On-blade smart structure actuators are capable of actively altering the aerodynamic loads on rotor blades. With a suitable feedback control law, such actuators could potentially be used to counter the vibrations induced by periodic aerodynamic loading on the blades with lower weight penalties than the previous actuation methods and without the bandwidth constraints. This paper covers the development and testing of a new, robust individual blade control methodology for rotor vibration suppression using piezo-actuated trailing edge flaps and active twist tip rotors. The controller uses a neural network to learn to actuate the trailing edge flap thus adaptively suppressing the blade or hub vibrations. In this application, no off-line training is performed. Instead, a neural network is used in real time to adaptively command the actuator deflections thus reducing vibrations. Closed loop experimental tests with piezo-actuated scale rotor systems were conducted on the University of Maryland hover test stand. The results include two different Mach scale smart rotor systems (trailing edge flaps and active tip twist) that were controlled by the same adaptive neurocontrol algorithm. These tests demonstrate the controller's robust ability to successfully learn to control the rotor vibrations with no a priori information about the blade/actuator structure or the aerodynamic loading.
Closed-loop hover test results with a neurocontroller on piezo-actuated smart rotor blades
2000
12 Seiten, 30 Quellen
Aufsatz (Konferenz)
Englisch
Closed-loop neurocontroller tests on piezoactuated smart rotor blades in hover
Tema Archiv | 2002
|Open-Loop Hover Testing of a Smart Rotor Model
AIAA | 2002
|