Research On Fuzzy Logic Algorithm For Flush Air Data Sensing System

This thesis studied the Fuzzy Logic Modeling Algorithm for the Flush Air Data Sensing System.Based on the pressure data on the nose of an aircraft model and the leading edge of the wing from the wind tunnel experiments,firstly the surface pressure of the model was calculated through my program using computational fluid dynamics(CFD)methods in this thesis,and good agreement with the experimental data and Fluent result could be noticed,validating the rationality of the mesh grid,the calculating methods and parameters of the numerical simulation.On this basis,the surface pressure data of the model at several altitudes and different flight states was further calculated and restored.Secondly,based on the wind tunnel experiments and CFD simulations,the structure of the fuzzy logic model,inner functions,membership functions,and the process to identify the structure of the fuzzy model was presented in thesis.Then a nonlinear mapping model for the relationship between the surface pressure and angle of attack,angle of sideslip,inlet velocity,and altitude was built.In contrast to most former research in which the method was often used in the modeling of the aerodynamics,this thesis introduced this method into the modeling of FADS system,which was essentially taking advantage of its characteristics that are applicable for nonlinear modeling.Moreover,comparing to the former adopted fuzzy logic method that only used the gradient descent method to identify the parameters of the inner functions in the modeling process,this thesis innovatively employed changeable membership functions that could change their shapes according to the change of inner functions' parameters simultaneously,forming a hybrid learning procedure which used both the gradient descent and the least squares estimate method to identify parameters of the system.The modeling accuracy and speed were greatly increased owing to the method we adopted and the final model obtained could better describe the nonlinear relationship within the original data.Finally,since dozens of pressure taps was placed on the experimental model surface,various schemes for the selection of pressure taps with different locations and number was studied.Considering the modeling accuracy,speed and equipment of pressure measurements,the optimized scheme for the selection of pressure taps was given at last,which could be a guide for the engineering applications of the FADS system.