X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
217–226
EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE
INSULATION PERFORMANCE OF FLAME-RETARDANT
MULTILAYER INSULATION MATERIALS
EKSPERIMENTALNA [TUDIJA DEJAVNIKOV, KI VPLIVAJO NA
IZOLATIVNE LASTNOSTI VE^PLASTNIH NEGORLJIVIH
IZOLACIJSKIH MATERIALOV
Xiaoyong Ma
1
, Shuping Chen
1*
, Lian Chen
2
, Yujie Wang
2
, Shufeng Jin
1
,
Yang Yu
3
, Maoyuan Mi
1
, Chaofan Shi
1
, Yagang Shi
1
1
School of Petrochemical Engineering, Lanzhou University of Technology, Lanzhou, China
2
Science and Technology on Vacuum Technology and Physics Laboratory, Lanzhou Institute of Physics, Lanzhou, China
3
Institute of Refrigeration and Cryogenic Engineering, Xi’an Jiaotong University, Xi’an, China
Prejem rokopisa – received: 2023-01-05; sprejem za objavo – accepted for publication: 2023-03-07
doi:10.17222/mit.2023.736
Flame-retardant multilayer insulation materials act as effective thermal insulation blankets of cryogenic containers that store
flammable and explosive cryogenic liquids. This study used standard static liquid nitrogen boil-off calorimetry to test the insula-
tion performance of eight groups of flame-retardant multilayer insulation materials with different wrapping parameters. The ef-
fects of four factors, namely the layer density, seaming process, number of reflector layer, and variable-density multilayer insu-
lation arrangement, on the insulation performance were analysed. Three layer densities were considered: 4.47, 3.08, and 2.50
layers/mm. Two types of seaming processes were discussed: the overlapped and fold-over seaming processes. Three numbers of
reflector layers were considered: 60, 70, and 80. Two variable-density multilayer insulation arrangements with similar thick-
nesses were discussed: 10-10-40 and 20-20-20 layers of reflectors allocated for low-, medium- and high-density segments. The
conclusions are as follows: Decreasing the layer density enhances the performance of multilayer insulation; Using the fold-over
seaming process results in less heat flux and lower apparent thermal conductivity; An increase in the number of reflector layers
weakens radiative heat transfer, resulting in better thermal insulation; Furthermore, for a given wrapping thickness, reducing the
number of reflectors appropriately in low- and medium-density segments improves the insulation performance; Optimizing and
controlling the layer density of each density segment are also essential for variable-density multilayer insulation effects. This
study provides supporting theories and reference data for practical engineering applications.
Keywords: flame-retardant multilayer insulation, layer density, seaming process, variable density
Negorljivi ve~plastni izolacijski materiali oz. materiali, ki vsebujejo zaviralce gorenja delujejo kot u~inkovite toplotno
izolativne blazine oz. obloge kriogenih zabojnikih v katerih so shranjene gorljive in/ali eksplozivne kriogene kapljevine.V
{tudiji so avtorji uporabili standardno stati~no kalorimetrijo s teko~im du{ikom za ugotavljanje izolativnih lastnosti osmih
skupin ve~plastnih izolativnih negorljivih materialov z razli~nimi parametri zaviranja gorenja. Analizirali in ugotavljali so
u~inke {tirih parametrov: gostote plasti, procesa {ivanja oz. robljenja, {tevila plasti oz.{tevila odbojnih plasti ter razporeditve
posameznih plasti. Ocenjevali so vpliv treh razli~nih gostot:4,47, 3,08 in 2,50 plasti/mm. Obravnavali oz. izbrali so dva tipa
procesov{ivanja oz. robljenja: s prekrivanjem in z zgibanjem. Ocenjevali so tri razli~na {tevila odbojnih oz. zaviralnih plasti: 60,
70 in 80. Obravnavali so dve razli~ni razporeditvi izolacijskih plasti s podobno gostoto in sicer 10-10-40 ter 20-20-20 odbojnih
plasti v delih z nizko, srednjo in visoko gostoto. Na osnovi rezultatov analiz so ugotavili, da zmanj{evanje gostote plasti
izbolj{uje izolativne lastnosti izbranih ve~plastnih materialov. Uporaba procesa {ivanja oz. robljenja ve~plastnih materialov s
postopkom zgibanja zmanj{uje toplotni tok in zni`uje navidezno toplotno prevodnost materialov. Pove~anje {tevila odbojnih
plastislabi prenos toplote s sevanjem, kar se odra`a v bolj{i toplotni izolaciji. Nadalje so ugotavili, da se pri dani (izbrani)
debelini materiala z zmanj{anjem {tevila odbojnikov primerno izbolj{ajo izolativne lastnosti materialov zdeli z nizko in srednjo
gostoto. Optimiranje in kontrola gostote plasti v vsakem delu je prav tako pomembna za u~inkovito delovanje ve~plastnih
izolacijskih materialov s spreminjajo~o se gostoto plasti. Predstavili so tudi teoreti~ne vidike problemov izdelave ve~plastnih
negorljivih materialov in uporabne reference za prakti~ne in`enirske aplikacije.
