Abstract: Considering the influence of various uncertainties, the parametric kinetic model of stage separation is constructed to deal with the problem of refined analysis under the condition of high dimensional uncertainties. The typical separation failure modes are identified and the reliability assessment model of vehicle separation mission is constructed. The active subspace method is employed to realize the dimensionality reduction of high dimensional uncertainty parameter space to deal with the problem. Combined with the maximum minimum distance sequential sampling method, the high order polynomial regression model with cross terms is used to achieve the reliability analysis in the dimensionality reduction space. It is shown that the influence of high dimensional uncertainty on separation reliability can be described exactly. Higher efficiency of reliability analysis is achieved under the condition of ensuring high dimensional uncertainties calculation accuracy. The separation reliability curves corresponding to different thresholds are consistent with the results of Monte Carlo method, which can further support the detailed design of the vehicle separation considering high dimensional uncertainties.

Key words: Launch vehicle, Stage separation, High dimensional uncertainties, Reliability assessment