Prediction about the vessel’s heave motion under different sea states based on hybrid PSO_ARMA model
Résumé
In order to solve the problem of control performance degradation caused by time delay in the wave compensation control system, the vessel heave motion displacement needs multi-step predictability because different wave compensation control systems have different delay response times. Inspired by the flexible and powerful optimization capability of the Particle Swarm Optimization (PSO) algorithm, this paper proposes a hybrid model of PSO algorithm and autoregressive moving average (ARMA) model to predict the vessel heave motion at different sea states. The fitness function for the ARMA model parameter optimization problem is formulated through the error function created by residuals squares sum, and the optimization of the ARMA prediction model parameter value is carried out with an effective global search technique based on the PSO algorithm. Simulations are performed with filed data, the results show that the PSO algorithm optimizing the ARMA model parameters to predict vessel heave motion is effective and it can get satisfying precision. In addition, the PSO_ARMA prediction model has excellent robustness, it can not only adapt to the prediction of vessel heave motion under certain conditions but also adapt to the prediction of various sea states.