Klju~ne besede: upo~asnjevanje ali zaviranje gorenja, ve~plastni izolacijski materiali, gostota plasti, proces {ivanja, spremi-
njanje gostote
1 INTRODUCTION
Fossil fuels are the main energy resource worldwide.
1
However, in recent years, the need for air pollution con-
trol, environmental conservation and geopolitical secu-
rity has shifted the attention to emerging cleaner energy
sources such as hydrogen and natural gas. Effective stor-
age is a key issue for these energy sources, with the liq-
uid storage being preferable. Hydrogen and natural gas
can be liquefied; in this form, they are referred to as liq-
uid hydrogen (LH
2
) and liquefied natural gas (LNG).
Liquefaction has the advantages of high energy density
and storage efficiency, convenient management and
transportation. However, LH
2
and LNG require low-tem-
perature storage as they have low boiling points and la-
tent heats of vaporisation, and they evaporate readily.
Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226 217
UDK 621.315.6:544.452 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 57(3)217(2023)
*Corresponding author's e-mail:
chensp@lut.edu.cn (Shuping Chen)

Therefore, the containers used for storing LH
2
and LNG,
called cryogenic liquid containers, must be thermally in-
sulated to prevent heat loads
2–5
.
The thermal insulation methods commonly employed
for cryogenic liquid containers include bulk insulation,
high-vacuum insulation, evacuated powder insulation,
and multilayer insulation (MLI). MLI, also known as
‘superinsulation’, is used most widely for cryogenic liq-
uid containers owing to its outstanding insulation effect
and typical apparent thermal conductivities of only
10
–5
–10
–6
W/(m·K) in a high-vacuum environment of
10
–3
Pa. In addition, MLI is lightweight and environmen-
tally friendly. As shown in Figure 1(a), cryogenic liquid
containers generally have a double-wall structure con-
sisting of an inner vessel, an outer vessel, and the vac-
uum-insulated annular space between the vessels. MLI
materials, wrapped and mounted on the outer surface of
the inner vessel as an insulation system, play an essential
role in the vacuum-insulated annular space, directly con-
tributing to the performance of cryogenic liquid contain-
ers.
6–8
As illustrated in Figure 1b, MLI materials are
placed between the inner and outer vessels of a cryo-
genic storage tank and consist of reflector layers with
low emissivities (grey vertical lines) and spacer layers
with low thermal conductivities (dashed red lines). These
inhibit thermal radiation (magenta arrows), gas conduc-
tion (blue arrows) and solid conduction (yellow arrows).
There is usually a temperature differential of hun-
dreds of degrees Celsius between the cold mass of a
cryogenic liquid and the surrounding environment,
which results in a strong radiative heat transfer between
the inner and outer vessels. The arrangement of reflec-
tors can effectively reduce thermal radiation. Increasing
the number of reflectors installed reduces thermal radia-
tion to a greater extent. The layer density – the total
number of reflector layers divided by the thickness of the
MLI system – is related to the vacuum-pumping effect
inside the MLI materials: the evacuation quality is re-
flected in the amount of the residual-gas conduction.
Moreover, the layer density is also related to the magni-
tude of the thermal contact resistance and contact area
among the layers of MLI materials. Thus, the layer den-
sity is an important factor to be considered in the design
of an MLI system.
