Optimal Classification of Hypersonic Inlet Start/Unstart Based on Manifold Learning

Risheng Zheng,Juntao Chang,Huixin He,Fu Chen
DOI: https://doi.org/10.4028/www.scientific.net/amm.274.200
2013-01-01
Applied Mechanics and Materials
Abstract:Inlet start/unstart detection has been the focus of researching hypersonic inlet, the operation mode of the inlet detection is the prerequisite for the unstart protection control of scramjet. Actually, due to computational complexity and high dimension discrete experimental data, all of these factors are against for the classification of real-time data. To solve this problem, firstly, the 2-D wind tunnel experiment is carried out, inlet start/unstart experiment phenomenon are analyzed; Secondly, isomap algorithm is introduced to reduce high dimensional data , the optimal classification method were obtained with the weighted embedded manifold learning algorithm, At last the superiority of the classification criterion is verified by decision tree algorithm.
What problem does this paper attempt to address?