The effect of two phase liquid-gas flow in 194 ft (59.1m) vertical 5 in. (127 mm) pipe on the outputs of a Fluid Density tool and an impeller type flowmeter tool was observed. Gas rates ranged from 2.6 to 249 MSCF/d. (3.6 × 10(-5) to 3.44 × 10(-3) Kmol/s). Liquid rates were 0 and 500 Bbl/d (3.31 m3/h). A correlation was sought to convert the flowmeter output to total surface rate. A detailed discussion of the obtained data and a suggested method of interpretation is presented here.


Most wells producing from partially depleted reservoirs have free gas production associated with the liquid at the producing depths. Most gas wells, except for those drilled with air and some very high rate producers, will have gas flow through a liquid column over at least part of the producing interval. Gas velocities of several thousand feet per minute are necessary to overcome the slip velocities at low liquid holdups. This was experimentally shown by J. Ros (ref. 1) and Govier et al. (ref. 2, pg, 325) among others.

In the majority of producing wells velocities are much lower. In these cases it becomes necessary to distinguish between gas flow through a static liquid column and further to quantify the gas flux rate and the liquid flux rate.


194 ft (59.1 m) of 5.047 in. (128.2 mm) I.D. was hung from the surface with three feeder lines for the gas and both liquids leading to the bottom of the string. Gas flow was measured through metering valves with .080 in. (2.03 mm) and .25 in. (6.35 mm) maximum openings, liquid flowrates were measured with 3/4 in. (19.05 mm) turbine flowmeters (see Fig. 1). Downhole a stack of three tools were run with centralizers on a 3/16 in. (4.76 mm) single conductor wireline. The tools were a Quartz Pressure Gauge, a Fluid Density tool and a High Sensitivity Flowmeter. The latter uses a low mass, jewel suspended impeller two inches (50.8 mm) long and 1.35 in. (34.29 mm) in diameter and can output signals for both the rate and the direction of rotation. A more detailed equipment description is beyond the scope of this presentation and can be found in ref. 3.


For all flow measurements with zero liquid flowrates the flow string was kept full of liquid until the gas rate had completely stabilized; then the liquid line was shut off, Three different measurements were recorded from 30 ft (9.1 m) to 170 ft (51. m) from the top of the string. Pressure was obtained from a Quartz Pressure transducer in a single run through the interval. Two runs, one up and one down were made with the Fluid Density tool and two to four runs in both directions at line speeds between 8 and 40 ft/min (.04 to .203 m/s) were made with the Flowmeter. In addition, 5 minute stationary readings at 20 ft (6.1 m) intervals were made with both tools. The output from the three tools was recorded in standard API Log format. (Fig. 2)


At the same depths as the stationary readings which were 8 points, 20 ft (6.1 m) apart, the superficial gas velocity (ug) was calculated, using the pressure reading from the Quartz gauge and the gas volume measured at the surface. The superficial liquid velocities in all tests were either 0 or 14.21 ft/min (.072 m/s). This was added to ug to give the superficial total velocity (ut) or also called mixture velocity (um). In Fig. 3, this velocity is plotted for all observed points vs. water holdup. Superficial water velocity (uw) for Fig. 3 points vs. water holdup. Superficial water velocity (uw) for Fig. 3 was zero. Fig. 4 shows the mixture velocity vs. holdup for the tests with air and diesel, again for a superficial diesel velocity (uo) of zero, Comparison of Fig. 3 and 4 shows that below a velocity of 20 ft/min (.102 m/s) the diesel shows a lower holdup than the water, however for velocities between 20 and 700 ft/min (.102 to 3.56 m/s), the diesel holdup is higher than the water holdup for the same gas flow. The liquid holdup discussed here is obtained from the fluid density tool and thus represents only the holdup in the center of the pipe (a sample size of 3-3/4 in. (95.25 mm) long and 3/8 in. (9.525 mm) in diameter). This then is the local holdup in the center and therefore it is influenced by the concentration profile. The phase concentration can be expressed by defining a distribution coefficient Co = / .

The average holdup was approximated from the recorded pressure log and is shown plotted vs. in Fig. 5 and 6 for the depth pressure log and is shown plotted vs. in Fig. 5 and 6 for the depth of 90 ft (27.4 m) from the surface.

This content is only available via PDF.
You can access this article if you purchase or spend a download.