9
This study experimentally evaluates
the influence of the layer density on the thermal insula-
tion performance of MLIs. Johnson et al. reported that
seam designs affect the insulation performance of MLI
systems.
10
Therefore, this study tested the thermal per-
formance of MLI using two seaming processes (Figure
2): the overlapped and the fold-over processes, with the
same number of reflector layers and similar wrapping
tightness. Variable-density multilayer insulation
(VDMLI) is a key factor in enhancing insulation perfor-
mance. In VDMLI, one layer of reflector and N layers of
spacers (N > 1) constitute a composite layer of the
cold-end segment, and a number of composite layers are
configured to reduce the layer density by adjusting the
number of spacers within a composite layer. In the
hot-end segment, a layer of the reflector and a layer of
the spacer are combined into a composite layer, and a
corresponding number of composite layers are arranged.
VDMLI is an advanced and notable technology in the
field of multilayer insulation. It can specifically reduce
solid conduction dominated by the cold-end segment,
weaken the thermal radiation playing the major role in
the hot-end segment, and decrease the total heat transfer.
The initial research conducted by NASA, comparing
VDMLI with uniform-density multilayer insulation
(UDMLI) showed that VDMLI provides superior ther-
mal protection in cryogenic storage applications.
11,12
Hastings et al.
13
observed that VDMLI can attain the
same insulation performance as UDMLI with less mate-
rial. Zheng
14
found that, compared with a 50-reflec-
tor-layer UDMLI, the heat leakage of VDMLI with the
same layer number was 13.53 % lower, and the heat flux
was 17.49 % lower, proving that VDMLI exhibits better
thermal insulation performance.
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
218 Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226
Figure 1: MLI material application and internal thermal-transfer mechanism

The standards governing the construction of vehicu-
lar cryogenic LNG containers and application of MLI
materials, such as GB/T 34510-2017 and ASTM
C740/C740M-13, require that the cryogenic liquid con-
tainers intended for the storage of flammable and explo-
sive energy sources like LH
2
or LNG are protected by
flame-retardant insulation materials placed in the annular
space.
15,16
Therefore, the thermal insulation performance
of flame-retardant MLI materials must be tested and elu-
cidated, and the factors affecting the insulation perfor-
mance must be investigated.
Most theoretical and experimental studies on the in-
sulation properties of MLI materials focussed on aero-
space applications, and the selected research materials
were mostly non-flame-retardant MLI materials with
aluminised Mylar/Dacron net combinations; experimen-
tal studies on flame-retardant MLI materials are rare.
17–19
In this study, the thermal performance of flame-retardant
MLI materials were experimentally investigated, consid-
ering the following aspects: layer density, seaming pro-
cess, number of reflector layers and VDMLI arrange-
ment. The findings and conclusions of this study may
facilitate the design of MLI structures for applications in
cryogenic liquid containers, which are especially suitable
for cryogenic containers for storing flammable cryogenic
liquids.
2 EXPERIMENTAL METHODS AND
MATERIALS PREPARATION
2.1 Experimental method
Static liquid-nitrogen boil-off calorimetry is a stan-
dard method for testing the thermal-insulation perfor-
mance of MLI materials, as shown in Figure 3. Its ex-
perimental platform can be divided into four subsystems:
liquid-nitrogen filling, vacuum pumping, data acquisi-
tion, and calorimeter. The liquid-nitrogen-filling subsys-
tem, responsible for introducing liquid nitrogen into the
LN
2
boil-off calorimeter, comprises two self-pressurised
cryogenic dewars, their corresponding pipes and several
cryogenic valves. The vacuum-pumping unit consists of
rotary-vane and turbo-molecular-vacuum pumps capable
of evacuating the vacuum chamber to 10
–5
Pa. The
data-acquisition subsystem includes a digital vacum-
meter, gas-mass flowmeter, thermometers and auxiliary
devices (such as a temperature-inspecting instrument and
a data-acquisition computer) for monitoring the opera-
tion of the experimental set-up and collecting experimen-
tal data. The technical parameters of the key instruments
of the data-acquisition subsystem are listed in Table 1.
