What's New in FLOW-3D 9.3
2008年8月30日 浏览：159106 评论：5
I. New Models and Features
Elastic membranes and walls
A limited fluid-structure interaction (FSI) capability is now available in FLOW-3D. It currently consists of two basic models: an elastic membrane and elastically flexing wall. The main assumption in both models is that the deformations in response to external forces are small compared to the lateral extents of the deforming objects. In this case, these deformations can be described by analytical functions. The deformations impact the motion of the adjacent fluid, while pressure in the fluid in turn affects the deformations. These interactions are described in the model in a fully coupled fashion.
The model is accessed through the Moving and deforming objects dialog in Physics and Deforming Components in Meshing & Geometry. See Flow Science Technical Note #79 for details.
Thermally active volume of geometry
The unstructured memory technique introduced in version 9.2 is a major benefit for users in the casting industry and elsewhere with its drastic reduction to solver memory and CPU time requirements. However, when the full fluid/solid heat transfer model was used, the memory savings were lost, as all computational cells were required to be 揳ctive.?In version 9.3, a user can now define the Maximum thermal penetration depth for each solid component under Solid Properties in the Geometry tree in the Meshing & Geometry tab. The thermal solution is then obtained within the volume along the component抯 open surface bounded by this depth. Everything outside of this volume is removed from the calculation resulting in faster simulations.
Casting users will benefit greatly from this feature. For example, during die casting filling, heat penetrates the die by only several millimeters and, therefore, modeling the whole die is unnecessary. If the user defines the thermally active layer thickness to be, for example, 5 mm, only cells within that layer are retained in the calculation, potentially reducing the total number of cells in the model several fold.
Residue model
The residue from dried liquid drops has implications for many useful applications, including coating processes, formation of pixel arrays of organic materials for video displays and for a variety of micro-electro-mechanical (MEMS) devices. A new residue model in version 9.3 allows users to model the formation of a solid residue during the drying process (i.e. the 揷offee ring?effect) and investigate the influence of such parameters as the initial solute concentration, fluid viscosity, volatility of the solvent, evaporation rate, surface tension and initial shape of the drop.
The model is accessed by directly editing the prepin file in the text editor. See Flow Science Technical Note #80 for details.
Two-component compressible gas with evaporation/condensation
The two-fluid, liquid/vapor phase change model has been extended to include a non-condensable gas component, allowing the users to describe, for example, a system consisting of liquid and gaseous hydrogen and an inert gas such as helium.
The model is accessed through the Bubble and phase change dialog in Physics. See Flow Science Technical Note #78 for details.
Running simulations
Navigator and Workspaces
The new Navigator feature in the FLOW-3D allows the user to manage several simulations at the same time, including building models, running the solver and post-processing results. The concept of Workspaces is introduced to help the user better organize his work. A Workspace generally contains a group of simulations that may be related to the same development project, including restart simulations. Existing and new simulations can be added to a workspace where they may be edited, executed and post-processed ?all within the same GUI session. All simulations in a workspace can be executed sequentially at a click of a button. Simultaneous execution of several simulations is also possible, as long as the simulations are not located in the same folder. Multiple Workspaces can be created within a single GUI session.
The introduction of Navigator represents a new approach to setting up and running simulations in FLOW-3D, and we have included with the distribution a detailed 揾ow-to?guide for your benefit. IT IS STRONGLY RECOMMENDED THAT ALL USERS REVIEW THIS GUIDE BEFORE USING VERSION 9.3.
Runtime changes to solver parameters
It is common for a user to want to make a change to the numerical settings of a simulation to make it run more accurately and efficiently. There could be more than one reason to change the settings, e.g., poor convergence, a small time-step size, a mentor tip, or just an omission in the model setup. The new Runtime Options tool in the GUI抯 Simulate tab is designed to help in such situations. It allows the user to change most of the numerical options during the simulation without the need to stop or even pause it.
Simulation restarts
There are significant changes in version 9.3 to how restart simulations are carried out. The changes are designed to simplify and generalize the restart function and require that all restart simulations from the earlier versions of FLOW-3D be redefined in accordance with the description below:
Restart logic
The prepinr file naming convention for restart simulations is no longer used; the restart simulations have now the same naming rules as the regular ones, that is, all simulations are named prepin with a unique file name extension. The user needs to rename all old restart input files to prepin.<extension> before opening them in the GUI.
