The Study Of Model Predictive Control For Shipboard Boom Cranes

As an important engineering equipment,the crane is not only widely applied on land,but also increasingly needed in the marine field,such as cargo handling,maritime engineering(offshore wind farms,drilling platforms,submarine cable/pipeline laying,etc.),wreck salvage,etc.Different task scenarios,complex and changing environment,etc.,lead to the shortcomings of manual operation,such as high training cost,low efficiency,and adverse effects of human factors on safety.Due to the introduction of attitudes of ships or floating platforms,cranes used in the marine field(shipboard cranes for short)have a more complicated structure,more complex dynamics,stronger couplings among states,more significant underactuation and nonlinear characteristics,and cranes are significantly influenced by the external environment(such as wind,wave,and current)simultaneously.It can be seen that the shipboard crane possesses a strong practical engineering background and a wide range of application fields,and presents urgent and realistic demands for the automatic control.Unfortunately,the design of the controllers is very challenging for shipboard cranes due to their own characteristics.It is found that shipboard boom crane is the most common one among various types of shipboard cranes.Therefore,shipboard boom crane is chosen as the research plant in this thesis.Among numerous functions,the stevedoring is mainly considered when one ship is berthing or alongside with other ships.Based on the operation task scenario and the advantages of model predictive control(MPC)in processing multi-variable system,explicit handling constraints and performance optimization,MPC is selected as the main control method to convey the payload to designated position fast and accurately while easing or eliminating its swing and satisfying the constraints as much as possible.Based on the chosen shipboard crane type and operational task,following research and attempts are made to achieve the research objectives:(1)Aiming at the control problem of the shipboard boom crane,this thesis realizes constrained optimization control based on nonlinear MPC.Firstly,various state constraints and input saturation in the shipboard boom crane are fully analyzed and considered,and then the nonlinear MPC mechanism is constructed according to the characteristics of the dynamics and the physical significance,and the control law is determined by solving the corresponding optimal control problem,thus achieving a better constrained optimal control effect.This shows the effectiveness of MPC in handling multi-variable systems,constraints and optimizing control,and preliminarily verifies that MPC method is very suitable for the dynamic characteristics and control needs of shipboard cranes.(2)For the control of shipboard crane affected by disturbances or noises,this thesis realizes state estimation and robust control based on moving horizon estimation(MHE)and MPC.Only considering the disturbances,the robust control of shipboard crane is realized based on tube MPC.Firstly,an MPC controller is constructed for the nominal system to obtain the corresponding nominal state trajectory,and then another MPC controller is constructed to drive the actual state to approach the nominal state trajectory,so as to finally approach or reach the expected value.If there are both disturbances and noises,and some of the states cannot be measured directly,the state estimation is carried out by MHE.On this basis,the MPC is further employed for control with strong robustness.It further proves that the MPC exhibits good applicability to the shipboard cranes.(3)Aiming at the stable control problem of nominal ship crane system,this thesis realizes the asymptotic stability guarantee in theory based on Lyapunov MPC while satisfying the input saturation or state constraint.The Lyapunov function is constructed based on the mechanical energy analysis,and the corresponding controller satisfying the input saturation or state constraints is further obtained.On this basis,the Lyapunov MPC framework is established,in which the contractive constraint of Lyapunov function is introduced to guarantee the asymptotic stability theoretically.Compared with the controller built directly using Lyapunov function,Lyapunov MPC has better coordination and optimization control effect and adaptability,and eliminates the conservation in tackling constraints.When designing the Lyapunov function and its controller,arctangent function and barrier Lyapunov functions are introduced to deal with input saturation and state constraints,respectively.(4)Aiming at the problem of parameter adjustment in the control of shipboard boom crane,this thesis constructs the gain auto-tuning mechanism based on MPC,and realizes the automatic optimal adjustment of controller parameters under the premise of considering the system constraints.The Lyapunov function is constructed based on the mechanical energy analysis,and the corresponding controller is designed.According to the core idea of MPC,the autotuning mechanism for gains is created.The optimal gains of the above controller can be determined online in real time,which greatly improves the efficiency of parameter adjustment and enhances the adaptability of the controller.In the process of gain regulation,the state constraints and input saturation are well considered,and the stability can still be guaranteed.To sum up,automatic control,state estimation,gain adjustment and other functions are realized for shipboard boom cranes by taking MPC as the core idea.Through the combination of MPC functions or the combination of MPC and other control methods,a relatively ideal control effect is achieved for the shipboard boom crane system under different conditions,and the advantages of MPC in handling constraints and optimization are fully utilized.