Comparison Of Frequency And Time Domain Identification Of A Helicopter In Hover

Abstract

This paper focuses on frequency domain system identification of a full-scale helicopter under both hover and forward flight conditions and comparison of the results with time domain identification. A high fidelity nonlinear simulation model resembling a 3T class helicopter is employed. Frequency-sweep input signals are applied to generate the required input-output datasets. The time domain signals are transformed into the frequency domain using the Discrete Fourier Transform, along with conditioning and composite windowing techniques to ensure high accuracy and reliability. A state space system based on a six degree of freedom quasi steady linear model is identified using the output error method combined with the Levenberg-Marquardt algorithm. To account for higher-order dynamics, time delays are introduced between inputs and outputs. In addition, the frequency domain based identified system is compared against an independently time domain identified model using data from the same truth model. The results demonstrate high identification accuracy, with the key aerodynamic derivatives estimated within a consistent and comparable range.