Research On State Estimation Of Ship Dynamic Positioning System

With the increasing exhaustion of land resources, human will gradually put theeye of development resources and energy to the marine field. The research anddevelopment on the marine areas has expanded from shallow to deep. Complexenvironment of the deep sea put forward higher requirements to oil and gasexploration ship and ocean development platform equipment. It also dramaticallyimprove the requirements for ships and offshore platforms in the ocean positioningwork, with the depth increases traditional anchoring equipment highlights installationcosts soared, poor precision work and other disadvantages. Dynamic positioningsystem(DPS) has the advantages of without depth restrictions, put and evacuatequickly, high positioning accuracy, more and more scientific research institutionsincrease the study and input of ship dynamic positioning system.The DPS is a closed loop control system, which rely on its own power plant tocounteract the wind,wave, current and other environmental loads, so it can keep at apredetermined position or follow the route set sail. Thus, the low-frequency locationinformation plays a key role in the entire ship DPS.This article mainly research state estimation algorithm and simulation, the majorwork is as follows:1)Firstly, we introduce the six degrees freedom of ship motion in surface, thenfocusing on the three degrees of freedom in the horizontal plane, and giving the shiphigh and low frequency motion mathematical model. Then the calculation methods ofwind, wave and current loads were introduced, which can add interference term forsubsequent simulation work to simulate real ship motion. Finally, introducing threekinds of more mature control algorithm (PID, LQ, Backstepping) and they all connectwith low-frequency motion of the ship being composed loop control for simulation,offering importation to state estimation algorithm.2)Secondly, elaborating Kalman and Passive two kinds filters from the theoryand filtering process. Then, we simulate in the same condition of input andenvironment. We observe the filtering effect from a qualitative point of view,analysising the affect results of parameters to each filter, and comparativing the advantage and disadvantage of this two filters.3)This article provides two improvement strategies to Kalman filter: Fadingmemory and H∞filter. Firstly, introducing the characteristics of these two filteringalgorithms and processes in theoretically. Then, analysising the two filters’ effectunder the same input conditions in qualitative and quantitative separately. Finally,these two filters and Kalman filter are all in the same parameters, adjust specificparameters of both filters, analysising of these two filters’ advantage compared withKalman filter in quantitative.