Horizontal slug flow is the most common flow regime in oil and gas pipelines. Slug flow causes severe internal corrosion problems due to its high turbulence. There is a need to study the physics and chemistry of the influences of slug flow on flow-accelerated corrosion and corrosion models.
This dissertation first describes a systematically experimental investigation of mass transfer and wall shear stress in horizontal slug flow in two-phase water-gas and three-phase water-oil-gas mixtures using the Electrochemical Limiting Current technique. Water and an oil of viscosity of 2 cP were used for the liquid phase and nitrogen was used for the gas phase. Full pipe flow, plug flow, and slug flow in a 10 cm I.D. 15 m long Plexiglas TMpipe was studied.
The effects of gas bubbles, the presence of an oil phase and droplets, and flow regime on average and instantaneous mass transfer coefficients in horizontal slug flow was studied. It is found that the cavitation type bubble collapsing exists in slug flow and causes an increase of up to several hundreds times the instantaneous mass transfer coefficient. The presence of the oil phase can further enhance the cavitation in slug flow. The analysis of the wall shear stress and the fracture mechanics indicates that the instantaneous wall shear stress produced by the cavitation can scrape away the corrosion film from the wall. The measured mass transfer coefficients were used to develop a mass transfer correlation highlighting the bubble collapsing effect in horizontal slug flow.
A CO 2corrosion mechanistic model is developed for the prediction of corrosion rates in carbon steel pipelines experiencing horizontal slug flow. The model considers chemistry, thermodynamics, enhanced mass transfer, electrochemical reaction kinetics, and corrosion films. The model can be used to examine the effects of the slug flow characteristic parameters, temperature, and pressure on corrosion rate. The determined influence of superficial gas velocity and Froude number on corrosion rate is found. A comparison of a laboratory corrosion database and other models proposed for slug flow shows that this new CO 2corrosion mechanistic model improves the prediction of corrosion rates significantly.