Data Transfer Mesh Mapping

To transfer data across a dissimilar mesh interface, the nodes of one mesh must be mapped to the local coordinates of a node/element in the other mesh. This section describes the settings that are available in Mechanical when data is mapped across two different meshes.

You can add the exported mesh and loads as external data in the project schematic and couple a new Mechanical analysis system with this external data. The Mapping Settings described below are available within Mechanical for Thermal-Stress coupling with dissimilar mesh, Submodeling, when temperatures or displacements are transferred from Mechanical to Ansoft, or when the source data comes from an External Data system.

Mapping Settings

The Settings category provides the following properties.

Mapping Control

By default, when Program Controlled is selected, the software will determine the appropriate algorithm and settings based on the source and target mesh data, as well as the data type being transferred. See Program Controlled Mapping topic below for additional information. You may choose to modify the advanced features by setting this to Manual.

Mapping

This read-only property displays the mapping algorithms the application selects. Options include:

  • Profile Preserving: Using this mapping option, the application simply takes the profile of the variable (for example, temperature) on one mesh, and matches or maps it to the other mesh as best as it can.

  • Conservative: Using this mapping option, the application makes sure that the profile is interpolated in such a way as to ensure that a total quantity passing across the interface is conserved, that is the same total passes out of one mesh and into the other. For example, a conservative interpolation of force ensures that the total force on one side of the boundary exactly matches the total force received by the other side of the boundary, even if the mesh resolution is poor. Conservative interpolation does not make sense for a variable such as temperature where there is no corresponding physical quantity to conserve.


    Note:  Conservative algorithms are only available for Imported Force loads. Conservative algorithms are not available for 2D to 3D data transfers. If conservative algorithms are not available, it is a read-only field displaying that a "Profile Preserving" algorithm is being used.


Weighting

Choose which type of weighting should be performed. This option can be changed only if Mapping Control is set to Manual.

  • Direct Assignment applies the source value directly on the nodes/elements identified by Node IDs/Element IDs in the External Data specification.


    Note:  This mapping only supports loads applied to nodes or elements. The following loads applied to element faces are not supported:

    • Pressure (on element faces)

    • Convection Coefficient

    • Heat Flux


  • Triangulation creates temporary elements from the n closest source nodes to find the closest points that will contribute portions of their data values. For 3D, 4-node tetrahedrons are created, and for 2D, 3-node triangles are created by iterating over all possible combinations of the source points (maximum number controlled by the Limit property), starting with the closest points. If the target point is found within the element, weights are calculated based on the target’s location inside the element.

  • Distance Based Average uses the distance from the target node to the specified number of closest source node(s) to calculate a weighting value.

  • Weighting:

    • Shape Function: Two mapping methods are available for a load transfer: “Profile Preserving” and “Conservative”.

      In a Profile Preserving mapping, each node on the target (receiver) side maps onto an element on the source (sender) side (α1). The transfer variable is then interpolated at α1. The transfer value is T1 = φ (α1). Thus, all nodes on the target side query the source side.

      Figure 34:  Profile Preserving Mapping

      Profile Preserving Mapping

      In a conservative mapping, each node X on the source (sender) maps onto an element on the target (receiver) side. Thus, the transfer variable on the source is split into two quantities that are added to the target nodes. As shown in the following figure, the force at node 4 splits into forces at nodes 3’ and 4’.

      Figure 35:  Conservative Mapping

      Conservative Mapping

      Thus profile preserving (conservative) version of Shape Function algorithm loops over the target (source) nodes and tries to locate a source (target) element that each target (source) node can be mapped to. Weights for each of the source nodes are then assigned based on the location of the target (source) node and the shape function of the element. For each target (source) node, the search efficiency can be improved by restricting the search to a subset of the source (target) elements. The search algorithm works by:

      • Distributing all source (target) elements into Cartesian boxes or buckets. The number of buckets is controlled by the Scale property.

