DTPL

Format

Example 1

Example 2

Definitions

Comments

1. The von Mises stress constraints may be defined for topology and free-size optimization through the STRESS optional continuation line on the DTPL or the DSIZE card. There are a number of restrictions with this constraint:
   • The definition of stress constraints is limited to a single von Mises permissible stress. The phenomenon of singular topology is pronounced when different materials with different permissible stresses exist in a structure. Singular topology refers to the problem associated with the conditional nature of stress constraints, that is, the stress constraint of an element disappears when the element vanishes. This creates another problem in that a huge number of reduced problems exist with solutions that cannot usually be found by a gradient-based optimizer in the full design space.
   • Stress constraints for a partial domain of the structure are not allowed because they often create an ill-posed optimization problem since elimination of the partial domain would remove all stress constraints. Consequently, the stress constraint applies to the entire model when active, including both design and non-design regions, and stress constraint settings must be identical for all DSIZE and DTPL cards.
   • The capability has built-in intelligence to filter out artificial stress concentrations around point loads and point boundary conditions. Stress concentrations, due to boundary geometry are also filtered to some extent as they can be improved more effectively with local shape optimization.
   • Due to the large number of elements with active stress constraints, no element stress report is given in the table of retained constraints in the .out file. The iterative history of the stress state of the model can be viewed in HyperView or HyperMesh.
   • Stress constraints do not apply to 1D elements.
   • Stress constraints may not be used when enforced displacements are present in the model.
     Note: The functionality of the STRESS continuation line to define topology stress constraints consists of many limitations. It is recommended to use DRESP1-based Stress Responses. Actual Stress Responses for Topology and Free-Size (Parameter) Optimization are available through corresponding Stress response RTYPE's on the DRESP1 Bulk Data Entry. The Stress-NORM aggregation is internally used to calculate the Stress Responses for groups of elements in the model.
2. It is recommended that a MINDIM value be chosen such that it is at least 3 times, and no greater than 12 times, the average element size. When pattern grouping, draw direction, or extrusion constraints are active, a MINDIM value of 3 times the average element size is enforced, and user-defined values (which are smaller than this value) will be replaced by this value. However, in cases where a MINDIM greater than 12 times the average element size is defined, irrespective of whether, or not other manufacturing constraints are defined, the value is reset to be equal to 12 times the average element size. If DOPTPRM,TOPDISC is present in the model, a MINDIM value equal to 2 times the average element size is enforced.

   If MINDIM is defined, but no other manufacturing constraint exists, MINDIM will not be reset to the recommended lower bound value for PTYPE=PSHELL or PSOLID, if the defined value is less than the recommended value. For PTYPE=PCOMP, MINDIM will be reset in the absence of manufacturing constraints.

3. MAXDIM should be at least twice the value of MINDIM. If the input value of MAXDIM is too small, OptiStruct automatically resets the value and an INFORMATION message is printed.

   The MAXDIM constraint introduces significant restriction to the design problem. Therefore, it should only be used when it is a necessary design requirement. A study without MAXDIM should always be carried out in order to compare the impact of this additional constraint.

   MAXDIM implies the application of a MINGAP constraint of the same value as MAXDIM, as well. Therefore, for MINGAP to be effective, it should be greater than MAXDIM.

   It is important to pay attention to volume fraction as the achievable volume is below 50% when MAXDIM is defined, and further decreases as MINGAP increases.

4. MTYP=ALIGN may be used in conjunction with draw direction or extrusion manufacturing constraints to indicate that a mesh is aligned with a draw direction or extrusion path.

   
   
   Figure 1. Draw Direction

   Mesh 1 is "aligned" for draw direction 1 in the example shown, but not for draw direction 2.

   MTYP=ALIGN may also be used in conjunction with manufacturing constraints (minimum member, maximum member, pattern grouping, and pattern repetition) other than draw direction and extrusion, and Mesh 1 is considered "aligned" for those manufacturing constraints, too.

   In both cases, this will enable OptiStruct to use a smaller minimum member size and smaller maximum member sizes. The default minimum member size is three times the average element edge length; with an "aligned" mesh, the default size can be two times the average element edge length.

   Mesh 2 in the example shown is not "aligned" in any case.

5. The stamping constraint is available for only one sheet, which is defined by the combination of STAMP and DTYP as SINGLE.

   It is recommended that the stamping thickness, TSTAMP, be chosen such that it is at least 3 times the average element size. If TSTAMP is defined less than the minimum recommended value, TSTAMP will be reset to the minimum recommended value.

   STAMP and NOHOLE can be a good combination as this helps to produce a continuous/spread shell structure.

