ABSTRACT

Existing pipe-filling models have primarily considered steady friction alone, tending to underestimate pressure attenuation. Various popular unsteady friction models are considered here to simulate pressures within a rapidly filling vertical pipe. Numerical predictions are compared to each other and to experimental results. The models considering unsteady friction better reproduce measured pressure oscillations. The greatest errors in predicted pressures neglecting unsteady friction occur after the first period, but numerical errors are markedly greater for cases with smaller initial air volumes and higher inlet pressures. Significantly, the largest errors in the prediction of maximum air pressure when ignoring unsteady friction effects occur near the critical value of initial air volume (here around 2–3% of pipe volume). Compared to accurate but computationally demanding convolution models, the Trikha-Vardy–Brown simplified model is shown to be indistinguishable in terms of accuracy; however, the improved Brunone model is even more computationally efficient and retains excellent numerical accuracy.

摘要

现有的管道填充模型主要只考虑稳定摩擦，倾向于低估压力衰减。这里考虑了各种流行的非定常摩擦模型来模拟快速填充垂直管道内的压力。数值预测相互比较，并与实验结果进行比较。考虑非定常摩擦的模型更好地再现了测得的压力振荡。忽略非定常摩擦的预测压力的最大误差发生在第一阶段之后，但对于初始空气量较小和入口压力较高的情况，数值误差明显更大。值得注意的是，在忽略非定常摩擦效应时，最大气压预测的最大误差发生在初始风量临界值附近（此处约为管道体积的 2-3%）。与准确但计算要求高的卷积模型相比，Trikha-Vardy-Brown 简化模型在准确性方面表现出无法区分；然而，改进的 Brunone 模型计算效率更高，并保留了出色的数值精度。