Optimal classification criterions of hypersonic inlet start/unstart

Inlet start/unstart detection is one of the most important issues of hypersonic inlet and is also the foundation of protection control of a scramjet. To solve this problem, the 2-D inner steady flow of a hypersonic inlet was numerically simulated in different freestream conditions and backpressures with a Reynolds-averaged Navier-Stokes solver using a renormalization group k-e turbulence model; two different inlet unstart phenomena were analyzed. The feature selection of the pattern classification of hypersonic inlet start/unstart was performed based on "numerical experimental" data by the support vector machine-recursive feature elimination algorithm. The optimal classification criterions of inlet start/unstart were obtained with the Fisher linear discriminant analysis by maximizing the between-class distance of the inlet/unstart sample set and minimizing the within-class, and the physical significance of the classification criterions was explained. The idea of classification criterions used in the Central Institute of Aviation Motors/NASA flight test and possible reasons why the control system could not properly sense inlet start/unstart were discussed. In conclusion, it is useful to introduce the support vector machine-recursive feature elimination algorithms and the Fisher linear discriminant analysis to acquire the optimal classification criterions of inlet start/unstart.