Analysis And Prediction Of Stowage Plan For Crane Ship Based On BP Neural Network

Crane ships have developed rapidly as a large-scale marine engineering and majorequipment for shipwreck rescue.Crane ships play a great role in the lifting construction of offshore bridges,salvage of shipwrecks,and the construction and demolition of offshore platforms.These construction requirements have attracted widespread attention for the development of large cranes.The operation of offshore engineering is very complicated and it is a large construction project.With the continuous development of offshore engineering,the requirements for crane vessels used in the construction process will become higher and higher.These requirements have also changed the development direction of crane ships,especially in terms of lifting capacity,operating efficiency,safety and stability.Through the mutual allocation of ballast water in each ballast tank of the crane ship,the safety and efficiency of the crane ship during operation can be guaranteed.The efficient and rapid allocation of ballast water plays an important role in the operation of crane ships.In recent years,there are many ballast water allocation schemes obtained through theoretical numerical algorithms.As a new machine learning algorithm,neural network has relatively little research content in this field.By analyzing the prediction of ballast water allocation plan,we know that the prediction is a complex non-linear dynamic system,and the traditional prediction methods and prediction results are not ideal.The BP neural network algorithm is a relatively mature algorithm.It has strong nonlinear approximation ability,adaptive ability and self-learning ability.Compared with other neural networks,the classification and prediction effect of BP neural network is better.This paper first uses the ballast water allocation scheme obtained by the traditional method as the input node of the BP neural network model,determines the optimal network structure of the BP neural network through experiments,establishes the BP neural network model,and compares it with traditional numerical algorithms.Secondly,complete the model development of BP neural network,design and make the human-computer interaction interface of BP neural network model.Finally,the network prediction results show that the prediction system based on the BP neural network crane ship stowage scheme has the advantages of fast convergence speed and high prediction accuracy,and the results have high practical value.However,the prediction system still has some shortcomings.For example,the prediction range is relatively single,and only the ballast water allocation scheme of a certain slewing crane ship is predicted.There are fewer factors to consider,and the accuracy of the prediction has room for improvement.The practicability needs further verification.The next stage of this paper is to study how to increase the application range of the system under consideration of more factors and how to improve the BP neural network algorithm in combination with other algorithms in order to obtain better prediction results.