A floating offshore platform motion forecasting approach based on EEMD hybrid ConvLSTM and chaotic quantum ALO

Abstract

Accurate prediction of floating offshore platform motion (FOPM) is very important to control the movement of the platform and the normal operation of the equipment on the platform. However, the accurate prediction of FOPM is very difficult, due to the coupling effect of the mooring system, operation system, wind, wave, and current. Therefore, in order to obtain more accurate prediction results, firstly, the Convolutional LSTM (ConvLSTM) network is introduced to simulate FOPM nonlinear dynamical system, design input vector coding rules, considering the characteristics of complex time-varying nonlinear and space non-stationary of FOPM. The EEMD is applied for modal decomposition for time series of FOPM to reduce the nonlinearity of time series. Then the FOPM-EEMD-ConvLSTM forecasting model is proposed by designing the input sequence matrix and prediction architecture. Considering the defects of ALO algorithm, based on quantum computing and chaotic mapping, the quantum global search algorithm (QRS) and ant lion trap Chaos reconstruction mechanism (ALTCR)are designed, and chaotic quantum ant lion optimization algorithm (CQALO) is proposed and use to optimate the hyperparametric of FOPM-EEMD- ConvLSTM. Consequently, a hybrid forecasting approach of FOPM integrating FOPM-EEMD- ConvLSTM and CQALO was established, namely FOPM-EEMD-ConvLSTM-CQALO. Finally, the sway and heave data of a floating offshore platform serving in the ocean is used to carry out prediction experiments to test the performance of the proposed new prediction approach. The test results indicate that the prediction model established in this paper has higher prediction accuracy and stronger robustness than the comparison model selected in this paper and the CQALO obtains more appropriate hyperparameters than the comparison algorithm applied in selecting the hyperparameters of FOPM-EEMD- ConvLSTM.