Frequency Domain System Identification of a Robinson R44 in Hover

Abstract

The civil light helicopter domain has not fully benefited yet from the advantages system identification methods can offer. The aim of this paper is to show that system identification methods are mature enough to be successfully implemented in the civil helicopter domain. To achieve this goal, a Robinson R44 Raven II is identified in this work. The identification focuses on the hover trim condition. A lean frequency domain identification method is adopted. Furthermore, a new procedure is proposed to limit the sensitivity of the state-space minimization algorithm to initial parametric values and bounds. The resulting state-space model presents good predictive capabilities and is able to capture high frequency rotor-body dynamics. The model is also validated with the help of a helicopter pilot by performing closed-loop control task maneuvers in the MPI CyberMotion Simulator. The g gravitational acceleration, [m/s 2 ] s Laplace transform variable φ, θ, ψ fuselage angular attitude (roll, pitch, yaw) earth-fixed coordinates, [rad] β 0 , β 1c , β 1s rotor coning, longitudinal and lateral flapping angles, [rad] τ f rotor flap time constant, [s] ν rotor inflow velocity, [m/s] ν0 trim inflow ratio γ Lock number ξ, ω damping and natural frequency of a second order system η Ct integrated perturbation thrust coefficient σ rotor solidity ρ atmospheric density, [Kg/m 3 ] δ lat , δ lon , δ ped , δ col helicopter control inputs (lateral cyclic, longitudinal cyclic, pedals rudder, collective lever), [deg] Ω rotor rotation speed, [rad/s] ω frequency, [rad/s]

A Cybernetic Analysis of Maximum Unnoticeable Added Dynamics for Different Baseline Controlled Systems

“…Recently, it has been stressed by Mitchell et al (2006) that such envelopes depend on the dynamic characteristics -i.e., bandwidth -of the baseline (reference) vehicle dynamics, a dependency that would greatly complicate the task of defining universal envelopes for allowable levels of mismatch. Still, given their straightforward applicability and their intuitive soundness, the MUAD envelopes have become a de facto standard to assess the adequacy of LOES (Field et al, 2003;Bosworth and Williams-Hayes, 2007;Geluardi et al, 2014).…”

Transforming Civil Helicopters into Personal Aerial Vehicles: Modeling, Control, and Validation

“…The flight tests focused on the hover trim condition. The identification procedure that was used to obtain the final model can be found in [8]. The final model is a fully coupled 12 Degrees Of Freedom (DOF) state-space model valid within the frequency range of 0.1-16 rad/sec.…”

“…In a previous study [6], these control strategies were designed and applied to an identified model of a Robinson R44 Raven II in hover. The two control strategies are designed and applied to an identified model of a Robinson R44 Raven II in hover [7].…”

“…In a previous study [6], the two control strategies had been applied on an initial identified model of the same helicopter. Since the previous study [6], an improved model of the Robinson R44 has been identified through an iterative validation process, performed with the help of a helicopter pilot and the use of the Max Planck Institute (MPI) CyberMotion Simulator [7].…”

Manual Control Adaptation to Variations in Short-Period Natural Frequency and Damping

“…The concepts of "maximum unnoticed added dynamics" (MUAD) [1][2][3] or "maximum allowable errors" (AE) 4 are traditionally used in studies of aircraft handling qualities, [5][6][7][8][9][10][11][12][13] to formalize the human pilot's sensitivity to the specific dynamics of his controlled aircraft. This sensitivity is generally expressed as a frequency-domain (MUAD or AE) envelope within which any added dynamics (e.g., augmented systems, system failures) to the original aircraft dynamics will remain unnoticed by the pilot.…”