Cross-basin aerodynamic parameter simulation method based on convolutional neural network

Abstract

The method comprises the following steps: step 1, obtaining a partial representative aerodynamic data result based on a gas kinetic theory unified algorithm of a Boltzmann model equation; 2, based on a partial representative aerodynamic data result, by means of a convolutional neural network algorithm, obtaining a shape convolutional neural network fitting model; and 3, according to a convolutional neural network fitting model, obtaining aerodynamic state result parameters of a required state and a flow field information cloud picture. According to the trans-basin aerodynamic parameter simulation method based on the convolutional neural network provided by the invention, compared with different calculation methods used in different basins in the past, the calculation result of the typical flight height streaming state of a unified algorithm is used as data support, and the full-basin aerodynamic characteristics can be quickly calculated by using one method; and the situation that results at junctions of drainage basin methods, especially thin transition flow areas, are not uniform is avoided, and the result reliability is high.

Description

technical field [0001] The invention relates to the technical field of aircraft aerodynamics, in particular to a method for simulating cross-watershed aerodynamic parameters based on a convolutional neural network. Background technique [0002] In the past, when the computer conditions were not advanced enough and the fluid dynamics calculation technology was not fully developed, aerodynamic researchers could only calculate independently according to the characteristics of typical flow states in different watersheds. region, slip flow region, thin transitional flow region, and free molecular flow region, and then use Euler and N-S equations to solve the continuous flow region problem, and use the Slip-N-S (N-S with slip boundary condition) equation algorithm to solve For the flow problem in the slip flow region, the bridging method and coupling method are used to solve the aerodynamic calculation problem in the thin transition flow region, and the DSMC method is used to solv...

AI Technical Summary

Problems solved by technology

[0005] However, due to the large amount of calculation and resource consumption of the unified algorithm, it is impossible to obtain the aerodynamic data results of the whole watershed with a small amount of computing resources in a short period of time. A bridge formula method based on the theoretical correlation between free molecular flow and continuous flow local bridge function. This method determines the correlation parameters according to the shape of the aircraft and the characteristics of the flow around it. Taking some representative aerodynamic data results of the unified algorithm as a reference, the free molecular flow and continuous flow are obtained. Continuous flow whole domain aerodynamic data results

[0006] Although the above method achieves the purpose of obtaining the aerodynamic data results of the whole watershed with relatively small resources in a short period of time, the associated parameters of the bridge formula method need to be manually adjusted, and the best fitting results may not be obtained by visual calibration, and the operation High level of personnel experience required

Images