In the Restart dialog of the GUI抯 General tab, activate the restart model, select the Restart source file, then pick the Restart time. The Restart source file is used as the source of the initial data for the restart simulation and can be located in a folder other than the input file. These changes make it easier to perform multiple restarts from the same Restart source file, or consecutive restart simulations using Navigator.
Reset fluid velocity and pressure at restart time
The option to reset all fluid velocities to zero and pressure to the void pressure PVOID at restart time has been added. This is useful, for example, when modeling thermal stresses in a restart run from a filling simulation, where the residual metal velocity and pressure in the mold is not desired as the initial condition to the stress simulation.
Moving Object Model
The applications for the General Moving Object model have been expanding with every release as the model matures. With its extension to cylindrical coordinate systems, the need for the old linear moving obstacle model has been completely eliminated, and it has been removed from FLOW-3D (see Removed models).
The definition of the GMO motion has been simplified in the GUI by adding a separate Edit dialog next to each moving component in the Geometry tree in the Meshing & Geometry tab.
Below is the list of new additions to the General Moving Object model in version 9.3:
Moving object assemblies
Moving components can now be made of several materials characterized by their density. Each solid subcomponent that makes up a moving component can be assigned its own density. The mass properties are then automatically evaluated from the distribution of mass within the component. This will allow the users to model coupled motion of multi-material bodies.
Mass density for each subcomponent is defined under Subcomponent branch in Geometry tree in the Meshing & Geometry tab.
Reference point for six DoF motion
The prescribed translational and rotational motion of a moving object can now be defined about an arbitrary reference point, instead of only its mass center. Together with the definition of the actual mass center, this addition simplifies the definition of the six DoF prescribed motion.
Partial slip boundary condition
Viscous boundary conditions between fluid and moving objects have been extended to include partial slip, bringing it up to the same level of capability as that for the stationary objects.
Motion limiters for prescribed-motion objects
Linear and angular motion limiters can now be applied to prescribed-motion GMO components, and not only to the couple-motion ones. This will make it easier to make a prescribed-motion GMO stop at a desired location. For example, when modeling the shot sleeve in the high pressure die casting, the plunger can be easily stopped at the end of the sleeve. Note that the motion limiters are defined in terms of the total displacement - angular or linear - from the initial location.
Mesh boundary conditions
Several refinements and additions have been made to the mesh boundary condition models. Some are designed for specific applications, while others will be useful in general cases.
Linear surface waves
The linear wave boundary condition has been extended to allow for up to a hundred linear wave components to be generated at the same vertical mesh boundary. The purpose of this addition is to enable the definition of non-linear waves, such as tidal waves, with several Fourier components.
The accuracy of the linear wave model for large-amplitude waves has been improved by removing some of the simplifying assumptions.
In another improvement to the wave boundary condition, a uniform velocity component normal to the boundary is automatically computed and added at the boundary to compensate for the net flow into the domain that occurs over a wave period. With this addition, multiple waves can be introduced over hundreds of periods without a significant change in the fluid volume within the computational domain.
Counter-rotating flow at inlets
In a centrifugal casting process the metal is poured into a rotating mold. Such processes are often modeled in FLOW-3D using the Non-inertial Reference Frame model where the mesh coordinate system is attached to and rotates with the mold. In such a setup the incoming metal has a rotational velocity component opposite to the rotation of the mold, which previously was not possible to define in a Cartesian coordinate system. The new counter-rotating flow feature automatically applies this flow component at any inlet mesh boundaries, providing more realistic flow conditions.
Volume flow rate boundary condition
A volume flow rate boundary condition has been added to the list of standard mesh boundary conditions. The flow rate can be defined as a piecewise linear function of time at any of the six mesh boundaries. It can also be combined with fluid height at the four vertical mesh boundaries. The flow rate is converted to a uniform fluid velocity using the total open wetted area at the boundary. In addition, a flow direction vector can be specified in case the required flow direction is not normal to the boundary.
Optional fluid inflow at outflow boundaries
Despite the name outflow, it is sometimes desirable to allow the fluid to re-enter the computational domain during the simulation, if only momentarily. The most common example is when modeling linear surface waves in an otherwise quiescent fluid. Since the horizontal component of the fluid velocity in the wave oscillates back and forth, it is important to allow the same motion at the outflow boundary so that the wave energy can be properly transmitted through it without deflection, at the same time helping to preserve fluid volume n the domain. In most other situations, no flow should be allowed to enter at outflow boundaries. The users are now able to choose between these two options when defining the outflow boundary condition, with the no re-enter option being the default selection.
