The Study Of BP Neural Network Optimized By Intelligent Algorithm In Wave Height Prediction

The South China Sea is an important maritime area along the 21 st Century maritime Silk Road.The study of ocean waves is of great significance for the construction of marine engineering,marine development,shipping operations and marine fishery.However,ocean waves are complex and often accompanied by dangers.Rapid and accurate wave height prediction can understand the sea conditions in advance and provide relevant data,so that correct decisions can be made to ensure the smooth operation of sea navigation and the safety of personnel.In this paper,the BP neural network optimized by intelligent algorithm is used to predict wave height in the South China Sea.The main research contents are as follows:(1)The WAVEWATCH-?(WW3)model was used to simulate the wave heights of the South China Sea from 2013 to 2017.In the absence of relevant area buoy data,the accuracy of WW3 model simulation results was verified by cross-comparing with altimeter data,and the wave climate in the South China Sea was analyzed.The simulation results show that there are regional differences in wave height in the South China Sea,which is higher in the north and lower in the south.The WW3 simulation data provide a data base for further prediction of wave height.(2)Combined historical wave height data and wind data constitute different parameter characteristics.The accuracy of BP neural network prediction results of South China Sea wave height under various input parameters is verified.The results show that the prediction accuracy is improved with the increase of parameters.However,the BP neural network is prone to fall into the local minimum,and the results need to be further improved.(3)To overcome the shortcomings of BP neural network,genetic algorithm,particle swarm optimization algorithm,cuckoo algorithm and beetle antennae search algorithm were used to optimize weights and thresholds of BP neural network respectively,and the prediction accuracy variation of each combination model with different input parameters is analyzed.The prediction results show that four intelligent algorithms can improve the prediction accuracy of BP neural network,and the prediction results of CS-BP model are the best.