Fault tolerant control of multirotor vehicles
Loading...
Authors
McKay, Michael Evin
Issue Date
2021-12
Type
Electronic thesis
Thesis
Thesis
Language
en_US
Keywords
Aeronautical engineering
Alternative Title
Abstract
The future of vertical takeoff and landing (VTOL) vehicles involves many exciting and novel configurations, among which are electric multirotor vehicles and high-speed coaxial helicopters. These concepts move away from the conventional ``pod and boom" single main rotor helicopters, instead utilizing multiple rotors to generate the forces and moments required to control the vehicle. In the case of electric multirotor vehicles, the anticipated use cases involve human transport (``air taxis"), package delivery, and surveillance (among others), while the high speed coaxial helicopter is being developed to serve military missions, including aerial scout missions (assault and reconnaissance) and troop transport. The missions are markedly different, but both require safe operation, even when faults occur, to be deployed in the future of VTOL aircraft.
In the electric multirotor field, aircraft range in size from small package delivery drones to large manned air taxis. Research on these platforms consider the entire range of scale, considering things like control design using simplified physics models that ignore much of the physics of the problem. Additional work applies high-fidelity computational fluid dynamics analyses that consider the entire complexity of the physics, but at the cost of high computational time. The present work applies a developed medium-fidelity analysis tool to analyze a small (2 kg) hexacopter trim and performance with single rotor failure. The application of medium-fidelity tools allows for more confidence in the predicted results relative to low-fidelity models, at a significantly lower cost than high-fidelity analyses.
For coaxial helicopters, limited publication is available in the open literature due to ongoing development of the aircraft at present. Established comprehensive codes have been utilized to conduct studies on these platforms, but limited flight test data is available to validate predictions against. For analysis in the present body of work, the RPI Coaxial Helicopter Analysis and Dynamics (CHAD) code is developed and validated against available rotor test stand measurement, aircraft steady trim performance and control setting data, and identified flight test dynamic characteristics. This code is developed using a coupled finite state dynamic wake model to predict aerodynamic interference between rotors, as well as a coupled fuselage and elastic flap-lag blade dynamic model to account for coupled blade and rigid body motion as the vehicle operates.
Safety in the operation of coaxial helicopters, as in any air vehicle, is imperative to the successful deployment of the platform. One of the interesting features of this configurations relative to conventional single main rotor aircraft is control redundancy, which (like the aforementioned electric multirotor vehicles) can allow for control reconfiguration and tolerance to control failure during operation. CHAD is utilized to analyze coaxial helicopter fault tolerance in two ways: steady trimmed flight and dynamic simulation.
The former approach considers the allowable variation in control settings in low, moderate, and high speed flight, identifying potential ranges of flight control settings where the aircraft could retain trim even when a locked-in-place fault has occurred, considering assumed geometric limits on controls and observed tip clearance limits between the coaxial counter-rotating rotors. The latter examination explores control reallocation for a locked-in-place flight control fault during dynamic simulation. To this end, an explicit model following closed loop flight control system is designed, and a pseudoinverse control allocation is implemented to distribute control effort among the available effectors. Various control faults are considered in low and high speed flight, the pseudoinverse allocation is recalculated based on the available effectors, and the vehicle behavior post-failure is examined and compared across different fault cases.
Description
December 2021
School of Engineering
School of Engineering
Full Citation
Publisher
Rensselaer Polytechnic Institute, Troy, NY