Passive acoustic technology to detect, locate, and characterize undersea hydrocarbon leaks

The U.S. outer continental shelf is a major source of energy for the United States. The rapid growth of oil and gas production in the Gulf of Mexico increases the risk of underwater oil spills at greater water depths and drilling wells. These hydrocarbons leakages can be caused by either natural events, such as seeping from fissures in the ocean seabed, or by anthropogenic accidents, such as leaking from broken wellheads and pipelines. In order to improve safety and reduce the environmental risks of offshore oil and gas operations, the Bureau of Safety and Environmental Enforcement (BSEE) recommended the use of real-time monitoring. An early warning system for detecting, locating, and characterizing hydrocarbon leakages is essential for preventing the next oil spill as well as for seafloor hydrocarbon seepage detection. Existing monitoring techniques have significant limitations and cannot achieve real-time monitoring. This project launches an effort to develop a functional real-time monitoring system that uses passive acoustic technologies to detect, locate, and characterize undersea hydrocarbon leakages over large areas in a cost-effective manner. In an oil spill event, the leaked hydrocarbon is injected into seawater with huge amounts of discharge at high speeds. With mixed natural gases and oils, this hydrocarbon leakage creates underwater sound through two major mechanisms: shearing and turbulence by a streaming jet of oil droplets and gas bubbles, and bubble oscillation and collapse. These acoustic emissions can be recorded by hydrophones in the water column at far distances. They will be characterized and differentiated from other underwater noises through their unique frequency spectrum, evolution and transportation processes and leaking positions, and further, be utilized to detect and position the leakage locations. With the objective of leakage detection and localization, our approach consisted of recording and modeling the acoustic signals induced by the oil spill and implementing advanced signal processing and triangulation localization techniques with a hydrophone network. Tasks of this project were: 1. Conduct a laboratory study to simulate hydrocarbon leakages and their induced sound under controlled conditions, and to establish the correlation between frequency spectra and leakage properties, such as oil-jet intensities and speeds, bubble radii and distributions, and crack sizes. 2. Implement and develop acoustic bubble modeling for estimating features and strength of the oil leakage. 3. Develop a set of advanced signal processing and triangulation algorithms for leakage detection and localization. The experimental data have been collected in a water tank in the building of the National Center for Physical Acoustics, the University of Mississippi from 2018-2020, including hydrophone recorded underwater sounds generated by oil leakage bubbles under different testing conditions, such as pressures, flow rates, jet velocities, and crack sizes, and movies of oil leakages. Two types of oil leakages (a few bubbles and constant flow bubbles) were tested to simulate oil seepages either from seafloors or from oil well and pipeline breaches. Two types of gases were investigated (nitrogen and methane). These data were analyzed for acoustic bubble modeling, oil leakage characterization, and localization. This dataset contains data for oil leakage source localization. Two localization algorithms were developed: TDOA-based and SpectraRatio-based algorithms. The folders of the dataset are described as follows: • the folders of “Signals” contain raw underwater sounds data used for localization • the folders of “Results” contain the results of true and predicted oil leakage source positions More details of this dataset can be found in the corresponding ReadMe files in each folder.美国外大陆架是美国的核心能源来源之一。墨西哥湾油气产量的快速攀升，使得深水钻井作业面临的水下溢油风险显著增加。此类烃类泄漏既可能由自然事件引发，例如海床裂隙的自然渗漏；也可能源于人为事故，例如井口头与管线破损导致的泄漏。为提升海上油气作业的安全性、降低环境风险，安全与环境执法局（Bureau of Safety and Environmental Enforcement, BSEE）建议采用实时监测技术。一套可实现烃类泄漏检测、定位与特征表征的早期预警系统，对于防范后续溢油事故以及海底烃类渗漏探测均具有关键意义。现有监测技术存在显著局限，无法达成实时监测目标。本项目致力于开发一套实用的实时监测系统，通过被动声学技术，以高性价比的方式实现大范围海底烃类泄漏的检测、定位与特征表征。 在溢油事件中，泄漏的烃类会以高速、大排量注入海水。混合了天然气与原油的烃类泄漏主要通过两种机制产生水下声响：一是油滴与气泡射流引发的剪切与湍流效应，二是气泡的振荡与溃灭过程。这些声学辐射可被水柱中远端的水听器（hydrophone）记录。通过其独特的频谱特征、演化与传播规律以及泄漏位置，可对这些声学信号进行特征提取并与其他水下噪声区分，进而用于泄漏点的检测与定位。 本研究以泄漏检测与定位为目标，通过水听器（hydrophone）阵列记录并建模溢油诱发的声学信号，并结合先进的信号处理与三角定位技术实现泄漏源定位。 本项目的研究任务包括： 1. 开展实验室研究，在可控条件下模拟烃类泄漏及其诱发的声学信号，建立频谱特征与泄漏特性（如油射流强度与流速、气泡半径与分布、裂隙尺寸等）之间的关联关系； 2. 开发声学气泡模型，用于估算油气泄漏的特征与强度； 3. 研发一套用于泄漏检测与定位的先进信号处理与三角定位算法。 实验数据采集于2018至2020年间，采集地点为密西西比大学国家物理声学中心（National Center for Physical Acoustics, University of Mississippi）的水槽实验室，数据涵盖不同测试条件（如压力、流速、射流速度与裂隙尺寸）下，水听器记录的油泄漏气泡产生的水下声学信号，以及油泄漏过程的影像。研究测试了两类油泄漏场景：少量气泡型与恒流气泡型，以分别模拟海床渗漏以及油井与管线破损泄漏；同时分别采用氮气与甲烷两种气体开展实验。上述数据可用于声学气泡建模、油泄漏特征表征与定位研究。 本数据集用于油气泄漏源定位研究，共开发了两类定位算法：基于时差到达（Time Difference of Arrival, TDOA）的算法与基于频谱比的算法。数据集的文件夹结构说明如下： • "Signals"文件夹：存储用于定位的原始水下声学信号数据 • "Results"文件夹：存储真实与预测的油泄漏源位置结果 各文件夹内的详细信息可查阅对应文件夹中的ReadMe文件。