【作者】 谢文博；

【导师】 施小成；

【作者基本信息】 哈尔滨工程大学 ， 控制理论与控制工程， 2013， 博士

【摘要】 深海油气管道铺设船舶的动力定位（Dynamic Positioning，DP）控制技术，是铺管船用动力定位系统设计的关键技术。本学位论文结合相关工程需求，从理论上对J型铺管动力定位系统的建模和铺管船动力定位输出反馈控制理论开展了深入的研究，其研究内容主要包括以下几个方面：建立了动力定位船舶运动学、动力学和海洋环境干扰模型。提出一种计算J型铺管作业中管道对船体作用力的悬链线非线性方程方法，将船舶的位置、艏向、运动速度和放管速度作为影响管道力的因素考虑。通过建立船舶、环境和管道模型给出了动力定位船进行铺管作业的仿真模型，并基于模型进行了仿真验证。DP铺管船输出反馈控制器的基础是设计铺管船状态观测器，对于此问题，提出一种鲁棒强跟踪扩展Kalman滤波器（Robust Strong Tracking Extended Kalman Filter，RST-EKF）。该滤波器在DP铺管船模型存在参数不确定性时，利用指数渐消因子放大当前位置和艏向测量值权系数，使滤波器快速收敛；同时，针对船舶位置参考系统和罗经存在野值的问题，引入根据新息不同情况进行分段的稳健矩阵函数，避免了指数渐消因子对野值产生的粗差不合理放大，保证了滤波器的鲁棒性。仿真结果表明，该算法可以保证模型参数具有不确定性时滤波器的收敛速度，同时对新息粗差具有较好的鲁棒性。进一步，为了克服EKF不具有全局稳定性的问题，在系统模型存在结构不确定性情况下，提出了一种自适应滑模无源非线性观测器。对滑模面增益上界，提出一种切换自适应律进行估计，避免了增益过大产生的观测器振荡和增益过小造成的系统不稳定，最后证明了观测误差是全局一致有界的。仿真基于带有非线性水动力矩阵的DP铺管船动力学模型进行，结果表明，观测器能够克服非线性水动力矩阵引入的模型结构不确定性，具有良好的估计精度和鲁棒性。针对铺管船DP作业时的最优跟踪控制问题，提出一种自适应逆最优反步循迹状态反馈控制器。首先针对系统线性化部分设计最优跟踪控制律，再利用线性最优跟踪控制律给出的时变代数Riccati方程结合反步法对系统进行状态变换，并针对慢变海洋环境干扰设计了自适应律进行估计补偿，随后给出了使系统稳定的自适应反步循迹控制律，最后根据逆最优思想，给出所设计控制律作用下的最优性能指标。仿真中利用PID控制器与两组不同参数的最优控制器进行比较，说明了所提出算法的有效性。在此基础上，为了使控制器更适合铺管过程中船舶的行进——航停循环作业工况，提出带有速度校正环节的逆最优反步法状态反馈路径跟踪控制器，该方法能够在线灵活调整期望速度。首先设计逆最优反步路径跟踪控制律使船体渐近跟踪期望路径，然后根据该控制律给出可最小化的性能指标，再设计梯度更新律使速度误差收敛于零。仿真结果表明该算法在实现最优控制的同时，可以满足船舶行进——航停工作模式的要求。针对船舶铺管时管道和船体期望路径不同的问题，假设管道与海底接触点（着地点）到船体间的距离保持不变，并通过坐标转换将管道着地点实际路径与海底期望路径映射为船体实际和期望路径，然后设计了反馈线性化循迹控制器用以验证所给出的坐标变换方法。仿真结果表明上述方法可以控制DP船将管道铺设于给定路径上。进一步，为了使控制器更好的适应DP船行进铺管时船体匀加速——匀速——匀减速的运动过程，提出一种带有加速度校正环节的反馈线性化状态反馈路径跟踪控制器，该方法能够在线灵活调整期望加速度。首先将原DP铺管船二阶系统的控制项求导，得到一个新的三阶增广系统并对其设计反馈线性化路径跟踪控制律，使系统输出渐近跟踪期望路径，再设计梯度更新律使加速度误差收敛于零，并通过调节期望加速度实现了定常速度的误差有界控制。仿真结果表明该算法具有良好的路径和加速度控制效果，可以满足船舶行进铺管时的运动状态控制要求。更多还原

