高含硫天然气集输管道腐蚀与泄漏定量风险研究  

Quantitative risk study on corrosion and leakage of high sulfur natural gas gathering pipeline

作　　者：梁国华[1] LIANG Guohua(Guangzhou Institule of Technology,Guangzhou 510075,China)

机构地区：[1]广州工程技术职业学院,广东广州510075

出　　处：《石油化工腐蚀与防护》2022年第5期24-27,共4页Corrosion & Protection In Petrochemical Industry

摘　　要：为了提升高含硫天然气集输管道系统腐蚀与泄漏定量风险分析算法效能,在管道原有的远程实时抄表、压力计量系统的基础上增加激光惯性加速度计探头系统,最终联合应用超限学习机和卷积神经网络构成数据分析系统,赋予天然气集输管道机器人本体知觉,最终形成高精度定量风险评价机制。该系统相比较之前系统,泄漏故障评价平均偏差降低了86.0%,报错周期降低了62.7%,敏感度提升3.6个百分点。腐蚀故障评价平均偏差降低了87.4%,报错周期降低了77.2%,敏感度提升7.8个百分点。In order to improve the efficiency of quantitative risk analysis algorithm for corrosion and leakage of high sulfur natural gas gathering pipeline system, a laser inertial accelerometer probe system was added in the original remote real-time meter reading and pressure measurement system, combined with an overrun learning machine and convolutional neural network to form a data analysis system, endowing the natural gas gathering pipeline robot with ontology awareness, which finally created a high-precision quantitative risk assessment system. Compared with the previous system, the average deviation of leakage fault evaluation was decreased by 86.0%, the error reporting period was decreased by 62.7%, and the sensitivity was increased by 3.6 percentage points;Meanwhile, the average deviation of corrosion failure evaluation was decreased by 87.4%, the error reporting period was decreased by 77.2%, and the sensitivity was increased by 7.8 percentage points.

关 键 词：天然气集输管道 定量风险分析 机器人本体知觉 管道泄漏 管道腐蚀 

分 类 号：TE988.2[石油与天然气工程—石油机械设备]