Fluid droplet sources
A new type of fluid source has been introduced in this version: fluid droplet source. A fluid droplet source generates spherical fluid droplets at a fixed location and rate. The droplets are emitted with a uniform initial temperature, density, pressure and velocity. Multiple sources, each with its own rate, droplet size and initial conditions can be defined. A source can emit Fluid #1 or #2 in two-fluid problems, as well as bubbles in one-fluid cases. Obviously, the mesh must be sufficiently fine to resolve the droplets or the bubbles along their trajectories. This addition can be employed for instance in droplet coating and droplet welding applications.
Gap formation in thermal stress model
The formation of a gap between solidifying metal and mold during the deformation of the part during cooling has been coupled with the heat transfer between the two media. The heat transfer across the gap can now include both radiation and thermal conduction in the gas (e.g., air) in the gap.
New solver output
A number of new output quantities have been added to the solution data file flsgrf. All spatial quantities are optional and require user抯 action to be stored, while the other quantities are always included provided the associated models are active.
Fluid volume within each porous component
The total volume of fluid #1 within each porous component is automatically computed and stored in the General history data catalogue as a function of time when a porous medium model is used.
Electric charge within each component
The total amount of charge within each conducting component is automatically computed and stored in the General history data catalogue as a function of time when the electric charge model is used.
Thermal power generated in each component by Joule heating
The total amount of heat per unit time generated by each solid component due to Joule heating is automatically computed and stored in the General History data catalogue as a function of time when the Joule heating model is used.
Vorticity
The vorticity vector is computed in every computational cell with fluid and stored as a spatial variable. Each of the components and the magnitude of the vorticity vector are available for contouring 2D and 3D plots. The vorticity output can be requested in the Output tab of the GUI whenever the full fluid flow model is used.
Hydraulics data: free surface elevation, fluid depth and Froude number
As an extension of the existing capability to plot surface height, the fluid surface elevation and depth (in the z direction), and the Froude number (based on the depth-averaged horizontal velocity) can now be computed as a function of x and y coordinates and stored as spatial variables when the user chooses to do so in the Output tab. The data can be used for color shading in 2D and 3D plots. This capability is most useful in the hydraulics applications of FLOW-3D.
Elastic membranes and walls
A limited fluid-structure interaction (FSI) capability is now available in FLOW-3D. It currently consists of two basic models: an elastic membrane and elastically flexing wall. The main assumption in both models is that the deformations in response to external forces are small compared to the lateral extents of the deforming objects. In this case, these deformations can be described by analytical functions. The deformations impact the motion of the adjacent fluid, while pressure in the fluid in turn affects the deformations. These interactions are described in the model in a fully coupled fashion.
The model is accessed through the Moving and deforming objects dialog in Physics and Deforming Components in Meshing & Geometry. See Flow Science Technical Note #79 for details.
Thermally active volume of geometry
The unstructured memory technique introduced in version 9.2 is a major benefit for users in the casting industry and elsewhere with its drastic reduction to solver memory and CPU time requirements. However, when the full fluid/solid heat transfer model was used, the memory savings were lost, as all computational cells were required to be 揳ctive.?In version 9.3, a user can now define the Maximum thermal penetration depth for each solid component under Solid Properties in the Geometry tree in the Meshing & Geometry tab. The thermal solution is then obtained within the volume along the component抯 open surface bounded by this depth. Everything outside of this volume is removed from the calculation resulting in faster simulations.
Casting users will benefit greatly from this feature. For example, during die casting filling, heat penetrates the die by only several millimeters and, therefore, modeling the whole die is unnecessary. If the user defines the thermally active layer thickness to be, for example, 5 mm, only cells within that layer are retained in the calculation, potentially reducing the total number of cells in the model several fold.
Residue model
The residue from dried liquid drops has implications for many useful applications, including coating processes, formation of pixel arrays of organic materials for video displays and for a variety of micro-electro-mechanical (MEMS) devices. A new residue model in version 9.3 allows users to model the formation of a solid residue during the drying process (i.e. the 揷offee ring?effect) and investigate the influence of such parameters as the initial solute concentration, fluid viscosity, volatility of the solvent, evaporation rate, surface tension and initial shape of the drop.
The model is accessed by directly editing the prepin file in the text editor. See Flow Science Technical Note #80 for details.
