Dynamic Response Prediction And Fuzzy Reliability Analysis Of Deepwater Explosive Vessel

The deep-water explosive test vessel is an important experimental equipment to study the mechanism of deep-water explosion and deep-water blasting engineering technology.As a dangerous domain-limiting device,to ensure its safety during service,it is necessary to conduct a reliability analysis of the deep-water explosive vessel.Due to the complexity of the response law of the container under the coupling of deep-water explosion shock wave and hydrostatic pressure,it is difficult to accurately describe it analytically,and the performance parameters of the container not only have random uncertainty,but also cognitive uncertainty as the service time increases.Therefore,this paper uses a machine learning intelligent model to predict the dynamic response of containers,and on this basis,a structural fuzzy interval reliability model and a random-interval hybrid reliability model are established respectively to analyze the reliability of the deepwater explosion test vessel.Firstly,after sorting out the dynamic response test data during container service,the recurrent neural network,long short-term memory neural network and gated recurrent neural network were used to establish the container dynamic response prediction model,and the comparison and analysis of the prediction results showed that the gated recurrent network had the best prediction performance.Secondly,due to the various uncertainties in the performance parameters of deep-water explosion test vessel,two reliability analysis models are established:(1)Considering that the ambiguity of parameters is an uncertain factor affecting structural safety,the fuzzy interval reliability analysis model of deep-water explosion test vessel can be established by using the parameters as fuzzy variables.(2)Based on the random and cognitive uncertainty of the parameters,the random-interval mixed reliability model of the deepwater explosion vessel is established by taking the parameters as random variables and combining the dynamic strain prediction interval of the explosion test vessel.Finally,for the reliability calculation of the two established reliability models,the fuzzy interval reliability analysis method can be used to introduce the structure-related uncertainty parameters into the structural fuzzy reliability calculation to obtain the interval variation range of the structural fuzzy reliability,which can provide more useful information about the structural safety degree than the traditional reliability calculation method.The random-interval mixed reliability method is used to convert random variables into interval variables by membership function,which can more intuitively quantify the reliability of the structure.