天然气泄漏扩散实验与高斯预测模型优化

Experiments of natural gas leakage diffusion and optimization of the Gaussian prediction model

众所周知,掌握天然气扩散规律是天然气泄漏监测方案设计的重要依据,其中利用高斯模型来预测天然气泄漏扩散规律已成为天然气事故风险评估和安全疏散的重要技术,但目前还没有适用于天然气等轻气泄漏扩散浓度测量的实验装置。为解决高斯预测模型对天然气泄漏扩散预测精度较低的问题,采用天然气扩散浓度监测实验平台,借由阵列布置的激光甲烷浓度探头测量不同风速和不同泄漏速率下,扩散下游空间内不同高度和位置的甲烷浓度,进一步修正了高斯扩散模型中的抬升高度和侧向、垂向扩散系数。研究结果表明:①自主设计的甲烷浓度监测实验平台,可实现下游扩散空间内任意位置的泄漏气云捕捉和分布式浓度测量;②天然气泄漏的气云抬升现象不明显,仅在无风、泄漏速率较小时出现了明显的气云向上抬升。结论认为,通过实验优化后的高斯扩散模型计算精度显著提升,对比查表获取的经验参数,本研究优化后的高斯模型预测扩散浓度偏差由33%降低至16%,显著降低了预测误差,成果对天然气泄漏风险评估与泄漏监测系统的设计具有重要指导意义。

It is well known that natural gas diffusion pattern is vital to the design of natural gas leakage monitoring.Determination of natural gas leakage diffusion using the Gaussian prediction model has become a key process of risk evaluation and safe evacuation in case of natural gas accidents.Nonetheless,there are no experimental facilities available to measure the leakage diffusion concentration of light gases such as natural gas.Considering the low accuracy of the Gaussian prediction model in natural gas leakage diffusion prediction,the natural gas diffusion concentration monitoring experimental platform was employed to measure the methane concentrations at various heights and locations in the downstream diffusion space under different wind speeds and leakage rates by using the array of laser methane concentration probes.Then,the lift height,lateral diffusion,and lateral/vertical diffusion coefficient in the Gaussian diffusion model were further corrected.The results show that the custom-designed methane concentration monitoring experimental platform can effectively capture leakage gas clouds and measure their distributed concentrations at any location within the downstream diffusion space.Gas cloud lift resulted from natural gas leakage is not significant,but only observed in instances of no wind or low leakage rate.It is concluded that the Gaussian diffusion model optimized through experiment yields a significantly enhanced accuracy.Comparison with empirical parameters obtained from gauge readings reveals a great decrease in the deviation of diffusion concentration predicted by the Gaussian model from 33%to 16%.The research results have important guiding significance for the design of natural gas leakage risk evaluation and leakage monitoring system.