 29 - June 2, 2006, Porquerolles, France.
-
Konferensbidrag (refereegranskat)abstract
- Swirling flows are very common in technical applications, especially in hydraulic machinery, and they require rather sophisticated modeling. At present, an applicative method for simulating unsteady flows is Very Large Eddy Simulation (VLES). In VLES, large turbulence structures are resolved in time and space, while the small scales are modeled with an adequate turbulence model. Therefore the turbulence model must be able to distinguish between the resolved and the unresolved scales. In order to accomplish this the method used in this work employs an adaptive and dynamic filtering technique that restricts the turbulence model to only predict the effect of non-resolvable turbulent scales. The results obtained from using the standard k-ε model is compared to those using the extended model of Chen and Kim, with and without the filtering approach. The modified k-ε model of Reif et al. is also investigated together with the filtering approach. The models are implemented in the FENFLOSS and the CALC-PMB CFD codes. The chosen test cases are swirling flow in a straight pipe and swirling flow through a straight conical diffuser.
|
|
| 4. |
|
|
| 5. |
- Gyllenram, Walter, 1971, et al.
(författare)
-
Design and Validation of a Scale-Adaptive Filtering Technique for LRN Turbulence Modeling of Unsteady Flow
- 2008
-
Ingår i: Journal of Fluid Engineering. ; Volume 130:Issue 5, s. 10-
-
Tidskriftsartikel (refereegranskat)abstract
- An adaptive low-pass filtering procedure for the modeled turbulent length and time scales is derived and applied to Wilcox' original low reynolds number k-omega turbulence model. It is shown that the method is suitable for complex industrial unsteady flows in cases where full large eddy simulations (LESs) are unfeasible. During the simulation, the modeled length and time scales are compared to what can potentially be resolved by the computational grid and time step. If the modeled scales are larger than the resolvable scales, the resolvable scales will replace the modeled scales in the formulation of the eddy viscosity. The filtered k-omega model is implemented in an in-house computational fluid dynamics (CFD) code, and numerical simulations have been made of strongly swirling flow through a sudden expansion. The new model surpasses the original model in predicting unsteady effects and producing accurate time-averaged results. It is shown to be superior to the wall-adpating local eddy-viscosity (WALE) model on the computational grids considered here, since the turbulence may not be sufficiently resolved for an accurate LES. Because of the adaptive formulation, the filtered k-omega model has the potential to be successfully used in any engineering case where an LES is unfeasible and a Reynolds (ensemble) averaged Navier–Stokes simulation is insufficient.
|
|
| 6. |
- Gyllenram, Walter, 1971, et al.
(författare)
-
Large Eddy Simulation of Turbulent Swirling Flow Through a Sudden Expansion
- 2006
-
Ingår i: 23rd IAHR Symposium, Yokohama, Japan, October 2006.
-
Konferensbidrag (refereegranskat)abstract
- Turbulent swirling flow through a sudden expansion is investigated numerically using Large Eddy Simulation (LES). The flow resembles the flow in a draft tube of a water turbine that is working at part load. The swirling inflow is subject to a strong adverse pressure gradient and the symmetry of the flow breaks down close to the inlet. This gives rise to an oscillating, helicoidal vortex core which in turn creates a highly unsteady and turbulent flow field. In this work, the large-scale turbulent structures are numerically resolved, and detailed information about the flow characteristics is obtained. The oscillating flow is analysed using Fourier transforms of the wall pressure at different downstream locations. The most dominant frequency corresponds to the rotational rate of the precessing vortex core, and it is found that this frequency is constant throughout the domain. The results of two simulations using numerical discretization schemes of different order are compared. It is shown that the frequency of the precessing vortex core is not sensitive to the choice of discretization. However, the lower frequencies of the flow depend to a higher extent on the numerical accuracy. To validate the results, the computed velocities are averaged and compared to experimental data. The agreement is good. The Reynolds stress tensor is also computed and analysed. It is found that large degrees of turbulent anisotropy are found only in the region that is dominated by the oscillating vortex core. Further downstream, the degree of turbulent anisotropy is almost negligible despite the relatively higher level of swirl.
|
|
| 7. |
- Gyllenram, Walter, 1971, et al.
(författare)
-
Numerical Investigations of Swirling flow in a Conical Diffuser
- 2004
-
Ingår i: 22nd IAHR Symposium, Stockholm, Sweden, June 29 - July 2, 2004. ; A, s. A09-2 (1-10)-
-
Konferensbidrag (refereegranskat)abstract
- When a hydraulic power plant is operating at off-design, a swirling flow exits the runner and is convected through the draft tube. The swirling flow gives rise to several features that decrease the efficiency and/or may damage the construction. As a step towards making reliable numerical predictions of the swirling flow in draft tubes, steady computations in an idealised model have been carried out. The model geometry is a straight conical diffuser and O. G. Dahlhaug of NTNU, Trondheim, Norway, has provided experimental data. In this paper, numerical 3D RANS investigations are quantitatively compared to the available experimental data. Good agreement with experimental data was obtained. The discrepancies are partly reminiscent of the nature of the k-omega turbulence model that was used in this work. Given symmetric geometry and boundary conditions, a fluid flow is most often thought to behave in an equally symmetric manner. If the flow is swirling, this is not generally true. Due to the unstable properties of the symmetric solutions for the averaged Navier-Stokes equations, the flow will collapse into an asymmetric mode. In the first computational cases of this paper, the disturbance that triggers the instability is shown to be imperfections in the CAD-geometry. Also included is a discussion concerning the development of counter-rotating vortices in the boundary layer of swirling flow in a circular pipe.
|
|
| 8. |
- Gyllenram, Walter, 1971, et al.
