基于Markov特征的油气管道泄漏检测与定位方法

发布时间：2018-10-12 07:30

【摘要】：针对传统的基于压力信号的管道泄漏检测方法误报率和漏报率偏高,同时定位误差较大的缺点,设计了一种基于Markov特征的管道泄漏检测与定位方法。首先,将管道压力数据构造为Markov链的形式,并提取其动态特征;然后,将所提取的特征应用于Neyman-Pearson异常检测方法之中,检测全部压力数据样本的状态,并对检测到的异常样本进行同源信号匹配,修正检测结果;最后,将相似性定位方法与连续小波定位方法结合,确定管道首末两端响应压力变化的时间差,并根据管道长度和压力波传输速度等信息,对泄漏源定位。所提方法能应用于小泄漏和缓慢泄漏的检测与定位,易于实现,误报率与漏报率显著降低,定位精度提高。通过对历史数据的分析,验证了所提方法的可行性和有效性。
[Abstract]:Aiming at the disadvantages of the traditional pipeline leakage detection method based on pressure signal, the false alarm rate and false alarm rate are on the high side, and the location error is large, a pipeline leakage detection and location method based on Markov feature is designed. Firstly, the pipeline pressure data is constructed as a Markov chain and its dynamic features are extracted. Then, the extracted features are applied to the Neyman-Pearson anomaly detection method to detect the status of all the pressure data samples. The homologous signals of detected abnormal samples are matched to correct the detection results. Finally, the time difference between the two ends of pipeline response to pressure changes is determined by combining the similarity localization method with the continuous wavelet localization method. According to the information of pipeline length and pressure wave transmission velocity, the leakage source is located. The proposed method can be used to detect and locate small leakage and slow leakage, which is easy to realize, and the false alarm rate and false alarm rate are significantly reduced, and the positioning accuracy is improved. Through the analysis of historical data, the feasibility and validity of the proposed method are verified.
【作者单位】：东北大学信息科学与工程学院;
【基金】：国家自然科学基金(61473069,61627809,61374124) 中央高校基本科研业务费专项资金(N140405002)项目资助
【分类号】：TE973.6

【相似文献】