   Note: Attention should be paid to the compatibility between thickness and target volume.
6. Extrusion constraints cannot be combined with draw direction constraints.
7. Pattern repetition allows similar regions of the design domain to be linked together so as to produce similar topological layouts. This is facilitated through the definition of "Master" and "Slave" regions. A DTPL card may only contain one MASTER or SLAVE flag. Parameters will not be exported for any DTPL cards containing the SLAVE flag. For both "Master" and "Slave" regions, a pattern repetition coordinate system is required and is described following the COORD flag. To facilitate reflection, the coordinate system may be a left-handed or right-handed Cartesian system. The coordinate system may be defined in one of two ways, listed here in order of precedence:
   • Four points are defined and these are utilized as follows to define the coordinate system (this is the only way to define a left-handed system):
     • A vector from the anchor point to the first point defines the x-axis.
     • The second point lies on the x-y plane, indicating the positive sense of the y-axis.
     • The third point indicates the positive sense of the z-axis.
   • A rectangular coordinate system and an anchor point are defined. If only an anchor point is defined, it is assumed that the basic coordinate system is to be used.

   Multiple "Slaves" may reference the same "Master."

   Scale factors may be defined for "Slave" regions, allowing the "Master" layout to be adjusted via the SCALE continuation line.

   For a more detailed description, refer to Pattern Repetition contained within the User Guide section Topology Optimization Manufacturability.

8. Pattern grouping is applicable for PCOMP, PSHELL, and PSOLID components only.
9. For historic reasons, the SYMM flag may be used in place of the PATRN flag.
10. Currently there are six pattern grouping options:

    1-plane symmetry (TYP=1)

    This type of pattern grouping requires the anchor point and first point to be defined. A vector from the anchor point to the first point is normal to the plane of symmetry.

    2-plane symmetry (TYP=2)

    This type of pattern grouping requires the anchor point, first point, and second point to be defined. A vector from the anchor point to the first point is normal to the first plane of symmetry. The second point is projected normally onto the first plane of symmetry. A vector from the anchor point to this projected point is normal to the second plane of symmetry.

    3-plane symmetry (TYP=3)

    This type of pattern grouping requires the anchor point, first point, and second point to be defined. A vector from the anchor point to the first point is normal to the first plane of symmetry. The second point is projected normally onto the first plane of symmetry. A vector from the anchor point to this projected point is normal to the second plane of symmetry. The third plane of symmetry is orthogonal to both the first and second planes of symmetry, passing through the anchor point.

    Uniform Pattern Grouping (TYP=9)

    This type of pattern grouping does not require any additional input. It only requires the TYP field to be set equal to 9. All elements included in this DTPL entry are automatically considered for uniform pattern grouping. All elements on this DTPL entry are set equal to the same element density with respect to one another.

    Cyclic (TYP=10)

    This type of pattern grouping requires the anchor point, first point, and number of cyclical repetitions to be defined. A vector from the anchor point to the first point defines the axis of symmetry.

    Cyclic with symmetry (TYP=11)

    This type of pattern grouping requires the anchor point, first point, second point, and number of cyclical repetitions to be defined. A vector from the anchor point to the first point defines the axis of symmetry. The anchor point, first point, and second point all lay on a plane of symmetry. A plane of symmetry lies at the center of each cyclical repetition.

    For a more detailed description, refer to Pattern Grouping Options contained within the User Guide section Topology Optimization Manufacturability.

11. The LT field can be used to specify the lattice type used in Lattice Structure Optimization.
12. The density thresholds are defined using the LB and UB fields on the LATTICE continuation line. Elements with densities below LB (real) are considered voids and removed for the second phase. Elements with densities above UB (real) are considered solid and are retained as solid elements for the second phase. Elements with densities between LB and UB are considered as porous phases and elements having these densities are replaced by lattice structures. The amount of intermediate densities (between 0.0 and 1.0) is controlled using DOPTPRM, POROSITY. Refer to Lattice Structure Optimization in the User Guide for further information.
13. FailSafe topology optimization runs in SPMD mode and requires the -fso script option. Refer to Failsafe Topology Optimization in the User Guide for further information.
14. Multi-Model Optimization requires the definition of the COORD continuation line to allow mapping of the design domains among multiple models. If individual pattern repetition is defined on all models, then this continuation line is not required as the COORD data from the pattern repetition section is used instead. The coordinate system can be defined in one of two different ways:
    • Four points are defined, and these are utilized as follows to define the coordinate system (this is the only way to define a left-handed system):

      A vector from the anchor point to the first point defines the x-axis.

