Optimal Input Design Method And Aircraft Parameter Identification

As the core and basic technology in the field of aircraft test and design application,the improvement of the accuracy of aerodynamic parameter identification has extremely important research significance.Based on the 6-DOF modeling and nonlinear aerodynamic modeling of aircraft,the key technologies such as parameter identification,test data preprocessing,identification algorithm improvement and optimal input design for aerodynamic parameters identification were studied.The following results were obtained:1.The mathematical model of the aircraft was Established and analyzed.(1)A 6-degree-of-freedom mathematical model and a non-linear aerodynamic model were established.(2)The state equations and the observational equations in the aerodynamic parameter identification system of the aircraft were defined,and the identifiability of the aerodynamic parameters was analyzed.(3)An in-depth study of the maximum likelihood estimation method was carried out to identify the aerodynamic parameters of the aircraft using the Newton-Raphson iterative algorithm as an identification algorithm.It was pointed out that because of the sensitivity to the initial value of the iteration and the difficulty of solving the Jacobian matrix,the identification results obtained by using Newton-Raphson iterative algorithm were not satisfactory or even convergent.2.The application of traditional optimization algorithm in aerodynamic parameter identification of aircraft was studied.(1)Converting the aerodynamic parameter identification problem into a least-square nonlinear least-squares optimization problem based on the least-squares criterion,and summing up it into an unconstrained optimization problem.(2)On the basis of unconstrained optimization problem,the Newton method and the basic Gauss-Newton method were introduced,the limitation in solving aerodynamic parameter identification of aircraft was pointed out.(3)In view of the problem that the iterative direction of basic Gauss-Newton method may not decrease,improved algorithms were introduced: variable-step Gauss-Newton method and Levenberg-Marqurdt method.(4)According to the characteristics of a single algorithm,a hybrid iterative algorithm combining GaussNewton method and Levenberg-Marqurdt method is proposed.A simplified updating algorithm of Jacobian matrix was proposed,which alleviates the difficulty of solving Jacobian matrix in iterative process.3.The particle swarm optimization algorithm was studied and improved.(1)The PSO algorithm was studied,and it was pointed out that the standard PSO algorithm can easily fall into a local optimum when solving high-dimensional nonlinear problems.And the inertia weights and learning factors are analyzed,and corresponding improvements were introduced to provide a comparison for subsequent improved algorithms.(2)The social emotion model was studied and the intelligent behaviors were analyzed.Then an improved particle swarm optimization algorithm: SEPSO algorithm was proposed,which improved the population diversity of the particle swarm optimization algorithm.Therefore,it was possible to apply it in the identification of aircraft aerodynamic parameters.4.The optimal input design method was studied.(1)The optimal design criterion was studied.Since the condition number is a very important criterion in matrix operations,a new optimal criterion based on D-optimal criterion and matrix condition number: Dc-optimal criteria was proposed.(2)The multiple sine orthogonal input method was studied,which provided a method for the optimal input design.(3)The optimal input excitation based on Dc-optimal criterion and multiple sine orthogonal input method was designed for the single motion direction of the aircraft linear model,and the accuracy of identification was improved.5.The aerodynamic parameter identification of aircraft based on optimal input design and particle swarm optimization was studied.(1)For the 6-degree-of-freedom nonlinear model and nonlinear aerodynamic model of the aircraft,the optimal input is designed.(2)The measurement data preprocessing technique was studied and an adaptive robust Kalman method was proposed.The results showed that the adaptive robust Kalman method can deal with noise and deviation more effectively and had better robustness.(3)For the nonlinear aircraft aerodynamic model,the designed optimal input excitation was used as the system input,and the SEPSO algorithm was used as the parameter identification method.The results of the identification were compared with the results of the hybrid iterative algorithm and the standard PSO algorithm,which showed that the proposed method can effectively improve the identification accuracy.The parameter identification method based on optimal input design proposed in this paper can effectively improve the accuracy of aerodynamic parameter identification,and has positive reference value for the development of related fields.It can also provide reference for parameter identification of other complex nonlinear systems.