埋地输氢管道泄漏扩散规律研究

Research on the leakage and diffusion law of buried hydrogen pipeline

为探究埋地输氢管道的氢气泄漏扩散规律,建立以埋地输氢管道为对象的流体域模型,讨论工作压力、泄漏孔径、土壤类型对氢气泄漏扩散的影响。结果表明,当管道发生泄漏后,气体扩散速度和泄漏量与管道工作压力、泄漏孔径呈正比关系,同时在不同的监测位置,氢气浓度具有不同的变化规律;氢气在低黏度的土壤中泄漏时有着较大的扩散速度,管道在砂土中的泄漏量是黏土中的40倍;因此当管道处于低黏度土壤、高压力的工况时具有较高的危险系数。为对不同工况下地面氢气浓度达到氢气爆炸下限的时间进行预测,以数值模拟结果为训练样本建立GABP神经网络模型,结果显示该模型平均误差仅为3.3%,性能远优于传统BP模型,能合理地对该问题进行预测。

To investigate the hydrogen leakage and diffusion law of buried hydrogen pipelines,a fluid domain model was established with buried hydrogen pipelines as the object,and the effects of working pressure,leakage aperture,and soil type on hydrogen leakage and diffusion were discussed.The results show that when a pipeline leaks,the gas diffusion rate and leakage amount are directly proportional to the pipeline working pressure and leakage aperture.At different monitoring positions,the hydrogen concentration has different variation patterns.Hydrogen gas has a greater diffusion rate when leaking in low viscosity soil,and the leakage rate of pipelines in sandy soil is 40 times that in clay.Therefore,when the pipeline is in low viscosity soil and high pressure conditions,it has a higher risk factor.In order to predict the time when the ground hydrogen concentration reaches the lower limit of hydrogen explosion under different working conditions,a GA-BP neural network model was established with numerical simulation results as training samples.The results showed that the average error of the model was only 3.3%,which was much better than traditional BP models and could reasonably predict this problem.