Research On On-line Prediction Of Ship Motion Posture

During the development of marine resources and special operations,the shipborne stable transportation platform is used to transport personnel and equipment.In order to ensure that the stable platform can timely compensate the ship motion posture and keep itself stable,and also to ensure that the stable platform can fully compensate the ship motion posture during the transportation period,this paper studies the on-line prediction of the ship motion posture in the extreme short-term and quiescent period respectively.The main research contents are as follows:Firstly,the law of wave motion is analyzed,and the ITTC wave spectrum is used to simulate the wave motion under different sea conditions;the ship is modeled by AQWA software,and the RAOs and phase differences of the ship’s response to the wave are obtained,and the motion of each degree of freedom of the ship under different sea conditions is obtained.Then,the extreme short-term on-line prediction of ship motion under the four level sea state is studied.The feasibility of using AR model and RBF neural network model to forecast is verified.In order to solve the problem of time-consuming in parameter estimation of AR model,the recursive least square algorithm with limited memory is used to update the parameter estimation;in order to solve the problem of difficult selection of RBF neural network model parameters,the improved particle swarm optimization algorithm is used to optimize.At the same time,in order to reduce the amount of calculation,the influence of different sampling periods on the prediction results is analyzed;in order to make the prediction data and the system control data have the same frequency,the fitting and interpolation ability of RBF neural network model is verified,and the AR model prediction and RBF neural network model fitting and interpolation model are proposed for extreme short-term prediction,which improves the real-time performance of the prediction algorithm.Finally,the on-line prediction of the quiescent period of ship motion is studied.The method of ship motion data prediction directly is used to predict the quiescent period,and the effective prediction time is short.In order to extend the effective prediction time,the method of ship motion data envelope prediction is used to predict the quiescent period,because of the characteristic period of the envelope is longer,in the extraction process of the envelope,the method of limiting the extreme value is proposed to reduce the nonstationary of the envelope.At the same time,the EMD+AR combined model is used to predict the quiescent period,which improves the safety factor.