      • Locating each of the target nodes in a box

      • Finding an element that each target (source) node can be mapped to by restricting the search with each target’s (source’s) box


      Note:
      • When there is a significant distance between target (source) node and the closest element, e.g. Shell-Solid submodeling, the node and the element may not be found in the same box. In order to improve mapping accuracy in such cases, the Pinball control may be used. See Pinball in the Advanced section for more details.

      • For conservative mapping, the value on a source node is distributed only on the nodes of the target element it is mapped to. Therefore, it is possible, especially if you are mapping from coarse to fine meshes, that some of the mapped target nodes get a zero value. This is because none of the adjacent elements are mapped to one or more source nodes.


    • Kriging is a regression-based interpolation technique that assigns weights to surrounding source points according to their spatial covariance values. The algorithm combines the kriging model with a polynomial model to capture local and global deviations. The kriging model interpolates the source points based on their localized deviations, while the polynomial model globally approximates the source space. See Kriging Algorithms in the ANSYS DesignXplorer User's Guide for more information.


      Note:  By default, the Kriging technique uses an adaptive algorithm and ensures that the interpolated values do not exceed specific limits. The adaptive algorithm starts by using the higher-order Cross Quadratic polynomial to interpolate data. If the interpolated value of each target point is outside the extrapolation limit you specified, the algorithm re-interpolates data by reducing the polynomial order and the number of source points. Target nodes whose values are outside the limits when the lowest polynomial type is used are not assigned a value.

      The Kriging algorithm, when used with the higher-order Cross Quadratic or Pure Quadratic polynomial, may fail to correctly interpolate data for a target point if multiple source points are spaced close to one another or if the target point is outside the region enclosed by the source points that are selected for interpolation. This may introduce gross errors in the estimation of the target value and manifests itself mostly when mapping data on surface or edge geometries. In such cases, you should change the Polynomial Type to Constant or Linear and, if necessary, reduce the number of source points to be included for the interpolation.


    • UV Mapping can be used to transfer data from one surface to another. Unlike other algorithms, UV mapping does not require the surfaces to be coincident. This allows for mapping between deformed and un-deformed geometries, as well as transfers between dissimilar geometry. Element data is required from both the source and the target mesh. If the source is an Mechanical APDL .cdb file containing volumetric element data, a nodal component must also be specified which will be used to define the surface from which the data transfer will occur.

      Polyhedral Surface Creation and Conversion to UV

      To map a mesh in UV space, the application first creates polyhedral surfaces from the given mesh data. If the source mesh is volumetric data, an associated node-based component must be selected such that the nodes consist of the surface area where the mapping takes place. Once the application creates the source and target surfaces, they are then ‘unfolded’ and converted into UV coordinates. The application defines the UV space as a parametric space where the axis data equals 0.0 to 1.0. Alignment points anchor the node locations to the corners of the 1x1 box.

      Interpolation

      Once the source and target data is converted to UV space, the target nodal UV locations are used to locate the source element that would contain the target node. The value for the target is then calculated based on the values provided from the source elements nodes.


    Note:  Available weighting options depend upon the data available from source and target and on the Mapping setting. Some of the weighting options may not be available for certain mesh data or Mapping settings. For example, when Mapping is set to Profile Preserving, Shape Function and UV are only available when the source provides element information. For Conservative Mapping, only Shape Function for Surface transfers is available.


Transfer Type

Enables you to choose the dimension of the transfer (for 3D profile preserving transfers only). This option is available only for Triangulation, Shape Function, and for adaptive Kriging. For best results, use the Surface option when mapping data across surfaces and the Volumetric option when mapping data across volumes.

When used with Triangulation:

  • The Surface option tries to map each target point by searching triangles that are created from the set of closest source points. The target point will be projected onto the plane relative to the triangle surface. If the point is found inside the triangle, the weights are calculated based on the target’s projected location inside the triangle.

  • The Volumetric option tries to map each target point by searching tetrahedrons that are created from the set of closest source points.