Two-component compressible gas with evaporation/condensation
The two-fluid, liquid/vapor phase change model has been extended to include a non-condensable gas component, allowing the users to describe, for example, a system consisting of liquid and gaseous hydrogen and an inert gas such as helium.
The model is accessed through the Bubble and phase change dialog in Physics. See Flow Science Technical Note #78 for details.
Running simulations
Navigator and Workspaces
The new Navigator feature in the FLOW-3D allows the user to manage several simulations at the same time, including building models, running the solver and post-processing results. The concept of Workspaces is introduced to help the user better organize his work. A Workspace generally contains a group of simulations that may be related to the same development project, including restart simulations. Existing and new simulations can be added to a workspace where they may be edited, executed and post-processed ?all within the same GUI session. All simulations in a workspace can be executed sequentially at a click of a button. Simultaneous execution of several simulations is also possible, as long as the simulations are not located in the same folder. Multiple Workspaces can be created within a single GUI session.
The introduction of Navigator represents a new approach to setting up and running simulations in FLOW-3D, and we have included with the distribution a detailed 揾ow-to?guide for your benefit. IT IS STRONGLY RECOMMENDED THAT ALL USERS REVIEW THIS GUIDE BEFORE USING VERSION 9.3.
Runtime changes to solver parameters
It is common for a user to want to make a change to the numerical settings of a simulation to make it run more accurately and efficiently. There could be more than one reason to change the settings, e.g., poor convergence, a small time-step size, a mentor tip, or just an omission in the model setup. The new Runtime Options tool in the GUI抯 Simulate tab is designed to help in such situations. It allows the user to change most of the numerical options during the simulation without the need to stop or even pause it.
Simulation restarts
There are significant changes in version 9.3 to how restart simulations are carried out. The changes are designed to simplify and generalize the restart function and require that all restart simulations from the earlier versions of FLOW-3D be redefined in accordance with the description below:
Restart logic
The prepinr file naming convention for restart simulations is no longer used; the restart simulations have now the same naming rules as the regular ones, that is, all simulations are named prepin with a unique file name extension. The user needs to rename all old restart input files to prepin.<extension> before opening them in the GUI.
In the Restart dialog of the GUI抯 General tab, activate the restart model, select the Restart source file, then pick the Restart time. The Restart source file is used as the source of the initial data for the restart simulation and can be located in a folder other than the input file. These changes make it easier to perform multiple restarts from the same Restart source file, or consecutive restart simulations using Navigator.
Reset fluid velocity and pressure at restart time
The option to reset all fluid velocities to zero and pressure to the void pressure PVOID at restart time has been added. This is useful, for example, when modeling thermal stresses in a restart run from a filling simulation, where the residual metal velocity and pressure in the mold is not desired as the initial condition to the stress simulation.
Moving Object Model
The applications for the General Moving Object model have been expanding with every release as the model matures. With its extension to cylindrical coordinate systems, the need for the old linear moving obstacle model has been completely eliminated, and it has been removed from FLOW-3D (see Removed models).
The definition of the GMO motion has been simplified in the GUI by adding a separate Edit dialog next to each moving component in the Geometry tree in the Meshing & Geometry tab.
Below is the list of new additions to the General Moving Object model in version 9.3:
Moving object assemblies
Moving components can now be made of several materials characterized by their density. Each solid subcomponent that makes up a moving component can be assigned its own density. The mass properties are then automatically evaluated from the distribution of mass within the component. This will allow the users to model coupled motion of multi-material bodies.
Mass density for each subcomponent is defined under Subcomponent branch in Geometry tree in the Meshing & Geometry tab.
Reference point for six DoF motion
The prescribed translational and rotational motion of a moving object can now be defined about an arbitrary reference point, instead of only its mass center. Together with the definition of the actual mass center, this addition simplifies the definition of the six DoF prescribed motion.
Partial slip boundary condition
Viscous boundary conditions between fluid and moving objects have been extended to include partial slip, bringing it up to the same level of capability as that for the stationary objects.
Motion limiters for prescribed-motion objects
Linear and angular motion limiters can now be applied to prescribed-motion GMO components, and not only to the couple-motion ones. This will make it easier to make a prescribed-motion GMO stop at a desired location. For example, when modeling the shot sleeve in the high pressure die casting, the plunger can be easily stopped at the end of the sleeve. Note that the motion limiters are defined in terms of the total displacement - angular or linear - from the initial location.
Mesh boundary conditions
Several refinements and additions have been made to the mesh boundary condition models. Some are designed for specific applications, while others will be useful in general cases.
