Jökull - 01.12.1966, Blaðsíða 33

diagram in which G’/X was plotted against
LXý/G’, the quantities having the definitions
given in the nomenclature. Baker recognized
more sub-divisions in the flow patterns than
did Kozlov, adding regions for “wave”, “strati-
fied” and “plug” types. It should be emphasiz-
ed, however, that for both diagrams the lines
dividing adjacent flow regions are ficticious in
the sense that there is no sudden change of
regime from one type into the other as the
nominal boundary is crossed, only a gradual
transition.
The Ivozlov diagram is shown in Fig. 2 and
the Baker diagram in Fig. 3. On each diagram
state point loci are shown for the following
data.
(A) Data from Iceland (Hveragerdi, Well G-7,
Run 8) Mass flux 356,000 lb/hr., Exit
pressure 40 psia, Base Temperature 421° F,
Exit mass flow dryness fraction 0.174.
(B) Data frorn New Zealand (J. H. Srnith
(1958); Run 1) Mass flux 347,000 lb/hr.,
Exit pressure 335 psia; Base temperature
508° F, Exit mass flow dryness fraction
0.089.
It will be observed that there is broad agree-
ment between the two diagrams. In each case
and for both wells the flow pattern changes
from one of slug flow to annular dispersed
or fully dispersed (mist) flow. Since the slip is
unknown for both cases, A and B, it has been
necessary to base the Kozlov diagram ordinates
on the assumption of homogeneous flow, and
the placing of the loci will be to some extent
in error, particularlv towards the duct exit.
(b) Methods of Evaluating Discharge
(i) Method due to R.James (1962)
James separated the phases at the well head
and subsequently remixed them in controlled
proportions before discharging the mixture to
atmosphere. He obtained values of the mass
flow G over a range of stagnation enthalpies
230—1200 Btu/lb for critical pressures measur-
ed at the pipe exit of 14—64 psia. Pipes 3",
6" and 8" in diameter were employed. James
was able, by observation of the free jet after
discharge to affirm the existence of mist flow
or of annular dispersed flow. Fle found he was
able to correlate his results to an error of less
JÖKULL 187