- T=ixing in Surge Tank 
| Tracer 1 : Astineter 
injector ‘yo : 
___ Theoretical : 
complete mixing 
Cou nts/ Second 
Determining Mixing 
Stock blending is a refining operation that has been fol- 
lowed widely with radiotracers. A radioisotope is in- 
corporated in one component of a blend, and mix activity 
is observed by sampling or by placing a counter in the 
mixture. Large fluctuations diminish until constant 
activity is reached, showing that mixing is complete. 
A study of mixing in a 50,000-barrel surge tank in a 
thermal-cracking-plant feed line illustrates the technique 
(7). Feed comes from two sources and varies in viscosity 
and other properties. Tank contents are mixed by pump- 
ing the liquid through a side loop at 10° barrels/day. 
Surge-tank effectiveness in minimizing feed variations was 
studied (Fig. 7) by injecting Sb!*4-tripheny] into the intake 
line and watching the activity at the loop and the outgoing 
line. * The tracer was introduced into the tank in a volume 
of 50 barrels. During 5.5 hours, 13.1% of the labeled 
batch left the tank, compared with 17.0% expected. How- 
Leak Testing 
Diversion of any liquid stream from one channel to 
another can be detected readily with radiotracers. Under 
the proper conditions such leaks can be measured quantita- 
tively. Leakage between cross-streams in a heat exchanger 
(Fig. 9) offers an excellent example of this technique (7). 
Tracer is introduced in a short surge into the heating- 
stream inlet. G-M counters are attached to the exit pipes 
of both the heating medium and process streams. Appear- 
ance of a tracer wave at the counter on the process stream 
indicates a leak. Leak size is measured by the number of 
counts in this tracer peak compared to the other peak. 
Leaks of a few tenths of 1% are detectable. 
Many of the processes described in this 
paper are patented or patent applications have 
been filed onthem. This article is based ona 
ing of the American Association for the Ad- 
vancement of Science, December, 1964. 
1. M. A. Elliott, U. S. Patent 2,506,585 (1950) 
Det 2. D. G. C. Hare, U. S. Patent 2,323,128 (1943) 
paper presented at the Berkeley, Calif. meet- 3. D. P. Thornton, Pet. Proc. 8, 941 (1950) 
4. R. B. Jacobs, L. G. Lewis, Oil and Gas J. 52, 
No. 21, 128 (1953) 
Mixing in Crude Stills 
i Theoretical — 
i ft n complete mixing 
ia 8 i ~-—— Theoretical 
i I no mixing | 
i e5 * Observed 
3 
Oo4 
os 
(¥3 ; 
_c 
3 j 
o2 
ever, fluctuations of 0-0.22 % of labeled batch in the effluent 
showed that mixing was considerably less than ideal. 
Tar in crude stills. In another refinery application, 
segregation of tar bottoms from different crudes fed to a 
distillation battery of four stills in series was studied (7). 
When switching to a new crude, Co®-naphthenate was 
added to the feed. The tar line leaving the battery was 
monitored for several days with a G-M counter. Activity 
began to appear in the outgoing stream only two hours 
after injection; whereas, if the stocks had followed a piston- 
displacement pattern through the stills, it should have been 
delayed for 21 hours. The tracer spread out into a broad, 
diffuse wave, with a peak at 14 hours and a tail persisting 
into the third day. The curve expected if complete mixing 
occurred in each still (Fig. 8) closely approximated the 
observed data, showing that the mixing was not much 
short of ideal. Thus, it was not possible to effect the 
desired sharp break between tars from the two crudes. 
BIBLIOGRAPHY 5. H. K. Hughes, J. W. Wilezewski, Anal. Chem. 
26, 1889 (1954) 
. G. H. Metcalf, U. 8S. Patent 2,631,242 (1953) 
. California Research Corporation (unpublished) 
D. E. Hull, R. R. Bowles, Ozil and Gas J. 81, 
No. 46, 295 (1953) 
9. G. Gester, et. cl., Pet. Proc. 8, 550 (1953) 
ORD 
161 
