Development of a Novel Wave-force Prediction Model based on Deep Machine Learning Algorithms

摘要

The future knowledge of the waves and force is indispensable for themodel identification and the real-time control of ocean engineeringdevices. In order to effectively control the motion of the offshorestructures in a real-time manner, it is required to have an accurate andefficient prediction of the waves. Machine learning has been widelyapplied in ocean engineering field as it offers compromise betweenprediction accuracy and computational cost. The present study focuses onwave-force prediction of offshore structures based on deep machinelearning algorithms. A novel wave-force prediction model is proposed,which makes full use of the efficient processing characteristics of LongShort-Term Memory Recurrent Neural Network (LSTM RNN) andNonlinear Autoregressive Exogenous Feedback Neural Network (NARXFNN) for time series data processing. The relationship between the waveheight and the wave height is non-causal and nonlinear which need futurewave height knowledge for current wave excitation force. Therefore, TheLSTM RNN is firstly utilized for multi-step prediction of the time seriesof wave elevation. The NARX FNN is used to address the model systemidentification between the wave heights and the wave force. Then, theLSTM RNN is further applied to predict the future force of offshorestructures for the real-time control of the structure motions. After that, theproposed deep machine learning algorithm is utilized for wave-forceprediction based on the experimental data obtained in KelvinHydrodynamic Laboratory and the optimal horizon can be specified forthe test model by comparing the performance of different predictionhorizons. The results indicate that LSTM-NARX model can successfullypredict the time series of the waves and force.