      The second point lies on the x-y plane, indicating the positive sense of the y-axis.

      The third point indicates the positive sense of the z-axis.

    • A rectangular coordinate system and an anchor point are defined. If only an anchor point is defined, it is assumed that the basic coordinate system is to be used.
15. Both solids and shells are supported in multiple materials topology optimization. The following two limitations apply to PSHELLs in the design space for multiple materials topology optimization.
    • For any PSHELL entry part of the design space, then all material reference fields (MID# fields) on each PSHELL entry should point to the same material entry.
    • Additionally, if multiple PSHELL entries are part of the design space, then all MID# fields on all PSHELL entries should point to the same material entry.
16. The original material defined by its property will be taken as one of the candidate material by default. Besides the original one, the maximum number of candidate materials is nine (9). Only isotropic material MAT1 is supported on the MID# fields.
17. The Rational Approximation for Material Properties (RAMP) method uses the following equation for penalization.(1)
    $$\tilde{K}\left(\rho \right)=\left(\frac{\rho }{1+p\left(1-\rho \right)}\right)K$$
    Where,

    $\tilde{K}\left(\rho \right)$

    Penalized stiffness matrix of an element (as a function of density)

    $K$

    Actual stiffness matrix of an element

    $\rho$

    Density

    $p$

    Penalization factor

    For information about Solid Isotropic Material with Penalty (SIMP) method, see the Design Elements section.

18. For overhang constraints, STEP=1 allows aggressive move limits and typically converges fast. It generally produces good results for a majority of situations. However, it may show large fluctuations in convergence. In such cases, STEP=2 can be tried, which moves conservatively and follows a smoother convergence curve. It may sometimes offer faster convergence and better designs.
19. This card is represented as an optimization design variable in HyperMesh.
20. If the METHOD field is set to CONSTR, then the following considerations are available:
    • Depending on the model, CONSTR method can pose a significant reduction of the design freedom for the optimization. Which may lead to a reduction in performance compared to the run without overhang constraints.
    • If a volume or mass constraint is used in addition to the overhang constraint with CONSTR, then the target may be too low for the optimizer to find a good design. In such cases, you can try increasing the volume or mass target.
    • If the impact on the performance or the volume/mass target is too large, then try the PENALTY method.
    If the METHOD field is set to PENALTY, then the following considerations are available:

    • This method is good at removing members that are overhanging but have a small or medium impact on the performance. This method may not remove members which are very important to performance. If the goal is to remove such high impact members, try the CONSTR method instead.
    • For PENALTY method, the violations of overhang angle are output to H3D file.
21. The ANGTOL and DISTOL fields can be used to define elements that are considered to always be supported during optimization. Candidate elements are all those elements in the first layer encountered when traveling in the build direction. If this layer is inclined more than ANGTOL, and lies within DISTOL, it is always supported.
    The elements which are always supported are output to the H3D file under the “Predefined Support” results type. For the actual manufacturing of the part, some sections of this predefined support might require support structure.

    
    
    Figure 2.

22. LATPRM,LATSUP can be used to define the maximum volume fraction of the lattice support regions when overhang constraints are used in Topology Optimization.
23. The level set method can merge existing holes but cannot nucleate new holes in the design domain. Therefore, creating an initial design with holes is necessary, especially for 2D design problems (For 3D design problems, new holes can be “tunneled” when two surfaces merged).
24. By default, OptiStruct will automatically create a Cheese-like initial design with holes adaptively distributed over the design domain (Figure 3) The default hole radius is 4.0 times the average mesh size.

    
    
    Figure 3. A Cheese-like Initial Design Generated . (left) with default setting; (right) double hole radius

25. Changing the value of HOLERAD can result in different initial designs. Figure 3(right) shows an initial design filled with holes possessing a doubled hole radius when compared to Figure 3(left). If you want to create an initial design with evenly distributed and well aligned holes (this may be preferable for regular design domains), HOLEINST can be set to ALIGN. The number of holes in each direction can be further specified by using NHOLESX, NHOLESY and NHOLESZ (Figure 4).

    
    
    Figure 4. A Cheese-like Initial Design with 3-by-5 Evenly Distributed Holes . (generated using: HOLEINST=ALIGN, NHOLESX=5 and NHOLESY=3)

26. Currently, level set supports both SINGLE and SPLIT draw direction constraints. When multiple DTPL cards are involved, the draw directions need to be the same. The information needed for draw direction constraint is read from the DTPL cards and thus, no extra settings are required.
27. This card is represented as an optimization designvariable in HyperMesh.