When used with the Shape Function:

  • The Surface option uses the bucket surface search algorithm to locate a source element that each target node can be mapped to. This option supports only triangle and quadrilateral source elements; do not use it if your source elements are other shapes because the algorithm does not account for these shapes.

  • The Volumetric option uses the bucket volume search algorithm to locate a source element that each target node can be mapped to. This option supports triangle, quadrilateral, tetrahedron, hexahedron, and wedge source elements.

When used with adaptive Kriging, the Surface option uses fewer surrounding source points to interpolate data than the Volumetric option does.

2D Projection

Available only for 2D to 3D data transfers from an External Data system connected to Mechanical. The default option is Normal To Plane. You will be able to choose between the default as well as all application and user input coordinate systems.

Rigid Transformation Controls

Rigid transformation properties enable you to apply a coordinate transformation to the source points. Two options are available through the Mesh Alignment property:

Use Origin and Euler Angles: The source locations are transformed by the coordinate system defined by the Origin and Theta entries. For example, applying a value of .1 meters to Origin X would modify the x locations of all the source points by adding .1 meters to their values.

Use Coordinate Systems: To use this option, choose two coordinate systems, (1) Source Coordinate System attached to the source mesh frame of reference and (2) Target Coordinate System attached to the target mesh frame of reference. The transformations are automatically calculated such that the Source Coordinate System is aligned with the Target Coordinate System after transformation. For example, when the source mesh is defined in the XY plane, whereas the target geometry is defined in a plane obtained by applying the Euler rotations RXY, RYZ and RZX to the XY plane. Then choosing Global Coordinate System as Source Coordinate System and the coordinate system created by applying the transformations RXY, RYZ and RZX to the Global Coordinate System as the Target Coordinate Systems, the source mesh is transformed such that it is aligned with the target geometry. This option is useful if the source points are defined with respect to a coordinate system that is not aligned with the target geometry system.

The option Display Source Points on an Imported Load or Imported Thickness object inside Mechanical respects this transformation and can be very helpful in ensuring proper alignment between the source and target points.

Graphics Controls

The Graphics Controls category provides the following properties.

Legend Controls

The Legend Controls category provides the following properties.

Named Selection Creation

The Named Selections category provides the following properties.

These settings are only available for profile preserving mapping.

Advanced

The application filters the properties of the Advanced category based on the settings made in the Mapping Control and Weighting properties in the Mapping Settings category. Properties include:

Advanced Shell-Solid

Advanced shell-solid settings are filtered based on the Mapping Control and Weighting type selected in Mapping Settings. They are only available for Shell-Solid submodeling. In the case of imported cut boundary conditions, Shape Function is the only available Weighting type.

Pinball Factor: This value is used to calculate the Pinball Value for shell-solid submodeling. The Pinball Value is calculated by scaling the maximum shell thickness with the Pinball Factor.

Figure 40:  Shell-Solid Submodeling with Pinball Factor = 1.0

Shell-Solid Submodeling with Pinball Factor = 1.0

Figure 41:  Shell-Solid Submodeling with Pinball Factor = 1.2

Shell-Solid Submodeling with Pinball Factor = 1.2

As shown in Figure 40: Shell-Solid Submodeling with Pinball Factor = 1.0 and Figure 41: Shell-Solid Submodeling with Pinball Factor = 1.2, the gap between the nodes in the filleted region is greater than the maximum shell thickness for the model. Hence using a Pinball Factor equal to 1 results in nodes in the fillet not finding appropriate matching elements.(1) When Pinball Factor of 1.2 is used, then additional buckets are included in the search resulting in better mapping results.(2)


Note:   Increasing the Pinball Factor increases the number of buckets searched to find the matching element hence, may decrease the efficiency of the mapping. An appropriate value should be chosen so that the resulting bounding region includes the matching element but not too big so as to negatively affect the efficiency of the search.

