Analysis And Prediction Of Ship Impact Environment Based On Intelligent Algorithm

Since the beginning of the 21st century,big data has become a resource depression for human competition.Intelligent technology based on big data has become the development trend of all industries in the future.As the backbone of national defense,the design performance of surface warships has been concerned by people.In recent years,with the high attention of various departments,people have carried out a lot of research work on the impact resistance of various types of ships.For this reason,with the high attention of various departments,people have carried out a lot of research work on the impact resistance of various types of ships.For this reason,a large number of real ship explosion experimental data and mathematical model calculation data have been accumulated.There is a huge value for mining hidden in these data.However,there is still a lack of effective methods and experience to complete the big data of ship impact environment Mining and exploration.For this reason,this paper uses the neural network model which is more mature and has better performance to explore the mining of big data of ship impact environment.The specific research content of this paper is as follows:(1)In this paper,we use the technology of design mother ship and APDL language to develop the corresponding parametric modeling software,through which we can quickly complete the development of the corresponding virtual ship;secondly,we complete the calculation of the massive impact environment conditions of the virtual ship and the real ship through our own Python small program;in this paper,we also calculate the ship type parameters,the calculation condition parameters and the calculation Results the parameters were selected,and MySQL software was used to design the architecture of the ship impact environment database,and based on the above architecture,the ship impact environment database was built.(2)In this paper,by analyzing the data distribution characteristics of various parameters of ship impact environment,comparing the sensitivity of PNN and RBF networks to different characteristics of data,we choose to use PNN model to mine and train spectral acceleration data,and use RBF network to mine and train spectral velocity and displacement respectively.At the same time,this paper also analyzes the scale of data and the influence of different data normalization methods on the prediction accuracy of the training model.(3)By analyzing the structure and characteristics of particle swarm optimization algorithm,gradient descent algorithm and clustering algorithm,this paper designs the network design of the three prediction models,and completes the construction of the ship impact environment database based on the above architecture.According to the characteristics of the related parameters,the training and prediction of the related network model are completed by adjusting the important parameters of the algorithm.(4)Through the prediction results,this paper analyzes the distribution rule and attenuation trend of ship impact environment under the parameters of ship scale and calculation conditions,and gives a brief prediction formula of ship spectral velocity.Finally,based on the ship database,this paper analyzes the rationality of the impact design input setting in the standard.