管廊内燃气管道泄漏口尺寸统计分析与预测

Statistical Analysis and Predictionof Leakage Port Size of Gas Pipelinein Utility Tunnel

为解决综合管廊燃气管道输运安全研究缺乏失效数据支撑的问题,基于管廊内燃气管道架空敷设的特征,收集156组不同类型架空管道的失效数据,统计分析管道泄漏口尺寸的分布。建立BP神经网络预测模型,预测管廊内燃气管道的条缝形泄漏口长度,有效解决管道泄漏口尺寸随机性大、难以预测的问题。研究表明:条缝形泄漏口是架空管道失效的主要形式,条缝形泄漏口长度大部分分布在0~1 000 mm范围内,不同长度条缝形泄漏口的数量占比整体呈现指数函数的分布趋势。随着使用年限增加,小尺寸泄漏口会发育并扩展为大尺寸泄漏口,[15,20) a是长度为(0,30)mm小尺寸泄漏口发育的高峰期;整体而言,长度大于等于150 mm条缝形泄漏口的数量占比随着管径增大而增高;相对承压能力处于[1.2,1.6)时,长度大于等于150 mm的条缝形泄漏口的数量占比高达40%,对应的运行压力范围是比较危险的。构建的基于BP神经网络的综合管廊燃气管道条缝形泄漏口长度预测模型,预测精度较高。

In order to solve the lack of failure data support in the safety research of gas pipeline transportation in utility tunnel,based on the characteristics of overhead laying of gas pipeline in utility tunnel,156 sets of failure data for different types of overhead pipelines were collected,and the distribution of pipeline leakage port size was statistically analyzed.A BP neural network prediction model was established to predict the length of slit-shaped leakage port in gas pipeline within utility tunnel,which effectively solved the problem of high randomness and difficulty in predicting the size of pipeline leakage ports. The results show that the slit-shaped leakage port is the main failure mode in overhead pipelines,the size of the slit-shaped leakage ports is mostly distributed in the range of 0?1 000 mm,and the proportion of the number of slit-shaped leakage ports with different lengths shows an exponential distribution trend as a whole. With the increase of service life,the small-sized leakage port can develop and expand to the large-sized leakage port. The peak period for the development of small-sized leakage port,ranging from 0 to 30 mm in length,is 15 to 20 years. Overall,the proportion of the number of slit-shaped leakage ports with a length greater than or equal to 150 mm increases with the increase of pipe diameter. When the relative bearing pressure capacity is at between 1.2 and 1.6,the proportion of the number of slit-shaped leakage ports with a length greater than or equal to 150 mm accounts for up to 40%,and the corresponding operating pressure range is relatively dangerous. The prediction model for the length of slit-shaped leakage port in gas pipeline within utility tunnel,constructed based on BP neural network,has high prediction accuracy.