Gas-Liquid Flow in an Upward Inclined Large Diameter Pipe Under Elevated Pressures

Abstract

This paper presents a unique gas-liquid experimental dataset acquired at large-diameter laboratory multiphase loop under elevated pressures. The dataset and corresponding model validations are useful to upscale available multiphase flow knowledge into large-diameter-high-pressure conditions commonly encountered in offshore facilities. Intermittent (slug and pseudo-slug) and segregated (stratified and annular) flow patterns were observed in the experiments. For given superficial liquid Froude number (FrSL), all flow pattern transitions scale with superficial gas Froude number (FrSG) within the experimental range, capturing the effects of pressure (gas density). The change in pressure gradient and liquid holdup across the intermittent to segregated transition is more pronounced at low vSL. In segregated flow, the pressure gradient (-dp/dL) increases with pressure and vSL. However, these effects are less noticeable in intermittent flow. In intermittent flow, -dp/dL is generally gravity dominated but may become friction dominated as vSL increases, owing to absence of film reversal. For given vSL, -dp/dL scales with FrSG. The relationship between dimensionless -dp/dL (P*) and Lockhart-Martinelli parameter (X*) scales the effects of pressure and vSL for segregated flow. Liquid holdup was observed to decrease with pressure and increase with vSL. As pressure increases, density difference between phases decreases and interfacial friction increases, thereby reducing slippage and holdup (HL). Two state-of-the-art models exhibit similar bias tendency. In the intermittent region the inaccuracy of -dp/dL and HL predictions increase with vSG, i.e.: deeper into pseudo-slug region. This error is larger at low vSL. For segregated flow, the models tend to underpredict -dp/dL as vSG increases. The magnitude of this error is larger at high vSL. This paper addresses the limitation of large-diameter-high-pressure data in multiphase flow literature. The presented data, scaling approaches, and model validation results are critical for model improvement. For practicing engineers, they can be used as an upscaled benchmark/practical guidance to design multiphase flow pipelines.