THERMAL PERFORMANCE OF CRYOGENIC PIPING MULTILAYER

INSULATION IN ACTUAL FIELD INSTALLATIONS

J. Fesmire 1, S. Augustynowicz 2
NASA Kennedy Space Center, YA-F2-T, Kennedy Space Center, Florida 32899, USA
2Dynacs Inc., DNX-3, Kennedy Space Center, Florida 32899, USA

ABSTRACT

A standardized way of comparing the thermal performance of different pipelines in

different sizes is needed. Vendor data for vacuum-insulated piping are typically given in
heat leak rate per unit length (W/m) for a specific diameter pipeline. An overall k-value for
actual field installations (koafi) is therefore proposed as a more generalized measure for
thermal performance comparison and design calculation. The koafi provides a direct
correspondence to the k-values reported for insulation materials and illustrates the large
difference between ideal multilayer insulation (MLI) and actual MLI performance. In this
experimental research study, a section of insulated piping was tested under cryogenic
vacuum conditions, including simulated spacers and bending. Several different insulation
systems were tested using a 1-meter-long cylindrical cryostat test apparatus. The simulated
spacers tests showed significant degradation in the thermal performance of a given
insulation system. An 18-meter-long pipeline test apparatus is now in operation at the
Cryogenics Test Laboratory, NASA Kennedy Space Center, for conducting liquid nitrogen
thermal performance tests.

INTRODUCTION

Ambient heat transfer into a cryogenic pipeline comes through several paths including
valves, connectors, instrumentation, and insulation. A common type of thermal insulation
system is MLI. MLI systems come in many varieties and must be tailored to the specific
application. As reviewed by Augustynowicz and Fesmire (2000), the performance of MLI
is known to be sensitive to localized compression effects and trapped residual gases
produced by the combined mechanical influences of bending and spacers. Bending-type
mechanical effects come from four sources: bending, as in handling and installation;
thermal contraction and expansion; line pressure reaction forces; and the weight of the line
(sagging). Spacers are employed in the design of vacuum-jacketed lines to keep the inner
line concentric within the outer line during manufacturing and to counteract these
mechanical effects during operation. Spacers are made from low-thermal-conductivity
materials to minimize heat conduction. In this experimental research study, a section of
insulated piping was tested under cryogenic vacuum conditions including simulated
bending and spacers.
 1. THERMAL INSULATION SYSTEM OVERALL PERFORMANCE

In reality no one really needs insulation. What is important is to save money on the energy
bill or to be able to effectively control a system. Thermal insulation systems therefore
provide energy conservation and allow system control for cryogenic process systems.

The overall efficiency is governed by four basic factors: 1) thermal conductivity, 2) vacuum
level, 3) density or weight, and 4) cost of labor and materials. The vacuum level is the
major cost driver for systems that must operate under the high vacuum condition. A
material's thermal conductivity is described in terms of milliwatt per meter-kelvin (mW/m-
K) for a small difference in boundary temperatures. An insulation material's performance
under a large temperature difference is also measured in terms of mW/m-K but is referred
to as an apparent thermal conductivity or k-value. Finally, the performance of a total
thermal insulation system as it is actually deployed in use is defined as the overall k-value
for actual field installation or koati (Fesmire et al., 2001). The total system includes the
inner piping, the insulation material layers, the outer piping, and other items such as spacers
and getters. Geometry, spacer design, manufacturing factors, vacuum maintenance,
outgassing, insulation materials, layer densities, and other factors all affect the overall
thermal performance of a double-walled piping system.

The performance of MLI systems varies widely. Augustynowicz et al. (2000) assert that
there are three levels of thermal performance of MLI: ideal, laboratory, and industrial.
Ideal performance using advanced materials and sophisticated technique can approach 0.01
mW/m-K while the laboratory performance using state-of-the-art methods will be around
0.1 mW/m-K. Industrial performance, including the effects of spacers and bending, is then
in the range of 1 mW/m-K as evidenced by vendor data and practical experience. This
study provides some quantitative data to assess this wide range of performance.

2. TEST RESULTS AND DISCUSSION

The experimental apparatus and methods of the liquid nitrogen cryostat have been
previously described by Fesmire et al. (2001) and Fesmire and Augustynowicz (2000).

All MLI test articles are comprised of the standard aluminum foil and fiberglass paper
arrangement. Test article C108 is 40 layers at a density of 1.8 layers/ram while test article
C123 is 60 layers at a density of 2.4 layers/mm. Test article C124 is actually the same as
C123 but with the addition of five circumferential rings. These 6.4-mm plastic tings were
installed at a spacing interval of 102 mm to simulate the compression effect of spacers in a
double-wall cryogenic pipeline construction. The compressed local density was measured
to be 8.4 layers/ram (or 71 percent compression). The results of the simulated spacers test
are given in Figure 1. The spacer simulation shows a significant increase in the rate of heat
transfer for the high-vacuum tests. For C123 in comparison to C124 the k-value increased
from 0.09 to 0.15 mW/m-K (a 67 percent increase in heat transfer).

