When drilling with aerated drilling fluids, bottomhole pressure is a strong function of gas-liquid flow patterns. The existing methods of predicting flow patterns are mostly based on the extrapolation of results from pipe flow to flow in annuli. Due to non-linear relationships between flow rate, pipe size and pressure drop, this practice may not be appropriate to apply in drilling operations. Therefore, in order to verify the applicability of the existing practice, the present study focuses on the flow of aerated drilling fluids through an inclined annulus, typical for drilling operations in directional wells.
Extensive experiments were performed in a unique fieldscale low-pressure flow loop (8 in × 4.5 in annular geometry, 90 ft long) in inclined positions (15°, 45° from vertical) with (100 rpm) and without drillpipe rotation. The liquids used were water and aqueous polymer solution (CMC+XCD+water), at flow rates in the range of 100–360 gpm and air in the range of 8–85 scfm (for 15°) and 250–1150 scfm (for 45°). Gas/liquid ratios were chosen so as to simulate similar gas/liquid ratios under downhole conditions, correcting for the fact that the actual pressure in the flow loop was much lower than the bottomhole conditions. Measurements of pressure drop and average liquid holdup over the entire annular section were carried out. To our knowledge, no such data has been published before.
The two-phase flow patterns were identified by visual observations. Bubbly flow and the slug flow are the two flow patterns observed over the ranges of the chosen test matrix. The presence of slug flow does not justify many of the existing simulation practices, which assume homogeneous gas-liquid flow.
Also, the flow pattern boundaries proved to be shifted as compared to pipe flow. The transition between bubbly flow and slug flow (for 15° inclination) was observed at a void fraction of 0.32 as compared to that reported for pipe flow (0.25). On the contrary, for air-aqueous polymer fluid flow, it was the same as for air-water pipe flow (0.25). For flow with drillpipe rotation, churn flow was observed instead of slug flow due to the churning of slugs by the rotating drillpipe.
While there was no significant effect of drillpipe rotation on the pressure drop for air-water flow, the pressure drop decreased in case of air-aqueous polymer fluid flow with drillpipe rotation. A higher pressure drop was observed in case of air-aqueous polymer fluid flow as compared to air-water flow.
An existing unified pipe flow model1 was modified based on the experimental results and was also evaluated against experimental and field data. The comparison shows that the pressure drop prediction by the modified model for inclined wells is better than some other existing models, though it still under-predicts the experimental pressure drop measurements.