Parametric sensitivity analysis of impact-induced vibration in revolving cannon systems using the PCE surrogate model

Abstract

Abstract

To solve impact vibration problems brought by impacts between sliding plates and rotating chamber structures when revolver shooting processes go on, a parameterized dynamic model which includes gas-driven mechanisms and curved chamber contact relations has been constructed. The accuracy of the model has been verified by means of chamber pressure measurement experiments and high-speed photography experiments. For overcoming the computation restriction problems of high-precision simulated experiments, a sparse polynomial chaos expansion method has been utilized by us for building a high-accuracy substitute model. In combination with the Sobol global sensitivity analysis method, this research has quantitatively carried out an evaluation of the influence that eight key geometric and physical parameters exert upon maximum collision forces. Our results have shown that the elliptical long axis of recoil chambers is the main parameter that contributes 67.2% to the main effect; what comes next is the elliptical minor axis and linear sections, which have contribution proportions of 17.1% and 10.3% separately. Other parameters have demonstrated influences which can be ignored. The optimization model which is gotten from sensitivity analysis has decreased maximum contact forces by 29.73%, thus its validation errors are under 1% via high-fidelity modeling. This study provides a theoretical basis and efficient analytical tools for structural optimization and impact vibration suppression of turret-mounted guns.