Vendor data for vacuum-insulated piping are typically given in heat leak rate per unit length
(W/m) and are useful then only for a specific diameter pipeline. The koafi is therefore given
as a more generalized measure for thermal performance comparison and design calculation.
For example, from a previous study of a 200-mm by 254-mm double-walled flexible pipe
 i0o

/" /

i
/
Ol

_ ---0_CI08, 40 _ MILl

---O--C123, _0 _ ]d_l

---_..--C124,60_ M_.I + @*ccrmO

o.ol I I t
0.01 OA i i0 100 I000 I0000 I00000 I000000

ColdVacuumP_sJure(m,dlli=_)

Figure 1. Variation of apparent thermal conductivity with CVP, spacer simulation results.

1 t i ,

--E}-- Koafi=0 01 -- &--- 0.05 X-'- 0 1

0.22 + 05 _ 1

O 2 -OK-- 5 --O-- Koafi=lO

I........ r'L
E IO(X) .... " ,, n
, [

,, !
a . ----....._

o01
io 1.1 t 2 ] 3 1.4 i5 16 17 1 8 19 2o

Double Wall Piping Diameter Ratio (Do/Ol)

Figure 2. Variation of heat leak rate with diameter ratio (Do/Di) for koan from 0.01 to 10
mW/m-K and boundary temperatures of 300K and 77K.
with 60 layersMLI, thek-value(athighvacuumwith boundarytemperatures
of 295K and
105K) wasreportedto beabout0.22mW/m-K. Nowconsidera 75-ram-by 125-ram-size
line operatingat boundarytemperaturesof 300K and77 K. The heatleakrate per unit
lengthfor thisparticularcasecanthenbedetermined
by equation(1):
Q = 2rdG,_ AT - 2.743ko_,_ = 0.60 W / m (1)

Figure 2 provides a convenient design tool for estimating heat loads (W/m) for different
lines sizes and different koafi. The experimental laboratory data can be compared with
manufacturers' typical data for a 60-mm by ll0-mm size line: 2.30 W/m (flexible) and
0.75 W/m (rigid). Converting these typical heat leak values into their thermal conductivity
equivalents, we obtain l%a_ of 0.99 mW/m-K (flexible) and 0.32 mW/m-K (rigid). The koan
also provides a direct correspondence to the k-values reported for insulation materials and
illustrates the large difference between ideal MLI performance and the performance of
actual systems with MLI.

CONCLUSION

In this experimental research study, a section of insulated piping was tested under cryogenic
vacuum conditions, including simulated spacers and bending. An overall k-value for actual
field installation (koan) has been defined for direct comparison of different types and
different sizes of double-walled insulated piping. The simulated spacers tests showed, at
high vacuum conditions, significant degradation in the thermal performance of a given
insulation system. The k-values for 60 layers MLI with and without the simulated spacer
rings were measured to be 0.15 mW/m-K and 0.09 mW/m-K, respectively. The apparent
thermal conductivity value for the ideal MLI under the same high vacuum condition is 0.05
mW/m-K. Considering a 75- by 125-mm-size line, for example, at the same boundary
temperatures of 300 K and 77 K, the heat leak rate per unit length is calculated to be 0.41
W/m. This figure for rigid piping can then be compared with vendor data of 0.75 W/m,
which corresponds to a koatq of 0.32 mW/m-K. An 18-m-long pipeline test apparatus is now
in operation at the Cryogenics Test Laboratory for conducting liquid nitrogen thermal
performance tests. Performance data for a variety of commercial and prototype pipelines
are currently being obtained.

REFERENCES

1. Augustynowicz, S. and Fesmire, J., 2000, Cryogenic insulation system for soft vacuum, Advances in
Cryogenic Engineering, VoL 45, Kluwer Academic/Plenum Publishers, New York, p. 1691-1698.
2. Augustynowicz, S., Fesmire, J., and Wikstrom, J., 2000, Cryogenic insulation systems, 20th International
Congress of Refrigeration, Sydney, no. 2000-1147, International Institute of Refrigeration, Paris.
3. Fesmire, J. and Augustynowicz, S., 2000, Insulation testing using cryostat apparatus with sleeve,
Advances in Cryogenic Engineering, Vol. 45, Kluwer Academic/Plenum Publishers, New York, p. 1683-
1690.
4. Fesmire, J., Augustynowicz, S., and Demko, J., 2001, Overall thermal performance of flexible piping
under simulated bending conditions, Cryogenic Engineering Conference, Madison, July (accepted for
publication).
5. Fesmire, J., Augustynowicz, S. and Demko, J., 2001, Thermal insulation performance of flexible piping
for use in HTS power cables, Cryogenic Engineering Conference, Madison, July (accepted for
publication).