System Identification Study Of Ship Motion Response And Wave Spectral Inversion

With the continuous development of ship and ocean technology,many projects require online acquisition of ocean wave information.For example,the ship dynamic positioning system and the stabilizer system often regard the real-time wave information as feed forward information to compensate the ship motion,which helps to improve the accuracy and efficiency of the control system and reduce energy consumption.The real-time wave information can be applied to predict the ship motion,which helps to improve the safety of aircraft landing or taking off.For large scale model tests which are conducted in ocean waves,real-time information of the wave in the area where the model voyages can help the researchers to understand the hydrodynamic characteristics of water condition,check the accuracy of the tests,and evaluate the seakeeping and maneuverability performance.Therefore,this paper proposes an effective method to obtain real-time information of ocean wave.This study concentrates on developing an adaptive wave calculating system which only use the history-time series of ship motion.Firstly,the AR model of the motion system is built based on the roll or pitch time-series of the ship which navigates in stationary random wave.Through the AR model,the amplitude frequency response operator can be roughly estimated.Three methods to estimate the parameter of AR model is introduced:the recursive least squares algorithm(RLS),the Kalman filtering method and the high order cumulant method.The normalized singular value method is applied to determine model order.Through the data of model tests in towing tank,three kinds of parameter estimating algorithms are compared.Then,the adaptive method is proposed to optimize the initial amplitude frequency response operator and calculate wave information at the same time.The mean square error of the estimated motion spectrum and the real motion spectrum drives the adaptive optimization system.When the optimal value approach,the adaptive system will stop and outputs the estimated wave spectrum.The classical gradient method is used in adaptive recursive optimization.Because the frequency response operator is nearly normally distributed of,the system transfer function can be parameterized by Gauss function.So the optimization problem for the frequency response operator can be simplified as parameter searching problem.The Wiener filtering algorithm is applied as regularized inversion method to calculate the wave spectrum.Finally,through the large scale model test in actual sea,AR model is built by the pitch time series,the amplitude frequency response operator is roughly identified,and the wave spectrum in the voyaging area is estimated by the adaptive optimization system.This application proves the research in this paper is feasible.In this paper,the ship motion signal is applied to build the adaptive optimization system for real-time wave spectrum inversion.This system can serve for the large-scale ship model in the actual ocean wave environment to analyze the motion response and predict the wave spectrum in real-time.Through the research in this paper,it can be seen that the adaptive method can be adapted to different wave states and operating conditions in middle and low sea conditions,which shows good robustness.This method can obtain wave information without using complex wave observation instruments.Thus,this paper has certain engineering application value.