Research On The Thrust Allocation Of Vessel Based On Improved Particle Swarm Optimization Algorithm

With the discovery and development of deep-sea resources, more and more vessels need to be equipped with dynamic positioning systems. The vessel's Dynanic Positioning System is that the vessel, under the action of control system, suppresses the interference of the external environment to remain a certain position and its heading with its own propelling system. As the important component of the vessel's DPS, the main function of the thrust allocation unit is to confirm each thruster's thrust and azimuth angle to produce the longitudinal, lateral and gyroscopic moment instructions of the vessel's control system.Taking "Hai Yang Shi You 201”,a vessel equipped with seven full-revolving thrusters, as research object, this paper deeply research the thrust allocation of the vessel's dynamic positioning system by simulation method.The main tasks and contents as follows:(1) Building the mathernatic model of dynaumic positiuiiniag systent, including the vessel motion state space model and the vessel motion environmental disturbance model (including the disturbance of wind, wave and current). This paper designs a PID control system,including three control loops: the longitudinal one, the lateral one and the heading. Each loop has a PID controller, and gives the vessel's final expected position and heading in linear superposition way to ensure the instructions issued by control system in a reasonable range.By vessel motion closed loop simulation experiment, this paper verifies that built model and designed PID control system.(2) By building suitable objective functions and constraint condition, this paper builds the thrust allocation optimization model of vessel to minimize the propelling system's energy consumption, the thrust error between propelling system and control system,the thruster's abrasion and other elements. It analyses three reasons for the vessel's thruster's performance degradation, and gives the setting method of each thruster's thrust restricted zone to minimize the vessel's thrust loss according to the interplay among thrusters and the thruster's position.(3) The thrust allocation method of vessel, based on the particle swarm optimization(PSO) algorithm, usually has the problems of low rate of convergence and low solving precision. Therefore, starting with the PSO algorithm's parameters and construction, this paper, combining with multi-agent scheme, puts forward the multi-agent PSO algorithm, in which the agent according its neighboring environment to conducted competition and collaborate mutation operation with its neighboring agents, simulate the evolutionary mechanism of speed and locality in PSO algorithm, and conducts self-learning by employing small scale searching technology to the global optimum agent to increase the capacity of getting exact solutions.(4) The thrust allocation method of vessel, based on the sequential quadratic programming, usually has local convergence, and depends on the deficiency of original value.Therefore, the sequential quadratic programming, acting as local search strategy, is combined with PSO algorithm in this paper, and the sequential quadratic programming and PSO algorithm is put forward. This algorithm, on one hand, keeps the simple and realizable features of the PSO algorithm,on the other hand,it solves the problems of local convergence and the dependence on the deficiency of original value of the sequential quadratic programming to provide the algorithm with the capacity of quick and exact problem-solving.(5) By simulation, this paper verifies two improved thrust allocation methods of vessel,compares and analyzes two improved methods and the fundamental PSO algorithm.