Speaker
Description
Bullheading is the process of pumping gas bubble downwards in a well. Since gas tends to migrate upwards on its own, a certain critical liquid rate is required to be able to move the gas downwards. The bullheading procedure is used for killing production wells and it is also a backup well kill solution if a gas influx is taken during drilling of new wells. It is also used in a special drilling system developed for handling large drilling fluid losses in carbonate formations.
One-dimensional two-phase flow can be described by the Drift-Flux model. Here the gas slip relation is essential to predict how gas moves relative to liquid. This relation depends on two parameters. One parameter is the gas rise velocity which describe how fast a gas bubble moves relative to stagnant liquid. The other parameter, the gas distribution coefficient is more related to the shape of the gas bubble. In literature, one will find that the models were mainly developed for co-current upward flow. Although, there are a few papers addressing how these parameters will change when considering co current and downward two-phase flow.
A medium size experimental arrangement has previously been built at University of Stavanger and experimental bullheading data from the loop was compared against the results achieved from the gas slip relation and the drift flux model. When they considered a Newtonian fluid and slug flow (Taylor bubbles), large differences were seen. It was much harder to push the gas down in practice compared to what the simulation model predicted.
In this paper the experimental loop will be described along with some improvements done before repeating the experiments. It will be shown that still there was a large discrepancy between model predictions and the experimental data. Here one will also emphasize how a simulation model based on the Drift-Flux model is needed to be able to predict the downward gas velocities when using a specific slip relation.
The large discrepancy led to a new investigation of what could be the cause. A deeper literature review considering an old paper from 1962 gave a hint that the gas distribution parameter can change dramatically when transitioning from co current upward flow to counter current and downward flow. Then gas distribution parameter was then calibrated to obtain better fit with the experimental data using both theoretical considerations and a workflow involving model simulations and comparison with the experimental data. The new value for the distribution coefficient differed significantly from other values proposed in literature. The reason for this large discrepancy is most likely relate to non-symmetric behavior of the Taylor bubble in this experimental setup. Literature seems to support that this kind of behavior exist and will have certain implications with respect to gas movement.
