This paper compares the measured wellbore pressure losses from a variety of gas-condensate and gas-water wells with the results from the following steady-state multiphase correlations:
Aziz, Govier and Fogarasi(1)
Hagedorn and Brown (2)
Beggs and Brill (3)
Beggs and Brill revised (4)
OLGAS (5)
OLGAS 2000 (2 phase from Scandpower) (6)
Gregory - AGF (7)
This paper describes the Gregory - AGF(7) model, which is a modification of the Aziz, Govier and Fogarasi (1) method. The differences between using the Duns and Ros (8) model and the Gray(9) correlation for annular-mist flow are outlined. A revision to the Gray (9) correlation is proposed to remedy an apparent typographical error in the publication of that method. The results from using both the original and revised version of the Gray (9) method (as part of the Gregory - AGF (7) model) are compared.
Modelling wellbore pressure losses accurately is especially important for gas field deliverability forecasting. The total amount that the well can produce over its life depends to a large extent on the ability of a well to lift its liquids. It is important to know when compression or smaller tubing is required to help the wells lift their liquids when planning gas field development. Thus correctly predicting multiphase wellbore pressure losses is crucial. This paper compares measured wellbore pressure losses in gas-condensate and gas-water wells with the predictions of several multiphase wellbore pressure loss correlations, including:
Aziz, Govier and Fogarasi(1)
Hagedorn and Brown (2)
Beggs and Brill (3)
Beggs and Brill revised (4)
OLGAS (5)
OLGAS 2000 (2 phase from Scandpower) (6)
Gregory - AGF (7)
The Aziz, Govier, and Fogarasi (1) method determines the flow regime (i.e. annular-mist, slug, froth, bubble) using the correlation of Govier and Aziz (10) and then the corresponding model is used to calculate liquid holdup and frictional pressure loss.
The Hagedorn and Brown(2) method was developed for multiphase wellbore pressure loss calculations using data from a 457 m deep producing test well with nominal tubing sizes of 25 mm, 31.75 mm and 38.1 mm. The liquid holdup was not directly measured; instead, it was inferred from the measured pressure losses. The original Hagedorn and Brown (2) method is independent of flow regime. The revised version used in this study determines if bubble flow exists according to the Griffith and Wallis criteria. If bubble flow exists, the Griffith and Wallis equations to compute liquid holdup and friction loss are used, otherwise the original Hagedorn and Brown procedure is followed. The DeGance and Atherton (11) equations are used instead of the graphs included in the original paper by Hagedorn and Brown(2). The holdup revision suggested by Abdul Majeed et al(12) is incorporated in this revised version.
