Using Tracers 
in Refinery 
Control 
Flow rate, volume, flow pattern, effi- 
ciency of separation and mixing, and 
leakage—all can be determined easily 
with tracers where other methods fail 
By D. E. HULL 
California Research Corporation 
Richmond, California 
SIGNIFICANT IMPROVEMENTS in product yield and quality, 
and safer operations result from using radiotracers in 
refinery control. The radiotracer applications described 
here are only a fraction of those studied. 
Radioisotopes are valuable in petroleum refining in two 
different manners: as fixed radiation sources, and as tracers. 
Level gages are the most common use of radioactive sources 
in the refinery. Because y-rays can penetrate several feet 
of petroleum stock, plus the container walls, both source 
and detector can be installed outside the container. Many 
types of radioactive level gages (1-3) are available. 
A newer use for radiation sources is in devices for 
analyzing hydrocarbons for their C/H ratio (4) or for 
their S content (6). These are based on the different 
radiation-absorption characteristics of different elements. 
In contrast to sources, radiotracers have not been widely 
exploited as implements for refinery control. 
Tracers Used 
Most frequently the material to be traced in refining is 
a liquid stream. The tracer radioisotope should be pref- 
erably a y-emitter; its half-life must be comparable with 
test duration; and it must be physically and chemically 
compatible with what it traces. Tracers most commonly 
used are triphenylstibine containing Sb!*4 and Co®°- 
naphthenate for oil streams, and Cs!4salts for water 
streams. For solid-material tracing, Zr®5 is used. Its 
refractory properties mean that cracking catalyst can be 
traced. Gas tracers would be very useful in refinery tests, 
but no completely satisfactory isotope is available. 
158 
1—Flow Rate by Integral Count. 
Trocer 
injector 
A millicuries 3 
Vv Barrels 
Second 
F Counts Millicuries _ 
Seond/ Serial Borrel 
Measuring Flow Rates 
In the continuous processing that is the backbone of 
present-day refining methods, flow-rate knowledge and 
control are essential. 
Two-point method. The general method (Fig. 2) for 
measuring liquid or gas flow rates is timing a surge of tracer 
between two points separated by a determinable volume 
(6). This condition is most easily satisfied on a straight 
section of pipe of known dimensions, free from branch con- 
nections. The radiotracer is injected quickly, at a point 
close to the section where it will be timed, so that sharply 
defined peaks in counting rate are obtained as the tracer 
passes the counters. The counters at the two points are 
connected together into a single amplifier so that they both 
record on the same chart. The volume between the two 
points, divided by the time between peaks, gives the flow 
rate directly. 
Integral-count method. A new tracer method (Fig. 1) 
for flow uses but one detector and eliminates the need 
to know the volume between two points (7), It is based 
on the principle that the integral number of y-rays regis- 
tered by a G-M counter on a pipe passing a definite 
quantity of radioisotope is inversely proportional to flow 
velocity. This integrated count is independent of the 
variation in isotope concentration along the stream, as long 
as the flow rate is constant. This principle can be demon- 
strated mathematically; it was first experimentally vali- 
dated in field pipeline tests. A given batch of tracer, 
observed at points so far apart that extensive spreading 
had occurred during transit from one to the other, was 
found to give the same total number of counts while pass- 
ing each point. 
To translate the integral count into an absolute determi- 
nation of flow rate, it is necessary to calibrate the counting 
set-up on the particular type and size of pipe involved. 
This is done by filling a cut section of the pipe with a radio- 
tracer solution at known concentration and measuring the 
counting rate on a counter tube attached to the pipe in the 
same way as in the field measurements. The dimensions 
of this calibration factor, F, are 
= (counts/sec) + (mc/barrel) 
These can be transposed into 
F = (counts/mc) X (barrels/sec) 
With the calibration factor known, a measurement of the 
