Integrated Control Of Path Following And Roll Motion Reduction For Marine Vessels

Marine surface vessels play an im portant role in marine transportation. In order to reduce the work strength of helm sman, shorten the voyage distance, save fuel, and ensure safety, it is essential to control s hip motion for accomplishing the transportation mission efficiently and safely.Ship motion has characteristics of lar ge inertia, time delay and strong coupling. The mathematical model of ship motion varies as sailing status(speed, heading, etc.) and external disturbance(wind, waves, a nd current, etc.) change. Thus, the ship motion control is a strong nonlinear, uncertain, and constrained problem. It is necessary to develop appropriate methods for this problem. The heading control and path following belong to the low frequency m otion control from the perspective of control objective, while roll m otion reduction belongs to the high frequency m otion control. The path following and r oll motion reduction w ere, traditionally, treated separately in the previous studies. However, roll motion could cause negative effects on the marine surface vessel during path following in seaways, and path following actions could cause undesired roll motion. It is m eaningful to study the integrated control of path following and roll motion reduction.Four parts are included in the thesis, na mely, the ship heading control, path following, integrated control of path fo llowing and roll m otion reduction by rudder, and integrated control of path followi ng and roll m otion reduction based on rudder and fin. In addition to their comm on features, each application has its own features. According to these features, different control theories and methods are applied.The first part of the thesis(Chapter 3) deals with the sh ip heading contro l, including course keeping and course cha nging, which is a singl e-input-single-output(SISO) control problem, and many researchers have contributed considerable results. The motivations of this study are as follows : firstly, it serves as an exam ple to illustrate the design procedure for dif ferent control methods. Secondly, most of the previous studies aimed at one specific c ontrol method and did not compare dif ferent control methods under sam e simulation environment. In this part, dif ferent control laws are proposed based on dif ferent control methods, and several simulations are carried out to illustrate the merits and demerits of each control method. The simplified controller design model for the ship heading control is established, and then sliding mode, integrated backstepping and neural network, MPC(model predictive control) methods are applied into design dif ferent control laws for the ship heading control. PID(Proportion Integration Differentiation) method is still widely used in engineering practices; therefore, its performance is used as the benchm ark. The linear quadratic regulation(LQR) and MPC are both optimal control method. In order to illustrate the advantages of MPC, the LQR control law is also tested. U sing a 4-Do F(degrees of freedom) high fidelity model as simulation model, five d ifferent control laws are simulated. The comparisons of simulation results illustrate the merits and demerits of each control law.The second part of the thesis(Chapter 4) deals with path following, including straight line path following and curve line path following. A ship is normally equipped with two actuators: rudder and propeller, while longitudinal displacement, transverse displacement, and yaw angle are the control outputs. Obviously, the dimensions of the control inputs are fewer than the dimensions of the control outputs, therefore, the path following is a classical underactuated problem. By combining Serret-Frenet frame and LOS(Line of Sight) m ethod, the control out puts of path f ollowing are transformed into cross-tracking error and heading angle e rror, and then a sim ple controller design model is established. In this way, the lim itation of underactuated characteristic is overcome and some classical control m ethods can be applied, a nd the robustness is also strengthened in relation to pure LOS method. Sliding m ode, integrated backstepping and neural network are em ployed to design dif ferent control laws for path following. The perform ance of PID c ontrol law is also included as the benchmark. Using a 4-Do F high fi delity model as simulation model, three different control laws are simulated. The comparisons of simulation results illustrate the merits and demerits of each control law.The third p art of the thesis(Chap ter 5) deals with integrated control of path following and roll motion reduction by rudder. From the perspective of path following, the control objective includes straight li ne path following and curve line path following; from the perspective of ro ll motion reduction, the contr ol objective includes the reduction of roll angle and roll rate. Based on the tr ansformation in the second part, the objectives of this problem are reduced to trac king error, heading angle error, roll angle, and roll rate. The control i nput is only the rudder whose amplitude and rate are constrained. The objectives of path following and roll m otion reduction are closely co upled; therefore, it is an optim ization problem of m ultiple objectives with constraints. The MPC m ethod is a natural control m ethod, given its advantages in dealing with m ulti-objective problem with constraints. Through the simulations using a 4-Do F hi gh fidelity m odel, the perform ance of the integrated system is evaluated by sim ulation. The trade-offs between path following and roll motion reduction are studied, performance sensitivity to key design parameters are analyzed, and the gu idelines for p arameter selection are d iscussed. The simulation results validate the effectiveness of the controller.The fourth part of the thes is(Chapter 6) deals with integrated control of path following and roll motion reduction by rudder and fin. The control objectives of this problem are the same as those of the third part; but the controllers are now rudder and fin. Three strategies are proposed in this pa rt:(1) the first strategy is nam ed as “separated strategy” in which the pa th following and roll m otion reduction are completely two loops. This formulation contains two scenarios: the first one uses PID method to get control law for r udder and fin; the other one uses integrated backstepping and neural network to desi gn control law for r udder and integrated sliding mode and neural network to design control law for fin.(2) The second strategy is named as “compact strategy” in which the rudder and fin are cooperated to control the path following and roll m otion reduction simultaneously. The control laws are designed using MPC method.(3) The third strategy is named as “sem i-separated strategy”, in which rudder is used to cont rol path following and roll m otion control, and fin is purely used to reduce of roll motion. Two scenarios are included, both of them use the MPC method to design rudder control law while the first control law for fin uses the sliding mode m ethod and th e second control law for fin uses the integrated sliding mode and neural netw ork. Through the simulations using a 4-Do F high fidelity m odel, the performance of the integrated system is evaluated by simulation. The comparisons of simulation results illustrate the metrits and demerits of each control strategy.Various control strategies are proposed in the thesis for the integrated control of path following and roll motion reduction pr oblem. The simulation result shows the feasibility and efficiency of the proposed c ontrol strategies and their advantages over the separated control strategies. The result s of this study provide the strong support for marine surface vess els to accomplish the transportation mission efficiently and safely.