代理模型在水下航行体空泡压力预示的应用研究

摘要

文章基于径向基神经网络(Radial Basis Function,RBF)和Kriging代理模型方法分别对水下垂直发射航行体表面附着空泡闭合区压力预示模型进行了研究,首先分析了影响空泡闭合区压力变化的物理因素,通过代理模型方法建立了影响因素与空泡闭合区压力最大值和压力空间分布波形之间的数学模型,形成了空泡闭合区压力分布预示方法,而后针对典型工况开展预示结果与试验数据比对,表明RBF方法和Kriging方法均能较为准确地获取空泡闭合区压力特征,在有限子样条件下Kriging代理模型对空泡闭合区压力峰值预示精度更高.%The Kriging surrogate method and the basic radial basis function(RBF) interpolation surrogate model were adopted to analyze the unsteady cavity pressure development of underwater vehicle vertical launching. And, based on the analysis of the influence factors of cavity pressure, using the experiment da-ta,the surrogate models between effect factors and cavity pressure and the described model of pressure distribution were determined by the approximate methods. Consequently the prediction method of unsteady cavity pressure based on the two surrogate models were put forward.The prediction results of typical cases agreed well with experiment results,which shows the validity of the prediction method used in cavity pres-sure development by the two surrogate methods,and the Kriging surrogate method exhibits the better pre-diction accuracy based on limited samples.