Research On Intelligent Processing Of Ship Test Data Based On Recurrent Neural Network

The ship model test runs through the entire process of ship development.The ship model test mainly includes resistance test,propeller open water test,self-propelled test and seakeeping test,etc.A large amount of original test data is generated in the process of ship test and system simulation..The processing of traditional test data mostly uses manual methods,and errors and low efficiency will occur in the actual operation.It is necessary to realize the summary and automatic realization of the method of processing the original data of the ship test;with the large amount of original data in the ship test Accumulation and deep learning research continues to deepen,and it is of great significance to use cyclic neural networks to study ship test data to realize the intelligent processing of ship test data.The main results of this article include:(1)Introduced in detail the research status of ship test data and the development of cyclic neural network,focusing on the rapidity test and seakeeping test data processing methods;(2)Ship test data is a kind of time series data,and the law is hidden in the time series data.This article discusses three deep learning methods for analyzing time series data obtained from ship trials,including the classification and prediction method of ship trial data based on support vector machines,the classification and prediction method of ship trial data based on BP neural network,and the time based on recurrent neural network LSTM.Sequence classification prediction methods are discussed from the principles and specific research methods;(3)Preprocessed the ship test data and summarized the methods,and designed a ship test data preprocessing platform using Matlab GUI.First,perform mathematical statistical analysis on the rapidity test and seakeeping test data;secondly,perform statistical processing and normalization operations on the ship test data;finally summarize the preprocessing methods of different types of ship test data,using Matlab GUI Design a ship test data processing platform and use corresponding types of test data to verify the accuracy and stability of the platform;(4)Build a recurrent neural network LSTM model to classify and predict ship test data,and compare it with the ship test data classification prediction method based on support vector machines and the ship test data classification prediction method based on BP neural network.The test shows that it is based on recurrent neural network.The LSTM model is more accurate.