Aerodynamic Parameter Identification For Aircraft

With the development of aircraft technology, to obtain accurate and reliableparameters of vehicle model has an important theoretical and practical significancefor the aircraft design. Aerodynamic parameter identification is a reliable andeffective means, which has become indispensable to modeling, simulation,controller design of vehic le. In this paper, the modeling of aircraft, identificationmethod and a novel approach in closed-loop identification are studied, specificallyin following aspects.Firstly, considering the flight characteristics, dynamic model of aircraft withsix degree of freedom (6-DOF) is built based on literature. Meanwhile, aerodynamicmodel of polynomial form is set up in preparation for the identification ofaerodynamic parameters.Secondly, maximum likelihood criterion is introduced, which is most widelyused in aerodynamic parameter identification. Newton-Raphson method and geneticalgorithm are based on this criterion and commonly used in parameter identification.Considering the advantages and disadvantages of the two methods, a hybrid geneticalgorithm based on local optimization operator in gradient method is proposed,combining the advantage of global optimization of genetic algorithm with that ofhigh precision and speed of Newton-Raphson gradient method. Simulation examplesverify the effectiveness of the hybrid genetic algorithm.Finally, two problems of identification emerges in the closed-loop flight test:one is low precision in the open-loop identification algorithm and the other isinformation loss due to the correlation of the input and output signal in the controlsystem. To solve the problems above, a novel closed-loop two-step identificationmethod based on hybrid genetic algorithm is proposed and the optimal input signaldesign is applied respectively, and numerical example verifies the effectiveness ofthe new approach.In conclus ion, on the basis of dynamic model of aircraft with six degree offreedom, a hybrid genetic algorithm based on local optimization operator in gradientmethod is proposed. Moreover, closed-loop two-step method the optimal inputsignal design solves the problems in the closed-loop identification. Finally, thesimulation example is provided to illustrate the effectiveness of the method.