Investigation of U-tube effect in drilling operation
MetadataVis full innførsel
The U-tube effect is present when drilling with a low riser return system(LRRS)or a dual gradient system(DGS). For the LRRS the U-tube effect will initiate asthe surface mud pump is shut down. The mud level in the drill string will drop until the hydrostatic pressure in the drill string and annulus are equal. When a DGS is applied, the U-tube effect will occur since the pressure at the inlet of the seafloor pump is maintained near the hydrostatic pressure of the seawater. Since the density of mud is greater than the seawater, the fluid level in the drill string will drop due to the pressure imbalance. The U-tube effect affects tripping operations when doing a connection since air will be injected into the drill string as the fluid level drops when breaking the pipe. The U-tube effect will also be dominant in a well control operation since the driller can not measure the shut in drill pipe pressure(SIDPP) due to the fluid level drop, unless a drill string valve(DSV) is installed. Formation integrity test(FIT)and leak of test(LOT), for the LRRS and the DGS, have to be done in a different way compared to conventional due to the U-tube effect. The widening of the pressure window, pore pressure versus fracture pressure, is one of the advantages with the LRRS and the DGS. The riser annulus consists of drilling fluid and air on top in a static situation when the LRRS is used. This will result in a pressure gradient that starts from the top of the mud column in the riser annulus and not from the rotary kelly bushing(RKB) as in conventional drilling. The widening of the pressure window for a DGS is a result of its dual density system, where the pressure gradient from RKB to the seafloor is equal to the hydrostatic pressure of seawater and the higher mud density below seafloor makes sure the bottom hole pressure(BHP) is greater than the pore pressure. Laboratory experiments for a small scale model of the LRRS show that the U-tube effect will last for 56 seconds when a 5/32" nozzle is used compared to 16 seconds when a 1032" nozzle is part of the system. The reduction of time to reach pressure equilibrium is due to the increased flow area of the 10/32" nozzle. A best fit numerical model is compared with the lab experiments where the damping constant, kT , the area change constant, kB;o and the nozzle constant,knozzle;o, are tuned to match the measured data. The damping constant is smaller for the experiment with the 5/32" nozzle compared to the experiment with the 1032"nozzle, since the friction pressure loss through the nozzle is greater due to the smaller flow area. Hence, the fluid level motion in the drill string for the 5/32"nozzle experiment, does not need to be as damped as the 10/32" nozzle experiment The best fit numerical model is verified by using different initial heights in the drill string and annulus and equal values for kT , kB;o and knozzle;o . For the 8/32" nozzle,all the experiments with initial different heights matched the numerical model.This indicates that the numerical model can be used for different initial heights in the future, assumed that water is used as drilling fluid and a 8/32" nozzle is part of the system. In the pressure calcualtions based on experimental data, the friction pressure drop in the annulus is neglected. To validate the assumption, the Reynolds numbers have been discussed. They show that turbulent flow will be developed for a system without nozzle, so the assumption was invalid for that case. It is in general controversial to neglect the effect of pressure loss in annulus, but by doing so the results are closed up to the observations. The friciton loss in the annulus will be strongly influenced by the size of the nozzle in the drill string. Future work should verify the numerical model for all the four nozzle sizes by using different initial heights. In this thesis, there were only available time for verifying the numerical model for the scale model with a 8/32" nozzle. Mud should also be used as drilling fluid, but the problem with mud is to figure out how to measure the level change in the drill string and annulus, since it is not possible to observe the level change visually due to settlement of mud particles on the pipe wall.