Shell Thickness Factor: For shell models with variable thickness, the gap between the target node, and matching element may be large. Shell Thickness Factor is used to exclude any matching element which has a gap greater than Thickness* Shell Thickness Factor. Thickness is the average element thickness of the matching element.


Figure 42:  Shell-Solid Submodeling with Shell Thickness Factor = 0.6

Shell-Solid Submodeling with Shell Thickness Factor = 0.6

Figure 43:  Shell-Solid Submodeling with Shell Thickness Factor = 1.2

Shell-Solid Submodeling with Shell Thickness Factor = 1.2


Note:  Increasing the Shell Thickness Factor to allow submodel nodes to be “found” can produce poor submodel results as shown in Figure 42: Shell-Solid Submodeling with Shell Thickness Factor = 0.6 and Figure 43: Shell-Solid Submodeling with Shell Thickness Factor = 1.2. where large Shell Thickness Factor causes the target nodes on the web region to be matched with the base (3), whereas the target nodes are more appropriately matched for a smaller Shell Thickness Factor (4).


UV Source Controls/UV Target Controls

Program Controlled Mapping

Refer to the following table for the appropriate settings for when you set the Mapping Control property to Program Controlled in order to determine which type of mapping algorithm to use. The application determines default settings based on the properties described above.

When Mapping is Profile Preserving:

Source mesh can provide:Target mesh can provide:Weighting that will be used:
Node IDs OnlyNodesUses Direct Assignment to assign values to target nodes.
Element IDs OnlyElementsUses Direct Assignment to assign values to target elements.
Nodes OnlyNodes OnlyUses Triangulation to calculate mapping data.
Nodes and ElementsNodes OnlyUses Shape Function to calculate mapping data.

When Mapping is Conservative

Source mesh can provide:Target mesh can provide:Weighting that will be used:
Nodes OnlyNodes and ElementsUses Shape Function to calculate mapping data.

Manual Mapping

When manual mode is selected, you will be able to control advanced settings for the mapper. Based on the mapping chosen (conservative or profile preserving) and mesh data provided from the source and target, you will be able to choose the type of weighting algorithm.

For Profile Preserving Mapping:

If the source mesh contains only points, you will be able to select from the following:

If the source mesh also contains element data, you will have the items listed above as well as:

For Conservative Mapping:

If the source mesh contains only points, you will be able to select from the following:

Element shapes supported during mapping when Shape Function is selected:

Element ShapeSupported
3 Node TriangleX (2D) (3D)
6 Node TriangleX (2D) (3D)
4 Node QuadrilateralX (2D) (3D)
8 Node QuadrilateralX (2D) (3D)
4 Node TetrahedronX (3D)
10 Node TetrahedronX (3D)
8 Node HexahedronX (3D)
20 Node HexahedronX (3D)
6 Node WedgeX (3D)
15 Node WedgeX (3D)

2D to 3D Mapping

Mapping point data from 2D to 3D analyses is possible using the External Data system connected to a downstream Mechanical system. This mapping is performed by collapsing the 3D mesh data into a 2D plane and calculating target point weighting factors from the source point data.

2D results in the XY Plane:

You will be able to select the 2D project plane to use based on the available coordinate systems as well as an option to select normal to the 2D source point data (Normal To Plane). Using the Graphics Controls described above, you will be able to turn on and off visualization of the source point data and the 2D projection plane.

Source point and 2D projection plane displayed:

When selecting Cartesian coordinate systems, the projection will be done on the XY Plane. If the coordinate system is cylindrical, the projection will be rotated about the Z axis into the ZX Plane. Normal To Plane will project the target points into the source point plane.

3D mapped data using cylindrical coordinate system projection:

3D mapped data using Normal To Plane:


Note:  Conservative mapping is not available for 2D to 3D transfers.


Notes

When mapping point cloud data, the mapping utility does not know where body boundaries are. If you have a model with contact between two bodies, the mapping may pick up points from both bodies causing undesired results.


Release 18.2 - © ANSYS, Inc. All rights reserved.