Discussed an approach to Multidisciplinary Design Analysis and Optimization (MDAO) research; a deliverable for Year 1 (to incorporate flight controls into the MDAO process) and a necessary step for the mAEWing2 design process, slides are attached to this page. The approach as presented consists of:

  • Use an inviscid CFD approach to explore a range of bending and twist deformations to the mAEWing2 planform. This study would be used to find lower drag shapes, similar to ​Nhan Nguyen’s “Elastically Shaped Future Air Vehicle Concept” paper. Results from this study could be a set of solutions or a parameterized model.
  • Run MDAO using this set of potential shapes:
    • Determine the amount of flexibility necessary to achieve the deformed shape.
    • Determine how this flexibility affects the aircraft flutter modes and frequencies.
    • Optimize the control surface and sensor layout, likely by placing surfaces away from node lines.
    • Generate reduced order models (ROMs) and perform some control analysis to assess the observability and controllability of the aircraft design.
    • Run the MDAO to maximize some performance criteria based on:
      • Reaching shapes that reduce drag of the vehicle.
      • Observability and controllability metrics.
  • Following preliminary design from MDAO:
    • Validate the aerodynamics from the determined shape with a viscous approach.
    • Optimize the structural design for the desired flexibility, primarily through the dimensions and placement of the spar.
    • Conduct a detailed control law design.
    • Build, validate, and flight test mAEWing2.

Our discussion led us to the following:

  • Likely, our study will find that twist, rather than bending, leads to a drag reduction
    • This finding will probably lead us to use leading-edge and trailing-edge control surface pairs to induce twist.
    • We’ll have to induce twist within the trim nullspace of the aircraft (i.e. add twist without adding an overall pitching moment).
    • We’ll want to make sure that the drag due to surface deflections isn’t so great that we counteract the lift benefits from deforming the wing - this means we’ll probably want to optimize control surface layouts to minimize necessary control deflections.
    • We should consider structural designs that would allow the aircraft to be stiff in bending and flexible in torsion; although, this may be limited since they will be somewhat coupled due to the wing sweep.
    • In addition to photogrammetry, we may consider using IMUs or root strain gages to validate that we modified the aircraft shape in flight.
  • Likely will be able to use VLM or DLM aerodynamic models for study of lower drag shapes rather than inviscid CFD, but would still use inviscid CFD to validate that VLM/DLM is giving the correct trends.
  • Rather than tune gains or optimize a control law in MDAO, probably just look at observability and controllability to make sure that the vehicle is controllable and to make the controls problem easier to solve (a metric for “easiness” requires some more thought, probably try to keep unstable structural mode frequencies lower and reduce the number of modes that go unstable).
  • Likely just include the effectors in the MDAO process and take care of sensor placement after the MDAO is complete.

Open Items and Next Steps

  • We’ll meet again to discuss specific tasks and breakdown work.
  • Need to see whether to use inviscid CFD or VLM / DLM for aircraft shape study to minimize drag. UMN is investigating some modifications to their VLM/DLM code to enable this study. Would still like to use CFD to validate VLM/DLM identified drag trends.
  • Need to determine metrics for observability / controllability performance
  • Will need to gather information on flight conditions and bounds for the optimization (should be relatively easy).

Meeting Presentations on: GitHub