The design of crude oil production systems involving the pipelining of unseparated reservoir fluids requires a detailed understanding of multiphase flow phenomena. Slug flow in particular may cause undesirable process upsets if the separation and processing facilities are not designed to handle the resultant gas and liquid flowrate variations.
Slug characteristics, primarily length, holdup and frequency, cannot be adequately predicted by existing empirical correlations. Mechanistic models are more satisfactory if the measured slug frequency is used as an input. However, existing methods for predicting slug frequencies are also unsatisfactory.
BP have collected a large database on slug flow characteristics in both low pressure test rigs and field production flowlines. From this database BP have developed an improved method for the prediction of slug frequency. This method correlates slug frequency against the two parameters which appear from test rig observations to be the most important in the initial formation of slugs from a stratified flow. These parameters are the liquid holdup in the stratified film at the start of the line, and the gas-liquid slip velocity. The results indicate a much improved comparison with the data. The phenomenological approach to the correlation development gives increased confidence of its applicability to conditions outside the range of data from which it was developed.