Surrogate Model Of High Temperature Plume Radiation And Identification Of Engine Parameters

With the rapid development of hypersonic and stealth technology,the technology of detecting and identifying targets based on the flame radiation of rocket engine of high speed aircraft is paid more and more attention.However,the generation mechanism of rocket engine plume radiation signal is very complex,and there are many signal transmission links in the detection process.At the same time,for non cooperative targets,the engine prototype parameters and flight parameters are not complete.As a result,the plume radiation signal obtained by space-based infrared detection system is highly degraded,and it is difficult to realize the classification and recognition of flight targets through the detection information.Therefore,through the whole chain simulation of rocket engine from prototype parameters to plume radiation signals,it is very important to study the identification of target prototype parameters from degraded detection signals.This research can provide theoretical and practical basis for the identification of rocket engine type spectrum parameters under the condition of incomplete parameters,and can also provide theoretical and practical basis for target classification based on optical radiation signals It provides new ideas for identification and diagnosis.In this paper,the typical formula spectrum of rocket engine exhaust plume is taken as the research object.The purpose of this paper is to construct the characteristic characterization system of exhaust plume infrared radiation signal from the phenomenological point of view,establish the multi factor correlation model of exhaust plume radiation signal transmission,and on this basis,carry out the uncertainty analysis and parameter sensitivity analysis of exhaust plume radiation signal,The influence mechanism and law of the engine type spectrum parameters on the exhaust plume radiation signal are clarified,and then the identification method of the engine type spectrum parameters is studied by extracting the typical exhaust plume radiation characteristics.The main research work of this paper is as follows:Based on the rocket engine plume radiation signal analysis tool developed by the research group,a multi factor correlation plume radiation signal transmission model under the typical formula pedigree is established by using Kriging model,support vector machine,polynomial chaotic expansion and other surrogate model methods.The results show that polynomial chaotic expansion method is more suitable for plume radiation calculation,This method can save a lot of calculation cost.At the same time,combined with the atmospheric radiative transfer software,a signal transfer model of plume radiation after atmospheric attenuation is constructed,which has high prediction accuracy.Based on polynomial chaos expansion(PCE),a surrogate model of typical height plume radiation signal of multi type rocket engine is established.The uncertainty factors in the process of plume radiation signal generation / transmission are analyzed quantitatively by using Monte Carlo method and sensitivity analysis method combined with Sobol index.The influence mechanism and law of rocket engine type spectrum parameters on plume radiation signal are clarified.By extracting the typical characteristics of the rocket engine plume radiation signal,a parameter identification method based on the typical characteristics of the rocket engine plume radiation signal is established by using the surrogate model inversion learning method and the improved support vector machine;It is found that the noise has a great influence on the parameter identification accuracy of rocket engine.