Research On Modeling Method Of Large Supersonic Wind Tunnel

Large supersonic wind tunnel is an important experimental platform for high-speed aerodynamic theory.In order to achieve higher control accuracy,an accurate mathematical model of the supersonic wind tunnel must be established.In this paper,the supersonic wind tunnel is taken as the research object,and the supersonic wind tunnel system model is established for the purpose of serving the design of high-precision controller.In this paper,the key parameters of wind tunnel modeling are obtained by analyzing the structure,operation mode,blow test process and control system of supersonic wind tunnel system.On this basis,the dynamic,strong nonlinear and internal of the supersonic wind tunnel system are obtained.Based on the characteristics of variable measurability,the research strategy of modeling the supersonic wind tunnel sub-module is proposed.The whole supersonic wind tunnel model is divided into two sub-modules:main pressure regulating valve system and air flow field.Different identification strategies are used to model respectively.The main pressure regulating valve is represented by a linear dynamic model.The order of the model is determined by the determinant method,and the parameter is identified by the batch least squares method.The linear model of the main pressure regulating valve is obtained.The reason why the accuracy of the linear model can not meet the requirements is analyzed.Combined with the experiment and analysis,the conclusion that the main pressure regulating valve system exhibits discontinuous nonlinear dynamic characteristics in a certain time is obtained.Therefore,in order to improve the accuracy of the model,the main pressure regulating valve system is represented as a block structure model with a nonlinear dynamic input part and a linear dynamic output part.Aiming at the special model structure of the main pressure regulating valve,this paper proposes a recursive identification algorithm based on Kalman filtering algorithm to estimate the unknown parameters of the model.The online identification results of the main regulator valve system verify the validity of the selected model structure and algorithm.Based on the analysis of air flow field characteristics,the air flow field is approximated by a dynamic and nonlinear characteristic of the air flow field,which is a Hammerstein block structure model composed of a nonlinear static input part and a linear dynamic output part.Aiming at this model structure,this paper proposes a method to estimate the order of Hammerstein model based on the determinant method.In order to realize the real-time control of the air flow field,the parameters need to be estimated in real time,so the parameters of the air flow field model are identified online by the recursive least squares method.Finally,the effectiveness of the model structure and algorithm for air flow location selection is verified by experiments.The modeling strategy of dividing sub-modules and the nonlinear system based on block structure are used to establish the supersonic wind tunnel model,which can make the model have higher prediction accuracy and is beneficial to the design of high-precision controllers.