【Abstract】 Dynamic positioning(DP) control method of pipelay vessel is the key technology ofdesigning a ship dynamic positioning control system for laying deep sea oil and gas pipe.Combined with related engineering requirements, this dissertation carries on theoreticalresearch about modeling of J-lay dynamic positioning vessel and output feedback controlmethod of dynamic positioning pipe lay vessel. The main contents of dissertation are asfollows:Kinematic and dynamic equations of DP vessel with sea environment model which isconsidered as the main disturbance are derived. A catenary theory based nonlinear equationmethod is proposed to calculate the pipe forces that acting on J-lay vessel: ship position,heading, motion velocity, rate of turn and the pipelay speed are all taken into consideration. Asimulation system comprised by the above three parts of models is given, and simulation testresults are shown based on the above model.The elementary problem of designing a DP pipelay vessel output feedback controller issystem states estimation. For solving this issue, a robust strong tracking extended Kalmanfilter(RST-EKF) is proposed. An exponent fading factor is added in EKF to form a ST-EKFfor making filter converge rapidly by enlarging the current position and heading weightsunder model parameter uncertainties. Simultaneously, a piecewise stable matrix functiondepending on different innovation situations is involved to enhance robustness when facinginnovation gross error caused by wild values from position reference system and gyrocompass,so one can avoid the unreasonable zooming out of gross error by the fading factor andguarantee the filter robustness. Simulation shows that the designed algorithm can improvestate estimation convergence speed and be robust to gross error from innovation in case ofmodel parameter uncertainties. However, EKF is unable to guarantee the global stability ofsystem. To overcome this weakness, an adaptive sliding mode passive nonlinear observer isproposed with the model structure uncertainties problem. A kind of switching adaptive updatelaw is derived to estimate the upper bound of sliding mode surface, so the estimation vibrationcaused by large gain and system unstability results from small gain situations are bothavoided, uniform global boundedness of estimation errors is proved. Simulation process isbased on DP pipelay vessel model with nonlinear hydrodynamic matrix, it shows that theobserver can give satisfied estimations and be robust to model structure uncertainties causedby unknown nonlinear hydrodynamic matrix. For the optimal following control problem of DP pipelay vessel, an adaptivebackstepping inverse optimal state feedback tracking controller is proposed: first, an optimalpath following controller is designed for linearized parts of the system, time varying algebraicRiccati equation of the linear optimal path following controller combined with backsteppingprocedure are adopted to carry on system transformation, adaptive law is designed forestimating ocean slow varing disturbance, then the adaptive backstepping tracking control lawwhich can make system stable is given. Finally, index function is specified based on thetracking control law with inverse optimal process. In simulation case, optimal controlmethods with two groups of parameters and a traditional PID algorithm are compared to showthe validity of proposed algorithm. Based on the results mentioned above, for the purpose ofgiving a more suitable controller for vessel moving----stoping pipelay situations, abackstepping inverse optimal state feedback path following controller with speed assignmentis proposed, which can adjust desired speed conveniently. Inverse optimal backstepping pathfollowing control law is designed for making vessel body converge to desired path, then theindex function based on control law is specified, finally gradient update law is given to makethe speed error converge to zero. Simulation shows that the proposed algorithm can controlpath and speed simultaneously and give satisfied optimal control effects on vesselmoving----stoping procedure.For the problem of position differences between DP vessel path and and seabed pipe path,the distance between pipe-seabed touch down point and vessel body is assumed to be fixed,then pipelay path is mapped to vessel path by coordinates transformation. A feedbacklinearization tracking controller is used for testifying the coordinates transformation,simulation results show that the above method can make the vessel and pipe both trackdesired path. In addition, for the purpose of giving a more suitable controller to vesselacceleration----constatn speed----decceleration pipelay process, a feedback linearizationstate feedback path following controller with acceleration assignment is proposed, which canadjust desired acceleration conveniently. Control term of DP pipelay vessel second ordersystem is differentiaed to get a new three order augmented system, then feedback linearizationpath following control law is given for this augmented system to force the system outputfollow desired path asymptotically, and gradient update law is designed to make accelerationconverge to zero. Finally a constant speed control strategy is given by adjusting the desired acceleration. Simulation case illustrates the effects on path following and acceleration control,it can make the vessel achieve desired motion states during moving process.更多还原