自主式水下机器人推进器弱故障诊断方法研究

Autonomous Underwater Vehicle (AUV) operates in the complex ocean environment without human intervention and cables. And thrusters are the most common and important force-providing device for underwater vehicles. Thruster fault diagnosis technology has played an important role on the AUV’s safety. However, the feature of weak thruster fault is faint, which also has only a few differences from the feature caused by external disturbance such as ocean current. Nowadays, there is no mature theory and unanimously accepted solutions for the weak thruster fault diagnosis in AUVs..At first, the project researches on the coordination between the transition frequency of Brownian particles and switching frequency of potential well. Based on energy transduction method, the project will obtain the mechanism of weak fault feature enhancement and disturbance suppression. Then with respect to statistical variance reconstruction of fault feature signal, the project plans to receive a solution to enhance the fault singular signal once again. Furthermore, as for the general projection synchronization among the fault information with different dimensions, the project is expected to achieve the nonlinear mapping relation between the fault feature matrixes and the magnitudes of fault based on the simultaneous lifting method for multi-source information. At last, the project researches on the confirmation problem for the fault identification results, and it obtains the coordinative operation mechanism between fault diagnosis and fault tolerant control based on the virtual adherent system. The project results will provide the corresponding theoretical and technical foundation for weak or incipient thruster fault of AUV, manned deep-sea submersible or other underwater equipment.

无人无缆自主式水下机器人（AUV）工作在复杂海洋环境，推进器作为AUV核心部件和负荷最重部件，其故障诊断技术对AUV安全性有重要影响。推进器弱故障特征微弱并且与海流等外部随机干扰特征相差较小，对于推进器弱故障诊断问题，目前尚无成熟的理论和一致认可的解决方法。.本项目针对布朗粒子跃迁频率与势阱切换频率的协调问题，采用信号能量转移方法，得到弱故障特征增强和干扰信号抑制机理；针对故障特征信号的统计方差重构问题，采用小波细节分量转移的方法，得到故障奇异信号再次增强方法；针对不同维度故障信息之间的广义射影同步问题，采用基于多源信息的同步升维方法，得到弱故障特征矩阵与弱故障程度的非线性映射关系；针对推进器辨识结果的进一步确诊问题，采用虚拟伴随系统的方法，得到故障诊断与容错控制的协同运作机理。本项目研究成果将为AUV、载人潜器等水下装备推进器弱故障或早期故障诊断提供相应的理论和技术基础。

无人无缆自主式水下机器人（AUV）工作在复杂海洋环境，作为AUV核心部件部件的推进器是AUV影响安全性的重要因素。推进器弱故障特征微弱并且与海流等干扰特征相差较小，推进器弱故障诊断问题目前尚无成熟的理论和一致认可的解决方法。研究推进器弱故障诊断技术对提高AUV安全性、加快其实用化进程具有重要的研究意义和实用价值。. 本项目针对AUV推进器弱故障诊断问题，主要从推进器弱故障特征增强和干扰信号抑制、弱故障奇异信号再次增强、弱故障程度辨识、弱故障辨识结果确诊四个方面进行研究，取得以下研究成果：. 在推进器弱故障特征提取和干扰信号抑制方面：1）得到外部干扰信号能量向故障信号能量转移的机理和方法。基于小波和二维卷积计算，实现信号在时频域能量集中区域增强；2）得到基于ISOMAP算法的推进器弱故障特征提取方法。解决了AUV不同类别数据融合后映射关系不唯一的问题。.在推进器弱故障奇异信号再次增强方面：提出同态隶属函数方法和低频趋势预测方法，对当前控制信号和速度信号进行前向预测，从预测信号中提取故障特征，实现推进器弱故障奇异信号再次增强并提高弱故障诊断精度。. 在推进器弱故障程度辨识方面：1）提出基于特征值相对变化量的归一化计算方法、基于故障信号类型计算相互关系数的方法，解决了不同故障信号间的差异性问题。提出基于正态分布函数计算不同故障信号间关联度的方法，解决了故障信号关联度处理不当的问题。2）得到基于灰色预测模型的水下机器人推进器弱故障程度预测方法，提高了辨识精度。. 在推进器弱故障辨识结果确诊方面：得到推进器故障重构及自适应容错控制方法。提出基于高阶滑模观测器的推进器故障重构方法，解决了适应容错控制框架下故障估计误差较大的问题，得到了故障诊断与容错控制的协同运作机理以及基于虚拟伴随系统的弱故障在线确诊方法。. 本项目研究成果为AUV等水下装备推进器弱故障诊断提供理论和技术基础。

内容获取失败，请点击重试