|The University of Adelaide||Home | Departments | Search|
(Commenced: 01-Jan-2005,Concluded: 30-Nov-2005)
This honours project is a continuation of the 2004 University of Adelaide Mechanical Engineering final year VTOL project, which aimed to develop a controlled and stable VTOL model aircraft based on the F-35 Joint Strike Fighter. Last years project group was unable to complete the project for several reasons, primarily because of the poor performance of the internal combustion engines. Due to advances in electric motor and battery technology, the use of electric motors in now a feasible option and has been chosen for this years project. An investigation was performed to re-evaluate the most suitable aircraft upon which to base this years design. This research identified that the V-22 Osprey was the most suitable platform. The primary advantage in using the V-22 Osprey as opposed to the F-35 was the higher thrust to weight ratio achievable using propellers.
Several designs were considered during the concept evolution, with each concept compared with the design requirements. The final concept which met all requirements was then modelled in SolidEdge where further details were considered. The design consists of an Aluminium chassis with rotating wing arms controlled by servo motors allowing for tiltrotor operation. Control simplifications use this wing arm rotation along with a rear tail-fan to directly control all three rotational degrees of freedom.
Thrust providing components were selected to best satisfy the constraints of cost, weight and power. Ultimately two-blade fixed-pitch propellers were chosen to be mounted radially to brushless motors via a planetary gearbox. These motors are controlled using an electronic speed controller and powered by on-board Lithium Polymer batteries. Using software based propeller theory these components were shown to produce adequate thrust.
To facilitate the creation of an appropriate control system using the physical characteristics of the model aircraft, a mathematical representation of the system was obtained by following several closely related examples. Using this mathematical representation a state space controller was developed using a Reduced-Order Observer and tuned using LQR Optimal Control. Another control technique, Proportional Integral Derivative (PID) control, was also used as a controller for the aircraft. While the PID controller was less capable than the state space controller, it was much easier to implement and tune.
The controllers were initially built in Matlab Simulink, which creates C code that is cross-compiled and downloaded to a dSPACE DS-1104 rapid prototyping control platform. This allowed for the easy implementation and tuning of these control systems. However this control platform is an undesirable final solution to control the aircraft due to its size and cost, so the control system was migrated to the on-board MiniDRAGON+ microcontroller. This microcontroller was specifically programmed to interface with all of the control peripherals, including a standard remote control and receiver which was used to control the model.
The model aircraft was attached to a gimble to allow the tuning of the control systems while in a tethered and safe configuration. However due to problems associated with this gimble the tuning was not successfully completed. The model was then moved to a semi-tethered configuration, where it was removed from the gimble but still tethered via the wiring loom which allowed relatively free movement within a fixed range. During this semi-tethered stage of controller tuning a gearbox shaft failed which prevented progress towards the project goals of controlled and stable hover.
The project was set back due to the late procurement of vital components and several mechanical failures. While this contributed to the project goals of controlled and stable hover being incomplete, the design was shown to be able to provide enough thrust to achieve hover and sufficiently controllable to achieve these goals. Furthermore due to the significant work undertaken in integration and control embedding the model is a solid control platform for future work.