【作者】 王仲；

【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， 2008， 硕士

【摘要】 现在动力定位(Dynamic Positioning,DP)系统在研究和应用上都已经非常成熟,但关于相对水面动目标的动力定位技术在国内还不曾具体研究过。本文提出了相对漂浮船舶动力定位的跟踪控制的概念,不是传统的定点定位和循迹跟踪,而是要求定位船跟踪一个在海面上自由漂浮的目标船,使两船始终保持相对固定的位置和艏向,并对该技术进行初步研究。本文建立了动力定位船舶的低频运动和高频运动数学模型、推进器模型、测量模型,同时对风、浪、流等外界环境扰动的数学模型进行描述。自由漂浮船采用了与定位船相同的数学模型进行仿真。给出了定轴推力器与全回转推力器的布置以及推力分配策略。介绍了数字式PID的控制算法,给出了神经网络的学习方式、网络结构等有关概念,着重分析了有关BP网络和对角回归网络(DRNN)的特性。结合相对自由漂浮运动船舶的动力定位系统设计了基于神经网络的BP—PID和DRNN—PID两种控制器用来对定位船加以控制。最后结合所设计的控制器进行仿真研究。同时将DRNN—PID控制器跟BP—PID控制器的控制效果进行比较。最后,通过对系统的仿真结果可以看出,DRNN—PID控制器具有控制精度比较高、响应时间比较短及稳定性比较性强的特点,对不同海流的干扰具有一定的适应能力。同时也验证了本文系统模型的正确性和准确性,表明该控制方法具有一定的理论意义及实用价值。更多还原

【Abstract】 Nowadays, the research on dynamic positioning system (DPS) is becoming more and more advanced. But the dissertation about the offshore target with random movement is not involved. The offshore ship inferred is freely floating in the control of the environment effect. The dissertation puts forward the idea about the research on the tracking control of the dynamic positioning system relative to a freely floating vessel. It keeps the DP ship with floating ship in the fixed location and heading, while the traditional DPS characteristic is that the DPS offshore target is a fixed location or track. If the method could be used in the practice, it will contribute to the exploitation of the ocean and the aggiornamento of the navy.A simple motion low frequency and high frequency mathematical model of vessel are established. The thruster, measurement and environment including wind wave and current models are also established. The floating ship adopts the model as the same as the DP ship and is simulated in the control of the environment.After the presentation of the digital PID algorithm, the neutral network framework is introduced, and the BP algorithm Diagonal Recurrent Neural Network’s characteristic is presented. As the new DPS is considered, the application of BP based PID and DRNN based PID hybrid inversed control techniques is designed and analyzed. Comparing the arithmetic of hybrid control, we can see that the DRNN—PID controller has the feature of high precision, quick-response, and high stability, and it also can be used in numerous environment. Finally, the correctness and accuracy of the DP model are validated, which showed that this method has high value for theory and utility.更多还原