A Case Study—Leaktesting a Sodium-Cooled Plant 
A KAPL-operated sodium plant (5) with a primary- 
coolant volume of several hundred cubic feet has shown 
unusually high freedom from leaks. To provide the plant 
with a primary coolant system as leaktight as possible, an 
all-metal, all-welded system is used. 
Previous experience with sodium showed that where 
high system-tightness standards were set, the occurrence 
of leaks during operation was reduced to near the vanish- 
ing point. Since some seemingly insignificant leaks are 
actually relatively large perforations that are partially 
masked by occlusions and will enlarge under operating 
conditions, any detectable leak was assumed to represent 
a flaw that should be repaired. Therefore, with com- 
ponents, and where possible with entire systems, mass- 
spectrometer-quality tightness was required. 
Leak Specifications 
Since the assembled system might not be accessible for 
high-quality mass-spectrometer testing, an over-all com- 
ponent, subassembly, and field weld-testing program was 
substituted. The program assured that no individual 
leak would exceed the lower limit detectable by the mass 
spectrometer leaktester under field conditions. This 
assured that the system would remain leaktight during 
operation and placed an approximate ceiling on over-all 
system leakage. 
The program specified that: (a) all small components 
tested to 0.005 mefh (ft3/hr at 1-micron Hg) and large 
components. to values consistent with their volumes; (b) 
piping components was necessary to 
obtain a leakproof system. Butt welds 
uwe/hr in a 107-ft® building (1). Al- 
though the activity of fission gases varies 
all small subassemblies tested to 0.005 mefh and large 
subassemblies to values consistent with their volumes 
before inclusion in the system; and (c) all field welds 
tested to 0.005 mcfh during assembly. Since there were 
about 100 subassemblies and field welds, the final system 
would be tight to about 1 mcfh over-all leakage. 
Inspection Procedure 
Rigorous handling, inspection, and testing procedures 
of all components and subassemblies were instituted: 
1. Component manufacturers supplied an outline of 
their cleaning procedures, leaktesting schedules, and 
packaging procedures. Leaktesting was conducted at 
several steps during manufacture since a leak in an in- 
accessible part of a component is difficult to discover and 
repair after assembly. Cleanliness, essential in the low- 
leak-rate ranges, must be maintained for repeat and sub- 
assembly tests; thus, packaging for shipment is important. 
2. Inspection on receipt of components included re- 
cleaning as necessary and retesting-for leaks. 
3. The subassemblies were inspected and leak checked 
just prior to inclusion in the system since many of 
the subassemblies and components were fabricated in 
the vendor’s shops and because fit-up work at the con- 
struction site involved some handling and cutting of 
subassemblies. 
4. Finally, the field welds were checked by the mass- 
spectrometer boot technique. 
cially designed for helium leaktesting 
attains a sensitivity of 1 part helium 
or other closures satisfactory for weld- 
ing fabrication will greatly reduce leak- 
testing and repair work after assembly. 
Leaktesting 
The degree of tightness required for 
a nuclear system can be appreciated 
from the table and Fig. 1. As an ex- 
ample, the maximum leak rate for 
fission gases that have a maximum per- 
missible level of 2X 10-§ ye/em* 
(based on Xe!*) would be 2 X 104 
£ 
= 
& 
in 
m 
3 
2 
3 
2 
$ 
a 
a 
Leok Role (cm3/day) 
FIG. 2. Radiation exposure 3 ft from 
point-source leak vs. leak rate 
116 
with reactor power and previous operat- 
ing history, an allowable leakage rate of 
10-§ em?/see is conceivable. Such 
low leak rates require specialized 
techniques for rapid measurement. 
Excessive leakage can produce a 
direct radiation hazard during main- 
tenance. Figure 2 shows the dose 
rate that might be expected from leaks 
in a sodium-cooled plant. 
Applications of the various leaktest- 
ing methods vary with the type of heat- 
transfer fluid and the materials of con- 
struction used. The _ leak-detection 
methods that have been used for pres- 
surized-water and liquid-metal systems 
are summarized on p. 34. 
Sensitivity 
The graph (p. 55) does not indicate 
ultimate sensitivity that can be at- 
tained by modification and the expendi- 
ture of greater amounts of time. For 
example, the dye or chemical-detection 
tests could be improved by using ab- 
sorbent strips, as in the radioisotope 
test, but man-hours would be increased. 
Mass spectrometry is capable of the 
greatest sensitivity (4). The unitespe- 
in 200,000 parts air, thus being capable 
of detecting a leak of about 2 1074 
mefh (micron ft?/hr—a leak rate of 1 
mcfh is such that when applied to 1 ft? 
under high vacuum, the pressure rises 
Max. Permissible Levels in Air* 
Radioisotope MPL (uc/cm?) 
H3 2X 10-5 
Na?4 2 X 10-6 
K* 2 X 10-5 
Cr®! 8 X 1075 
Mn*é 4 X 1076 
XKel36 2 X 1078 
Po?!0 7X 107! 
U (natural) 1.7 *% 1051 
Pu23? 2 1052 
Beta-gamma emit- 
ter (isotope un- 
known) 1X 105° 
Alpha emitter (iso- 
tope unknown) OO 10st? 
* Reproduced from Radiological Hand- 
book, U. S. Dept. Health, Education and 
Welfare (1954). 
