基于AUV平台的海管泄漏实时监测关键技术

The ocean provides rich oil and gas resources for human beings. Scientifically monitoring pipeline leakage for effective early warning, contemporaneous alarming and hazard assessment is significantly demanded to ensure the safety of our national ocean environment and the security of the Ocean Economy Demonstration Region in Zhou Shan, Zhejiang. Considering the peculiarity of the layout of submarine oil and gas pipelines and that of the ocean environment around them, we will construct an underwater mobile platform equipped with autonomous underwater vehicles (AUVs) carrying multi-beam sonars and various types of other sensors. First, the key technologies of the AUVs for autonomous cruising and of multi-beam sonars for detection are researched, so that the pipe targets can be effectively identified and positioned. Second, this project will utilize high precision terrain surveying combined with Bayesian estimation methods for sea tube state detection and tracking. Then the approach to real-time detection and classification of the leakage is developed using high resolution water imaging and Convolutional Neural Networks (CNN). Third, this project will rate the leak hazard by fusing the measurement information of the sensors and visualize submarine pipeline environment in 3D. Finally, this project will design and construct an experiment system for pool and sea experimental research to further optimize the proposed theories. In summary, the outcomes of this research will not only solve the international problems in the field of real-time monitoring of oil and gas pipelines, but also provide a crucial technical support to the smart ocean strategy being advanced by the Zhejiang Province and our nation.

海洋提供人类丰富的油气资源，对油气管道泄漏进行科学监测实现有效预警、及时报警和危害评估，切合国家海洋环境安全和浙江舟山海洋经济示范区安全保障的重大需求。项目根据浙江海底油气管道布局及周边海洋环境特点，构建由AUV搭载多波束声呐和各类传感器组成的水下移动平台。基于该平台，首先，开展自主巡航的AUV和多波束声呐探测的关键技术研究，有效辨识定位管道目标；接着，结合高精度测深和贝叶斯估计方法，实现海管状态检测与跟踪；同时通过高分辨水体成像和卷积神经网络，完成泄漏物实时检测与分类；然后，融合AUV搭载的各类传感器的量测信息实现泄漏危害等级评估，同时进行海底管线环境的三维可视化展示；最后，构建实验系统分别进行水池实验和海上试验研究，根据实验结果优化理论方案。项目的研究成果不仅解决了海底油气管道实时监测技术领域的国际难题，而且为国家和浙江省正在推进的智慧海洋战略提供关键技术支撑。

海洋提供人类丰富的油气资源，对油气管道泄漏进行科学监测实现有效预警、及时报警和危害评估，切合国家海洋环境安全和浙江舟山海洋经济示范区安全保障的重大需求。项目围绕近海油气管道的移动式监测存在的水下环境杂、探测对象多、探测反应快、自动程度高、评价标准优等难点，围绕基于AUV的移动检测平台关键技术、基于声学探测的海管状态实时检测技术、海管泄漏物声学识别技术、海管环境可视化与泄漏危害评估，以及实验系统构建与试验研究等五条研究内容开展研究。.项目组针对研究内容一提出了基于快速随机搜索树的多目标点路径规划方法，能够快速有效地实现给定路径协同跟踪控制；针对研究内容二提出了有效的接收阵列校准技术、及用以提高声呐水平分辨能力的高分辨波束形成技术和用以提高声呐航向分辨能力的孔径合成技术，从而获得海管图像和轮廓信息以及海底信息；利用多次检测技术同时检测水中目标和海底目标并通过自适应门限进行区分，利用三维点云信息实现了海管状态的实时检测与特征提取；另外，还提出了利用地形匹配辅助定位与导航以及利用海管特征的AUV位姿定位的高效鲁棒方法。针对研究内容三，研究了基于多次检测的静态特征提取和基于光流法的三维声呐图像序列动态特征提取方法，不仅可以有效检测和识别水下泄漏目标，而且还具有对泄漏规模进行定量的评估能力；针对研究内容四，开发了海底管道环境多信息可视化系统及基于多源异质传感器融合的泄漏危害评估系统；并基于多波束测深声呐和AUV构建了移动监测平台，开展了水下泄漏实时探测与识别实验，有效实现了水下泄漏的实时报警。.项目组在研究过程中注重理论与实践相结合，通过水池、湖上和海上的充分实验检验了研究成果的有效性和先进性，有效提升了我国海油工程中的海管泄漏声学检测技术水平，相关研究成果还可以拓展应用于其它水下特定目标的巡检识别，提升军民领域水下智能化水平，推进智慧海洋战略实施。