Phase 3: Ply Stacking Sequence Optimization

This algorithm is aimed at providing a global view of what the optimal stacking sequence could be. An input deck for the ply stacking sequence optimization, fairing_size_shuffling.*.fem, was generated from a previous design stage. Each ply bundle is divided into multiple PLYs whose thickness is equal to the manufacturable thickness (0.1 in this case), and the STACK card is updated accordingly. In this design phase, composite plies are shuffled to determine the optimal stacking sequence.

It is important that design performances are preserved. Hence, the optimization problem is retained as previously formulated in the size optimization phase. Two manufacturing constraints are applied:
  • The maximum successive number of plies of a particular orientation does not exceed 4 plies.
  • The + 45s and - 45s are reversed paired.

Import the Model

  1. Click File > Import > Solver Deck.
    An Import tab is added to your tab menu.
  2. For the File type, select OptiStruct.
  3. Select the Files icon files_panel.
    A Select OptiStruct file browser opens.
  4. Select the fairing_size_shuffling.*.fem file you saved to your working directory. Refer to Access the Model Files.
  5. Click Open.
  6. Click Import, then click Close to close the Import tab.

Set Up the Optimization

Update the Composite Strain Response

Since the ply bundles were divided into multiple plies in the shuffling model, the ply information in CSTRAIN response needs to be updated, as well.
  1. From the Analysis page, click the optimization panel.
  2. Click the responses panel.
  3. Click response= and select cstrain.
  4. Using the plies selector, select all of the plies.
  5. Click update.
  6. Click return.

Create the Manufacturing Constraints for Shuffling

A DSHUFFLE card was created automatically during the sizing phase. Two manufacturing constraints will be added for the shuffling optimization.
  1. From the Optimization panel, click the composite shuffle panel.
  2. Select the create subpanel.
  3. Click dshuffle= and select DSHUFFLE1.
    Review the type and stack ID.
  4. Select the parameters subpanel.
  5. Click dshuffle = and select DSHUFFLE1.
  6. Select pairing constraint.
  7. Set the pair type to reverse.
  8. In the ply angles1 field, enter 45.0.
  9. In the ply angles2 field, enter -45.0.
  10. Click update.
  11. Click edit.
  12. Define the MAXSUCC constraint, as shown in Figure 1.

    Figure 1.
  13. Click return.
  14. Click update.
  15. Click return twice to go back to the Analysis page.

Run the Optimization

  1. From the Analysis page, click OptiStruct.
  2. Click save as.
  3. In the Save As dialog, specify location to write the OptiStruct model file and enter fairing_shuffling for filename.
    For OptiStruct input decks, .fem is the recommended extension.
  4. Click Save.
    The input file field displays the filename and location specified in the Save As dialog.
  5. Set the export options toggle to all.
  6. Set the run options toggle to optimization.
  7. Set the memory options toggle to memory default.
  8. Click OptiStruct to run the optimization.
    The following message appears in the window at the completion of the job:
    OptiStruct also reports error messages if any exist. The file fairing_shuffling.out can be opened in a text editor to find details regarding any errors. This file is written to the same directory as the .fem file.
  9. Click Close.

View the Results

Open the fairing_shuffling.shuf.html file in your Internet browser.
The history of the shuffling optimization displays. The columns represent the global trend of the ply stacking sequence at a particular iteration, with the last column being the final solution. The plies are color coded based on their fiber orientations. The weight of the fairing has not been changed during the shuffling design phase.

Figure 2. Shuffling Optimization History
Reviewing the results from this process:
  • Lowest natural frequency = 0.02 KHz (>0.02 KHz)
  • Maximum strain = 9.947e-4 (<1.e-3)
This light weight design therefore meets all of the performance requirements, is feasible and manufacturable.