Research On Laser-Induced Plasma Ignition And Flame Stability Of A Rocket Engine

Reliable re-ignition and combustion status monitoring of a new type of rocket engine using non-toxic,non-self-igniting clean propellant is an urgent problem in the aerospace field.Laser-Induced Plasma Ignition(LIPI)is a new type of ignition technology,which has attracted much attention because of its ignition energy and time precision adjustable,flexible ignition position,and the ability to achieve multi-point ignition.The development of artificial intelligence makes it possible to monitor flame stability based on flame images and spectral information.This article first introduces the significance of laser-induced plasma ignition and flame stability research,investigates the research status at home and abroad,and proposes the research content of this article.Secondly,the feature parameters of commonly used flame images are introduced,and the theoretical basis of Hu moments,Zernike moments,and Krawtchouk moments used in feature extraction are briefly explained.The classification algorithms such as linear discriminant analysis,knearest neighbor classifier,support vector machine,and confusion matrix used to evaluate the algorithm model are studied.Thirdly,using a new gas-oxygen-gas-methane rocket engine as a platform,a laser-induced plasma ignition experiment and an optical diagnosis system were established,and laser-induced plasma ignition experiments under different working conditions were carried out,successfully verifying the application of laser-induced plasma ignition technology The feasibility of this type of rocket engine further expands the application range of laser-induced plasma ignition technology.Based on high-speed camera imaging and emission spectroscopy technology,dynamic evolution images and spectral data of laser-induced plasma ignition were obtained,and four typical flame modes of rocket engines were discovered.The Hu moment,Zernike moment,Krawtchouk moment,feature parameters and the intensity of spectral characteristic peaks of the four flame modal images are extracted and analyzed.The curve of the OH group’s spectral intensity over time was extracted,and it was found that the change of the OH group’s intensity over time can reflect the change of the engine combustion chamber pressure,which provides a new technical means for monitoring the combustion chamber pressure.Finally,based on the Hu moment,Zernike moment,Krawtchouk moment,feature parameters of the flame image and characteristic peak intensity of the flame spectrum,combined with the three classification algorithms of k NN,LDA,and SVM,a flame stability discrimination model is established.The Zernike-SVM model has the best discrimination effect,with an accuracy rate of 99.96%,specificity,sensitivity,and AUC of 0.982～1.The flame stability discrimination model established by the SVM algorithm has the best effect,and the accuracy of the spectrum-SVM model and the feature-parameter-SVM model can reach 99.8% and 99.7%.The accuracy of SVM based on image and spectrum multi-source information fusion can reach 98.5%.This article further expands the application range of laser-induced plasma ignition technology,provides strong technical support for rocket engine flame stability monitoring and active control.It provides reference and experience accumulation for the engineering application of laser-induced plasma ignition and flame combustion stability monitoring based on digital image processing technology.