Study On A Method For Helicopter Flight Dynamics Model Identification

Helicopter flight model established by identification method is more credible compared with mechanism modeling method. It can be used not only to evaluate and design the flying quality, but also simulate and verify the flight control laws. It is of important meaning to aircraft design and flight control system. In this paper, helicopter identification model is established by flight test data preprocessing, non parameter model identification and parametric model identification technology.Based on these methods and technology, the identification software for helicopter flight model is developed.Firstly, the flight test data preprocessing methods such as implementation method, restructuring and filtering technology are studied against low signal to noise ratio and strong inter axis coupling which are inherent characteristics of helicopter flight test data. The error of test data is effectively eliminated. The quality of flight test data is improved.Secondly, the cross axis decoupling technique is realized by using linear frequency modulation conversion technology and improved spectral estimation. The composite window technique is used to obtain good frequency response. Then non parametric system identification model is developed.Thirdly, the transfer function model is developed with least square method and prediction error method. Then, the state space model with clearly physical meaning is developed by using the subspace method, prediction error method and the prior knowledge of the helicopter flight dynamics model, and the simulation data of UH60 is used to verify the identification method.Finally, the system identification software is developed based on MATLAB. The flight test data of XV15 is used to test the software. The identification results are compared with that of commercial software CIFER. The results show that the model has good precision. The methods and technology are verified to be feasible. The software is testified to be practical.