Linear surface waves
The linear wave boundary condition has been extended to allow for up to a hundred linear wave components to be generated at the same vertical mesh boundary. The purpose of this addition is to enable the definition of non-linear waves, such as tidal waves, with several Fourier components.
The accuracy of the linear wave model for large-amplitude waves has been improved by removing some of the simplifying assumptions.
In another improvement to the wave boundary condition, a uniform velocity component normal to the boundary is automatically computed and added at the boundary to compensate for the net flow into the domain that occurs over a wave period. With this addition, multiple waves can be introduced over hundreds of periods without a significant change in the fluid volume within the computational domain.
Counter-rotating flow at inlets
In a centrifugal casting process the metal is poured into a rotating mold. Such processes are often modeled in FLOW-3D using the Non-inertial Reference Frame model where the mesh coordinate system is attached to and rotates with the mold. In such a setup the incoming metal has a rotational velocity component opposite to the rotation of the mold, which previously was not possible to define in a Cartesian coordinate system. The new counter-rotating flow feature automatically applies this flow component at any inlet mesh boundaries, providing more realistic flow conditions.
Volume flow rate boundary condition
A volume flow rate boundary condition has been added to the list of standard mesh boundary conditions. The flow rate can be defined as a piecewise linear function of time at any of the six mesh boundaries. It can also be combined with fluid height at the four vertical mesh boundaries. The flow rate is converted to a uniform fluid velocity using the total open wetted area at the boundary. In addition, a flow direction vector can be specified in case the required flow direction is not normal to the boundary.
Optional fluid inflow at outflow boundaries
Despite the name outflow, it is sometimes desirable to allow the fluid to re-enter the computational domain during the simulation, if only momentarily. The most common example is when modeling linear surface waves in an otherwise quiescent fluid. Since the horizontal component of the fluid velocity in the wave oscillates back and forth, it is important to allow the same motion at the outflow boundary so that the wave energy can be properly transmitted through it without deflection, at the same time helping to preserve fluid volume n the domain. In most other situations, no flow should be allowed to enter at outflow boundaries. The users are now able to choose between these two options when defining the outflow boundary condition, with the no re-enter option being the default selection.
Fluid droplet sources
A new type of fluid source has been introduced in this version: fluid droplet source. A fluid droplet source generates spherical fluid droplets at a fixed location and rate. The droplets are emitted with a uniform initial temperature, density, pressure and velocity. Multiple sources, each with its own rate, droplet size and initial conditions can be defined. A source can emit Fluid #1 or #2 in two-fluid problems, as well as bubbles in one-fluid cases. Obviously, the mesh must be sufficiently fine to resolve the droplets or the bubbles along their trajectories. This addition can be employed for instance in droplet coating and droplet welding applications.
Gap formation in thermal stress model
The formation of a gap between solidifying metal and mold during the deformation of the part during cooling has been coupled with the heat transfer between the two media. The heat transfer across the gap can now include both radiation and thermal conduction in the gas (e.g., air) in the gap.
New solver output
A number of new output quantities have been added to the solution data file flsgrf. All spatial quantities are optional and require user抯 action to be stored, while the other quantities are always included provided the associated models are active.
Fluid volume within each porous component
The total volume of fluid #1 within each porous component is automatically computed and stored in the General history data catalogue as a function of time when a porous medium model is used.
Electric charge within each component
The total amount of charge within each conducting component is automatically computed and stored in the General history data catalogue as a function of time when the electric charge model is used.
Thermal power generated in each component by Joule heating
The total amount of heat per unit time generated by each solid component due to Joule heating is automatically computed and stored in the General History data catalogue as a function of time when the Joule heating model is used.
Vorticity
The vorticity vector is computed in every computational cell with fluid and stored as a spatial variable. Each of the components and the magnitude of the vorticity vector are available for contouring 2D and 3D plots. The vorticity output can be requested in the Output tab of the GUI whenever the full fluid flow model is used.
Hydraulics data: free surface elevation, fluid depth and Froude number
As an extension of the existing capability to plot surface height, the fluid surface elevation and depth (in the z direction), and the Froude number (based on the depth-averaged horizontal velocity) can now be computed as a function of x and y coordinates and stored as spatial variables when the user chooses to do so in the Output tab. The data can be used for color shading in 2D and 3D plots. This capability is most useful in the hydraulics applications of FLOW-3D.
点赞 评论 5 收藏
关注
查看更多评论 >