In this paper we present a detailed analysis of large scale experimental data from the SINTEF Multiphase Laboratory on high-rate low liquid loading flows. The experimental work [1] was funded by Equinor as part of the Tanzania gas field development project [2] [3] [4], and SINTEF was granted access to use the data for improving the accuracy of the pressure drop predictions in LedaFlow. The experimental results showed that a key element for predicting high-rate low liquid loading flows accurately is to account for the droplets that deposit on the walls in the gas zone, creating a wall film. This wall film can have a profound effect on the hydraulic roughness experienced by the gas, and subsequently the frictional pressure drop. Furthermore, the data showed that this effect was particularly important for high liquid viscosities and in three-phase flows, and simulations showed that LedaFlow had a clear tendency to under-predict the pressure drop in such scenarios. To improve this situation, we used the data to derive a model for predicting this complex phenomenon. This paper summarizes the main parts of the data analysis and the development of the wall film model. We show that by introducing this new model into LedaFlow, we were able to significantly improve the agreement with the measurements.
Low liquid loading generally refers to flow conditions where the superficial liquid velocity is small compared to the superficial gas velocity. This is a typical scenario for wet gas lines, where the reservoir produces mostly gas, but where changes in the pressure and temperature along the pipe causes condensation of water and hydrocarbons, so that the liquid rate increases with the distance from the well.