(författare)
-
On the failure of the quasicylindrical approximation and the connection to vortex breakdown in turbulent swirling flow
- 2007
-
Ingår i: Physics of Fluids. - : AIP Publishing. - 1070-6631 .- 1089-7666. ; 19:4
-
Tidskriftsartikel (refereegranskat)abstract
- This paper analyses the properties of viscous swirling flow in a pipe. The analysis is based on the time-averaged quasi-cylindrical Navier-Stokes equations and is applicable to steady, unsteady and turbulent swirling flow. A method is developed to determine the critical level of swirl (vortex breakdown) for an arbitrary vortex. The method can also be used for an estimation of the radial velocity profile if the other components are given or measured along a single radial line. The quasi-cylindrical equations are rearranged to yield a single ordinary differential equation for the radial distribution of the radial velocity component. The equation is singular for certain levels of swirl. It is shown that the lowest swirl level at which the equation is singular corresponds exactly to the sufficient condition for axisymmetric vortex breakdown as derived by Wang and Rusak [J. Fluid Mech. {340}, 177-223 (1997)] and Rusak et al. [AIAA J. {36}, 1848-1853 (1998)]. In narrow regions around the critical levels of swirl, the solution violates the quasi-cylindrical assumptions and the flow must undergo a drastic change of structure. The critical swirl level is determined by the sign change of the smallest eigenvalue of the discrete linear operator which relates the radial velocities to effects of viscosity and turbulence. It is shown that neither viscosity nor turbulence directly alters the critical level of swirl.
|
|
| 9. |
- Gyllenram, Walter, 1971, et al.
(författare)
-
Very Large Eddy Simulations of Draft Tube Flow
- 2006
-
Ingår i: 23rd IAHR Symposium, Yokohama, Japan, October 2006.
-
Konferensbidrag (refereegranskat)abstract
- The objective of this work is to improve numerical predictions of unsteady turbulent flows in the draft tubes of hydraulic power plants. The standard two-equation turbulence models are known to have a strong damping effect on the resolved turbulence in this type of flow. In order to reduce this negative influence of the model, while retaining the usually satisfying near-wall behaviour, a dynamic filtering technique of the turbulent length and time scales is generalised, employed and evaluated. The filter limits the influence of the modeled turbulent length and time scales on the mean flow in regions where unsteadiness can potentially be resolved. The Wilcox (1988) $k-\omega$ turbulence model was chosen as basis for the investigations, and the effects of five different filter widths were examined. The original non-filtered model is also evaluated. A swirling flow through a straight axisymmetric diffuser was chosen as a test case and detailed measurements carried out by Clausen et al. were used to validate the numerical results. The influence of the filtering approach on the resolved frequencies and the time averaged solutions were analysed. It is shown that the filtering procedure gives better predictions of the time-averaged velocity field and more information on the large scale unsteadiness.
|
|
| 10. |
- Lehr, C., et al.
(författare)
-
Unveiling the two-proton halo character of 17 Ne: Exclusive measurement of quasi-free proton-knockout reactions
- 2022
-
Ingår i: Physics Letters, Section B: Nuclear, Elementary Particle and High-Energy Physics. - : Elsevier BV. - 0370-2693 .- 1873-2445. ; 827
-
Tidskriftsartikel (refereegranskat)abstract
- The proton drip-line nucleus 17Ne is investigated experimentally in order to determine its two-proton halo character. A fully exclusive measurement of the 17Ne(p,2p)16F→15⁎O+p quasi-free one-proton knockout reaction has been performed at GSI at around 500 MeV/nucleon beam energy. All particles resulting from the scattering process have been detected. The relevant reconstructed quantities are the angles of the two protons scattered in quasi-elastic kinematics, the decay of 16F into 15O (including γ decays from excited states) and a proton, as well as the 15O+p relative-energy spectrum and the 16F momentum distributions. The latter two quantities allow an independent and consistent determination of the fractions of l=0 and l=2 motion of the valence protons in 17Ne. With a resulting relatively small l=0 component of only around 35(3)%, it is concluded that 17Ne exhibits a rather modest halo character only. The quantitative agreement of the two values deduced from the energy spectrum and the momentum distributions supports the theoretical treatment of the calculation of momentum distributions after quasi-free knockout reactions at high energies by taking into account distortions based on the Glauber theory. Moreover, the experimental data allow the separation of valence-proton knockout and knockout from the 15O core. The latter process contributes with 11.8(3.1) mb around 40% to the total proton-knockout cross section of 30.3(2.3) mb, which explains previously reported contradicting conclusions derived from inclusive cross sections.
|
|
