Simulation and prediction of main rotor, tail rotor, and engine parameters from flight tests

Abstract

In the framework of this research project, the main rotor torque, tail rotor torque, engine torque, and main rotor speed of a helicopter in forward flight are estimated by using a state-space model from flight tests data. The state-space model inputs are non-linear terms made of combinations of pilot controls and helicopter states. The model simulates the helicopter outputs while knowing the states and controls at all times. It was also implemented as a prediction tool, for possible use in an envelope protection flight control system in which the states, controls, and outputs are known at the present time, and predict the future helicopter states and controls following to pilot controls time history. The state-space model parameters are identified by using the subspace identification method, a relatively recent non-iterative algorithm, which constructs an observability matrix from input and output data and uses this matrix to obtain the state-space matrices. The obtained parameters are then optimized with the Levenberg—Marquardt output-error method. A comparison of the results with and without optimization is also conducted. The results show that the subspace method provides a good estimate of the outputs within the FAA tolerance bands and that these results can further be improved by use of the minimization algorithm. The generated model using the subspace method is found to be very good for prediction applications, which makes it a promising model for flight control simulator applications.