天然气管道黑色粉末粒度分布模型评价研究

掌握天然气管道黑色粉末粒度分布（PSD）信息对于解决黑色粉末问题十分关键。如今常用的颗粒PSD模型较多,但缺乏较为成熟的模型评价机制。基于某一实际天然气管道内的黑色粉末数据,引入了S_RMSE、R²、I^AIC等评价指标和混淆矩阵、ROC曲线分别对7种常见PSD模型的拟合优度和预测能力进行了评价,结果显示,对数正态模型兼具描述集中分布和平均分布的能力而在拟合优度方面更具优势;同时,对数正态模型在颗粒全尺寸范围内[0.30 μm,7.25 μm]都有有效的预测效果。因此,该模型是一种综合预测能力最强的分布模型。

Assessment of Particle Size Distribution Models for Black Powders in Natural Gas Pipelines

Understanding the particle size distribution (PSD) of black powders in natural gas pipelines is critical to resolving the black powder issue. There are now many PSD models available; however, there is a lack of established methods for assessing them. In this study, seven common PSD models were assessed for their goodness of fit and prediction capacities, on the basis of black powder data of a real natural gas pipeline, by employing assessment indexes such as S_RMSE, R², and I^AIC as well as a confusion matrix and ROC curve. The results showed that the log-normal model not only is capable of both concentrated and even distribution, but also exhibits better goodness of fit. In addition, the log-normal model is capable of effective prediction in the full range of particle sizes (0.30~7.25 μm). Therefore, it is the PSD model with the most comprehensive prediction capability.