【作者】 张强；

【导师】 张显库；

【作者基本信息】 大连海事大学 ， 交通信息工程及控制， 2018， 博士

【摘要】 为了解决欠驱动船舶港内操纵时受限水域、不确定数学模型、相对较强环境扰动下自动靠泊控制器设计的关键性问题,本文以“船舶自动靠泊简捷非线性鲁棒控制”为选题,充分考虑航海实践的要求,进行系统性的理论探索研究。期望为船舶靠泊辅助支持系统或智能船舶全航线自动控制提供理论指导和技术支撑,从而进一步提高靠泊效率,降低安全风险。本文分别针对船舶自动靠泊控制任务中的港内操纵运动数学模型、控制理论算法、控制工程应用3个关键点进行深入探究,采用“简捷”+“鲁棒”+“自适应”的技术路线,即控制律要简捷、有效,易于工程实现,控制器能够在外界环境扰动和模型参数摄动有界条件下具有鲁棒性,在扰动和模型参数不确定下具有适应性。目标是通过解决船舶港内靠泊操纵控制中所涉及的几个难题,实现考虑船舶控制工程实践需求(执行器输入饱和、算法实时性等)的自动靠泊控制。本文首先针对船舶港内操纵时操控性和连贯性能变差,存在低速、大漂角运动,且风、流等干扰相对增大的特点,充分考虑浅水、低速、风、流等港内环境,以及船体、舵、桨等各部分的相互干扰和倒车特性,以应用广泛的右旋单桨船为研究对象,进行模型优化和参数修正,并利用三次样条插值算法对在常速域(漂角小于20°)与低速域(漂角大于30°)之间的过渡速域的流体力进行了计算,给出了一种适用于船舶港内操纵的、考虑右旋单桨倒车特性的船舶运动数学模型。最后通过对某液化气(LPG)船的仿真试验(旋回试验、Z型试验、停船试验、浅水试验、低速试验),验证所建模型的有效性。在控制算法方面,针对欠驱动船舶自动靠泊控制中模型动态不确定和有界扰动未知等问题,采用附加控制方法进行坐标转换,解决了欠驱动问题,并利用航海动态深度信息的神经网络自适应方法重构不确定的模型动态与未知扰动;针对港内螺旋桨非定常转速下航向保持控制问题,通过非线性环节来构造状态反馈或输出反馈的反馈(或修饰)方式,提出了基于非线性函数驱动的简捷非线性反馈和非线性修饰算法,并利用李雅普诺夫直接法及闭环系统响应分析法进行稳定性分析。在工程应用方面,为了加强船舶港内操纵与控制的安全性(从减小舵幅和舵机的负荷角度考虑),利用易于工程实践的闭环增益成形算法设计PID控制参数,选取双极性S函数驱动的非线性反馈技术设计控制器;为解决多参数自动靠泊神经网络控制器的计算负荷问题,提出一种考虑靠泊实践要求的简捷神经网络控制器设计的优化方法,该方法通过优化网络输入参数和有效信息提取频率,减少了计算维数;为解决自动靠泊自适应神经网络算法的计算负荷大的问题,在考虑到舵、桨执行器的输入饱和的基础上,利用动态面(DSC)技术与最小学习参数(MLP)的方法降低计算负载度,易于工程实现。本文所有实验采用Matlab/Simulink编程实现,验证了控制器设计方法的有效性。该研究对于推进高技术船舶制造业的发展具有重要的现实意义,为建设低碳、节能、环保和安全的海上交通运输奠定了重要的理论基础。更多还原

【Abstract】 In order to solve the key problems that the design of underactuated ship automatic berthing controller under the circumstance of the restricted waters in port,the uncertainty of mathematical model parameters and the relatively strong environment perturbation,this research topic is selected as "concise and nonlinear robust control for ship auto-berthing".The theoretical exploration based on the requirements of navigation practice has been carried out systematically in this thesis,which provides the theoretical guidance and technical support for ship berthing support system or automatic full-routed control of intelligent ship so as to improve the berthing efficiency and reduce the safety risk.This research work includes 3 key aspects:ship maneuvering mathematical model in port,control algorithm and control engineering applications.By employing the concise idea,robust control and adaptive control,the control low can be concise,efficient and easy to implement,and the controller has robustness under environmental disturbances and the perturbations of model parameters,at the same time the control parameters have adaptabiliy under uncertainties of disturbances and model paramerters.The objective is to realize the automatic berthing control considering the practice requirements(actuator input saturation,real-time algorithm,etc.)by solving several key problems for ship maneuvering and control in port.A kind of practical ship maneuvering modeling group(MMG)model is discussed.The correction of shallow water,low speed and external environmental disturbances together with the mutual interference of the hull,rudder,propeller are taken into account in this model,which introduced the practical reversing propeller thrust mathematic model reflected the characteristics of the right-handed single-screw propeller.Moreover,the cubic spline interpolation algorithm is used for the interpolation calculation of hull fluid force when drift angle between 20° to 30°.The effectiveness of the proposed model is verified by the simulation of a LPG(Liquid Petroleum Gas)ship(turning test,zigzag test,crash astern test,shallow water test and low speed test).In order to solve the uncertainties of disturbances and model paramerters of the underactuated ship control,an additional control method is adopted for the underacruated problem,while the navigation dynamic deep-rooted information based the robust neural networks(NN)adaptive approach is used to reconstruct the dynamic and unknown disturbances of uncertain models.In order to solve the problem of course keeping control with the propeller stopping or reversing at slow speed for berthing,a concise nonlinear feedback and nonlinear decoration algorithm driven by nonlinear function is proposed by using a feedback(or nonlinear decoration)method of state feedback or output feedback.Then,the stability analyses are carried out by using the Lyapunov direct method and response analysis method of the closed-loop system.In order to enhance the safety of ship maneuvering and control in port(reducing the rudder angle and the load of steering gear,etc),a novel nonlinear feedback PID(Proportional,Integral and Derivative)algorithm,combined with the closed-loop gain shaping algorithm and the nonlinear feedback technology driven by bipolar sigmoid function is proposed;In order to reduce the computational load and shorten the training time of the multi-parameter NN auto-berthing controller,a concise NN controller is proposed by optimizing input parameters and extraction frequency of effective information;In order to solve the computational load problem of adaptive NN auto-berthing control with input saturation of control actuator(rudder,propeller,etc),the dynamic surface control(DSC)technique and the minimum learning parameter(MLP)are used for the engineering realization easily.All the experiments in this thesis are implemented by using Matlab/Simulink to verify the effectiveness of the controller designed.This research has important practical significance for promoting the development of high-tech shipbuilding industry,and has laid an important theoretical foundation for building the low carbon,energy saving,environmental protection and safety marine transportation.更多还原