【作者】 于会民；

【导师】 邱天爽；

【作者基本信息】 大连理工大学 ， 通信与信息系统， 2011， 硕士

【摘要】 随着石油管道运输的不断发展,石油管道安全预警问题成为当前石油管道运输的一项重要内容。输油管道由于建造工艺、自然灾害以及人为破坏等原因所造成的泄漏及破损,极大地影响了石油管道的正常运输,给人们的生产生活带来了诸多不便。石油管道安全预警系统是解决该问题的一个行之有效的途径。该系统的技术核心是如何及时有效地对威胁管道安全的事件信号进行检测和定位。本文针对这一难点问题,结合课题组的研究内容,抓住预警系统的数据采集、分类、处理三个环节,利用时延估计的方法对管道安全事件进行检测定位,取得了较好的效果。具体工作内容为：(1)根据光纤管道安全预警系统采集的数据特点,提取出光纤传感器检测信号中事件信号的特征,并提出了同步性参数概念作为识别事件信号的重要特征,该特征对于确定原始信号中是否包含事件信号以及确定事件信号发生的时刻具有重要意义；(2)研究了适用于该系统的时延定位算法及其改进算法(WGCC算法),提高时延估计准确性及精度；(3)根据统计学原理,制订了时延估计结果的置信度方案,能够实时给出每个异常事件时延值的可信度；(4)通过时延结果概率密度融合技术克服单次估计不稳定的缺点,得到的定位精度基本控制在与准确时延值偏差0.3个点之内,大幅提高系统定位的可靠性。更多还原

【Abstract】 With the continuous development of the oil pipeline transportation, the pipeline safety problems are becoming the important part of transportation task. The oil pipeline leakage and sabotage due to low manufacturing process technology, natural disasters or human destruction greatly affect the normal transport of pipeline, which cause great inconvenience to the normal production and life. The pipeline safety alarm system is one of the methods to solve the problems. The core technology of the system is how to detect and locate the harmful events to pipeline in time. The paper related to a practical project aims to solve the difficulties. The alarm system including three parts (data detection, classification and processing) uses the time delay algorithms to detect and locate the abnormal events. The main works include the following issues:(1) After analyzing the features of system data, a new method is proposed to detect event signal using extracted signal characters, such as the synchronism parameter. The detection method is helpful to discover and pick up the event data;(2) The WGCC algorithm is used to estimate the time delay and position more accurately precisely;(3) According to the statistics theory, the confidence level scheme is used to judge the TDE results;(4) Because one single TDE result is unstable, the probability density fusion technology is used which significantly improve the reliability of position system. The position precision is generally controlled within 0.3 sampling interval.更多还原