Breakdown of Part
The initial part has been broken down into subparts as follows, see Figures 17.17 and 17.18:
- 1. wing spar tube
- 2. boom tube
- 3. skin
- 4. outer rib
Figure 17.17 The input geometry modeled as partitioned parts.
Figure 17.18 The assembly generated from reference geometry.
- 5. inner rib
- 6. foam insert spigot
- 7. boom tube fairing
- 8. skin stringer stiffening.
Ideally, the detailed geometry should automatically reconfigure/scale to any changes to the input reference geometry and other explicitly defined input variables/dimensions. In practice, it is very hard to achieve this from manually constructed models. An alternative way of achieving this is to capture all the logic and associated relationships in software code and use this to fully generate the detail model. This requires a substantial amount of effort, and for component topologies/functions that are repeatedly modified and reused, this can be worth the effort. However, this is beyond the scope of this example, and in this model manual construction has nevertheless been used. With care, manual models can, within limits, be made parametric. In other words, they can be made sufficiently robust to allow a limited range of changes to the inputs. Where manual models most frequently fail is when changes to inputs cause features to be eliminated. This then causes errors in the relationships between parts and other features.
To test the initial robustness of this emerging detailed model, a scaling feature has been inserted into the initial input reference geometry. This simulates the changes that might be made to the conceptual design. The scaling feature can be used to make global changes to the input geometry (e.g., making the entire geometry 20% larger). It can also be used to “stretch” the geometry in less than three dimensions. Hence, you could choose to stretch the length of the part in only one direction. The problem with this is that it can cause the downstream model to fail because circular features become ellipses, for example. For this reason, testing of this model was restricted to global 3D scaling tests.
By scaling the input reference geometry, the detailed geometry model can be “debugged” to ensure that it faithfully responds to the changes without generating errors. This is shown in Figure 17.19, which shows that the input reference geometry has downscaled by 30% (inner)
Figure 17.19 “Debugging” the detailed model.
Figure 17.20 Trimming of boom tube fairing.
and the previously generated detailed geometry (outer) has not yet been rebuilt. On pressing the rebuild button, successful scaling of the detailed model can be checked and any errors understood and fixed.
A number of steps now need to be carried out, including the following:
- • Modification of internal geometry (see, e.g., trimming of the boom tube fairing, Figure 17.20);
- • Ensuring that the part is capable of being manufactured; in the case of the nylon SLS process, closed voids need to be eliminated;
- • Editing of part geometry to meet functional goals consistent with the process capabilities of the selected manufacturing method - in other words, the geometry needs to allow parts to fit together;
- • Addition of further detail that does not exist in the conceptual input geometry.
All parts can now be edited as independent objects. The advantages of this compared with creating the detail model as one monolithic or multibody part are twofold. First, it allows very large and complex geometries to be manipulated more easily. Very large, complex geometries can create very large file sizes. It can be tedious having to load a very large geometry when the user only wants to modify a small feature. Second, it allows a model to be decluttered quickly and easily. Just having to open the item you are interested in allows you to focus only on that part of the geometry. A disadvantage of using this approach is where relationships between many “parts” are required. This may require the whole assembly to be loaded and only the relevant “parts” made visible.