Three flutter-suppression designs for a flexible flying-wing research drone are discussed, along with the modeling and flight-test results. The drone, with wing span of 10 feet, aspect ratio of almost 9, and control surfaces along the entire wing trailing edges, was developed for flight research into the modeling and control of highly elastic aircraft. The lowspeed vehicle was designed to exhibit body-freedom flutter in its flight envelope. The all-important dynamic modeling of the vehicle, used for analysis and control design, is based on a mean-axis formulation and quasi-steady aerodynamics, with the nondimensional aerodynamic/aeroelastic coefficients updated from flight tests. Flutter-modeling results were found to agree with those from NASTRAN and flight tests. The three different flutter-suppression approaches include both fairly classical and multivariable methods, all fixed gains but with different architectures. The primary controldesign objective was to augment the damping of the eventual flutter mode at a design condition below the open-loop flutter speed, for safety of flight, and to achieve a closed-loop flutter speed at least as high as the open-loop flutter speed. Analysis and flight tests revealed that all three designs achieved these design goals. In addition, two of these designs actually expanded the flutter boundary. The theoretical stability robustness of all three controllers at the design flight condition is quite good, but there are differences in controller complexity. Research is currently underway to further expand the flutter boundary using all three design approaches.
Bibliographical noteFunding Information:
This work was conducted as part of a multiyear NASA Research Announcement (NRA) program (Cooperative Agreement No. NNX14AL63A) led by the University of Minnesota, with Systems Technology, Inc., Virginia Polytechnic Institute and State University, D. K. Schmidt and Associates, CMSoft, Inc., and Aurora Flight Sciences as partners. The authors would like to acknowledge the Lockheed Martin Skunkworks for their generous donation of their body-freedom flutter (BFF) vehicle to the University of Minnesota after conclusion of their flight testing. That BFF vehicle was the starting point for the test-bed vehicle used in this research.. The authors would also like to acknowledge all partners as well as NASA for valuable technical support and resources. John Bosworth and Jeff Ouellette of the NASA Armstrong Flight Research Center have served as Technical Monitors. The authors would also like to recognize the reviewers who contributed significantly to this paper. Finally, the authors would like to remember and acknowledge our late colleague Gary Balas, formerly of the University of Minnesota, who founded the unmanned aerial vehicles laboratory at the university, and was the original principal investigator on this project.