Research On The Reliability Of Underwater Explosive Vessel Considering Random And Cognitive Uncertainty

The underwater explosion vessel is an important experimental equipment for the study of deep water explosion mechanism and deep water blasting construction technology.The response mechanism of the underwater explosion vessel is very complex because of the coupling effect of hydrostatic pressure,underwater explosion shock wave and bubble pulsation.In order to ensure the safety of the container during use,it is necessary to analyze its reliability.Based on the in-depth study of reliability theory and intelligent algorithm,the dynamic response of containers is intelligently predicted,and the reliability of containers is deeply studied by means of probabilistic and non-probabilistic methods in this paper.The main contents of the study are as follows:Firstly,based on the dynamic response test data of the underwater explosion vessel,the dynamic response prediction model of the vessel was established using BP neural network and the generalized regression neural network(GRNN).The dynamic characteristics of the container response are reflected by introducing the variable of the number of experiments in the model.The prediction results of the two models show that the prediction performance of GRNN model is slightly better than that of BP model.Secondly,based on the in-depth analysis of the uncertainty of the parameters of the underwater explosive vessel,two reliability analysis models are established in this paper:(1)The probability reliability model of the underwater explosive vessel is established based on random uncertainty of parameters;(2)For the performance evolution of the container during service,the random-interval mixed reliability model of underwater explosive containers is established considering the random and cognitive uncertainties of the parameters.Finally,different methods are used to calculate the reliability of the established reliability model.For the structure with implicit limit state equation,the implicit limit state function is reconstructed by using support vector machine regression to obtain an approximate explicit function.The reliability index and failure probability of the container were calculated by Monte Carlo method and First Order Second Momentmethod(FOMS).For probabilistic-nonprobabilistic hybrid reliability model,FOMS is an effective method for reliability calculation when the amount of data is limited.