【作者】 孙秀军；

【导师】 王树新；

【作者基本信息】 天津大学 ， 机械电子工程， 2011， 博士

【摘要】 混合驱动水下滑翔器集成了自主水下航行器、水下滑翔器及剖面浮标的优点,是海洋环境观测和海洋资源探测的重要平台。滑行操作模式下,周期改变自身净浮力和重心位置,借助机翼的液动升力前进,完成海洋中的锯齿形剖面运动,此时具有工作效率高、范围大、噪音低的特点;推进操作模式下,通过螺旋桨的推力快速前进,借助尾舵或内部姿态调节实现转向,完成海洋中的快速穿梭运动,此时具有速度快,机动性高,受海水深度及流速影响小的特点。本文在混合驱动水下滑翔器系统设计的基础上,利用浮基多刚体理论推导了其动力学方程,定制了水动力模型和参数,分析了其机动性与稳定性,完成了三维空间的滑翔运动和推进运动的仿真。而且,本文设计了针对该混合驱动水下滑翔器的神经网络运动控制器,并仿真验证了其稳定性和可行性。本文主要研究成果和创新点为:1.研究设计了首台小型混合驱动水下滑翔器样机,吸收了传统自主水下航行器、典型水下滑翔器和ARGO自持剖面浮标三者的优点。给出了混合驱动滑翔器的设计流程和计算方法,提供了部分单元试验及水域试验的重要数据,为混合驱动水下滑翔器的设计和制作提供参考。2.采用浮基多刚体理论和鱼雷水动力学模型建立混合驱动水下滑翔器的动力学方程。该方程考虑了内部质量的平移和旋转对姿态的控制作用以及高机动性运动时所受水动力的影响。分析了混合驱动水下滑翔器的机动性和稳定性,并给出了设计参数的选择范围,同时完成混合滑翔器在三维空间的推进操作和滑行操作模式下全状态运动仿真。3.针对混合驱动水下滑翔器裁制了基于BP神经网络的动力定位和轨迹跟踪控制方案,并构造李亚普诺夫函数分析了其控制稳定性。同时定制了基于PID神经网络的定深运动,目标跟踪和姿态解耦控制器,仿真验证了控制方法的可行性,提高了自适应控制和容错能力。更多还原

【Abstract】 The winged hybrid-driven underwater gliders combine the best features of conventional autonomous underwater vehicles, legacy underwater gliders and Argo free-drifting profiling floats, and can be served as important platforms for marine environment observation and ocean resource exploration. Actuated by net buoyancy and hydrodynamic lift from fixed wings, the hybrid gliders perform glide operation and roam the ocean in a zigzag path with high efficiency, long endurance and low noise. Driven by one single prop and steered with deflection of tail fins or redistribution of internal masses, the hybrid gliders undertake thrust operation and penetrate the strong current in the estuary’s littoral zone with high maneuverability and burst speed.In this paper, the architecture desigh and the operating principle of the hybrid gliders are introduced, the simplified and accurate dynamic models are built by using the theorem of multi-rigid-body with a floating base, the stability and maneuverability are analyzed and the three dimensional motion simulation under both thust operation and glide operation are performed. According to the paticular hydrodynamic profile, the hydrodynamic model was researched and the appropriate pramenters are seleted and calculated. In addition, the control model of the hybrid vehicle is developed based on the neural network algorithm, and its feasibility was verified by the control simulation. The main contributions are summarized as follows:1. One mini prototype of the hybrid-driven underwater glider is designed which tend to integrate together all the advantages of the conventional autonomous underwater vehicles, legacy underwater gliders and Argo free-drifting profiling floats. The critical design and calculation is presented and the fundamental data of the unit experiment and field tial are given which can serve as guidance for the fabrication of the marine platforms.2. The dynamic model of the hybrid gliders is built by using the therom of the multi-rigid-body with a floating base, and the impacts of the translational and rotational motion of internal masses on its attitude are taken into consideration. The hydrodynamic efficiencies of torpedo are introduced to add the accuracy of the dynamic model of the hybrid glider. Based on the analysis of the stability and maneuverability and the three dimensional motion simulation in both thrust operation and glide operation, the ranges of the design parameters of the hybrid-driven underwater glider are proposed.3. The BP neural network based controller for dynamic positioning and trajectory tracking of hybrid driven underwater glider are developed and anaylized, and the PID neural network based controller for fixed-depth navigation, target tracking and attitude decoupling are tailored. The capability of adaptive control and fault tolerance are brought about by the neural networks based control method.更多还原