Study On Ocean Wave Inversion Based On Ship Motions

When a ship is sailing at sea,it is affected by waves and will produce a six-degree-offreedom swaying motion.Excessive swaying motion will threaten the safety of the ship’s navigation.Real-time acquisition of sea wave information in the sailing sea area can provide important environmental information for the ship’s navigation operation decision-making,help the ship to rationally arrange the sailing plan and operation plan,and improve the safety and operation efficiency of the ship.At present,the main ways to obtain sea area wave information are wave buoys,satellites,and shipborne maritime radars.Wave buoys can achieve high-precision measurement of sea waves,but due to its point measurement,inability to move,and high deployment cost,it cannot provide real-time wave data for marine vessels at any location;satellites can provide large-scale wave information for the vessel’s location However,its characteristics such as slow data transmission speed,extremely high equipment cost,and insufficient precision restrict its use as a wave information sensing method that is universally applicable to ships;shipborne maritime radar can realize real-time measurement of waves on board the ship.There is a technical level that requires the "one machine,one calibration" wave measurement verification of maritime radars.Therefore,the current wave sensing methods onboard ships cannot achieve a balance in terms of accuracy,followability,real-time,and cost.In view of the shortcomings of the current mainstream methods,considering the relationship between ship motion and incident waves,using the ship’s own motion to invert the waves to solve the problems of cost,followability,real-time,precision,etc.,this method is also called "buoy analogy method" ".The basic idea is to use the ship as a filter for input(wave)and output(motion)signals,and realize the inversion of ship motion to sea waves by establishing a filter function.The method has two branches: model-based method and data-based method.The model-based method combines the ship’s mathematical model and motion data to construct a filter function to achieve wave information inversion.This type of model-based method mainly focuses on frequency domain inversion;the data-based method directly uses sensor data to pass The machine learning model fits the real filter function and infers wave information based on this.This paper first conducts research based on model methods,establishes a wave inversion model based on Kalman filtering,uses numerical simulation data to verify the results of frequency domain and time domain inversion,and finds that this method can correctly perform wave inversion.Wave inversion is performed based on the model method.The accuracy of the inversion essentially depends on the accurate ship motion response transfer function.However,the uncertainty and change of the ship motion response transfer function in the actual environment make it difficult to apply this method to the actual environment..Aiming at the problem of model-based methods,this paper uses machine learning methods to establish an intelligent wave inversion model based on the fully connected model and the time convolution model,optimizes the two models,and designs a tank test to preliminarily verify the inversion results.On this basis,the generalization of the model under unknown conditions is analyzed and researched,and an optimization method based on data mixing is proposed,which improves the generalization ability of the model and enhances the application value of the model in the actual environment.Finally,analyze the various factors that affect the model based on the water tank test data,and further design the actual sea test to verify the inversion performance of the established model under real sea conditions.The results show that the method proposed in this paper can better establish the ship swaying motion.The mapping relationship with wave information provides a feasible way for accurate inversion of real-time waves on the sea.