Improved Rotor Track And Balance Algorithm And System Identification Of Rotorcraft Using Flight Data

Abstract

Rotor track and balance is an essential part of maintenance for the rotorcraft industry. Methods for solving rotor track and balance problems to minimize undesirable vibrations in the cockpit and cabin of rotorcraft are analyzed. Additionally, different approaches to system identification for airframe response are discussed with emphasis on real-life practical applications. Rotor track and balance vibration optimization methods using least squares, weighted least squares, linear quadratic regulator, two-sensors only in the fundamental frequency, all sensors in the fundamental frequency, and all sensors in all frequencies up to N-1 of the fundamental frequency are formulated within this study and used to calculate control adjustments quickly. The resulting vibration reduction is simulated by approximating the effects of rotor adjustments utilizing a sensitivity matrix that maps the rotor track and balance controls to the response of the rotor. Finally, experimental data involving the use of the proposed method on actual experimental flights are shown. System identification of a helicopter airframe is conducted by capturing flight data from sensors throughout the cockpit and cabin as a result of a rotor track and balance control changes. A conventional approach, which plots data acquired from single control type adjustments on polar charts to aid in the proper identification of sensitivity coefficients, is discussed. The proposed recursive least squares algorithm is developed and implemented against a database containing hundreds of flights and has flexibility in accepting input data. Results show that modeling airframe responses from sensitivity coefficients using large data sets compared to the few gathered during flight test on a test aircraft are more accurate in terms of predicting the vibration level after control adjustment changes. This proposed methodology allows sensitivity coefficient development to continue as the aircraft is being produced. The method converges quickly and produces accurate results while successive measurements fine tune the sensitivity coefficients. This leads to the ability to better model airframe response and ultimately translates to lower flight counts. As a result, time and money can be saved during business operations.