Table 1: Technical parameters of the key instruments of the
data-acquisition subsystem
Instruments
Digital
vacuometer
Gas-mass
flowmeter
Thermometer
Brand EDWARDS ALICAT TPQE
Model nWRG
SCIENTIFIC
M-Series
PT100
Measurement
range
10
–7
to 10
5
Pa
0to300
SCCM
70 to 720 K
In the calorimeter subsystem, the inner barrel of the
LN
2
boil-off calorimeter consists of three cylindrical ves-
sels: the upper-guard, test, and lower-guard cylinders, all
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226 219
Figure 2: Two seaming processes used for wrapping MLI materials
Figure 3: Schematic diagram of the standard static liquid-nitrogen boil-off calorimetry

of which are installed in the vacuum chamber (inside the
outer shell). The flame-retardant MLI materials must be
wrapped around the surface of the inner barrel for test-
ing, as shown in Figure 4a.
In this study, the flame-retardant MLI materials were
pre-treated based on the relevant standards such as GB/T
31480-2015. They were wound evenly on the inner bar-
rel of the calorimeter using the clamshell approach: each
MLI quilt, which contained 5 layers of reflectors and
corresponding spacers, was wrapped individually around
the inner barrel to be insulated.
20,21
The number of wind-
ing layers was recorded, and the wrapping thickness of
the MLI structure was measured. Thermometers were ar-
ranged and fixed on the surfaces of the inner barrel and
outermost reflector to record the cold and warm bound-
ary temperatures of the MLI structure, respectively. Sub-
sequently, the inner barrel was placed into the outer shell
of the calorimeter using a crane to ensure that the sur-
faces of the MLI materials were not damaged during the
insertion process.
After sealing the calorimeter, the vacuum-pumping
unit was used. After the vacuum chamber was evacuated
to less than 0.1 Pa, the guard and test cylinders were
filled with liquid nitrogen. After the liquid-nitrogen fill-
ing, the pressure in the vacuum chamber, also known as
cold vacuum pressure, smoothly decreased to an order of
magnitude of 10
–4
Pa and was always maintained below
1×10
–3
Pa throughout the testing process. After the first
filling operation, the calorimeter remained idle for 30
min while the evacuation system continued to run. After
the inner barrel of the calorimeter and the MLI materials
cooled down completely and the internal system ap-
proached thermal equilibrium, liquid nitrogen was
re-added into the inner barrel to compensate for the
evaporation loss, and then the system was ready for data
acquisition.
In the data-recording stage, liquid nitrogen was filled
into the guard cylinders of the calorimeter approximately
every 10 min to ensure the presence of sufficient liquid
nitrogen in the guard cylinders to fully block the heat
leaking into the test cylinder of the calorimeter from the
upper and lower longitudinal directions, ensuring the ac-
curacy of the test data. Figure 4b illustrates the testing
site of the insulation-performance analyses for the
flame-retardant MLI materials.
The test data collected included the cold and warm
boundary temperatures of the MLI structures and gas-
eous nitrogen flow rates passing through the gas mass
flowmeter after the vaporisation and evaporation of liq-
uid nitrogen. When the variation in the gas flow between
any two time intervals was less than 5 %, the calorimeter
system was considered to be stabilised, achieving ther-
mal balance. The flowmeter values and boundary tem-
peratures of the MLI materials were then recorded every
10 min for the next hour.
2.2 Materials preparation
For this study, the flame-retardant MLI materials,
widely used on the Chinese mainland, were purchased
from a related manufacturer. The properties of the mate-
rials are listed in Table 2.
To investigate the factors affecting the thermal per-
formance, eight experimental groups were tested in this
study, which included UDMLI and VDMLI arrange-
ments. Layout schemes were also designed for the
flame-retardant MLI materials. Table 3 shows the de-
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
220 Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226
Figure 4: Wrapping material and the experimental set-up
Table 2: Material properties
MLI Material type Properties
Reflector Aluminium foil
Compatible with oxygen;
Thickness of 0.0065±6%mm;
Longitudinal tensile strength = 0.27 kN/m;
Bright and continuous, consistent light transmittance, uniformity greater
than ±5 %, no black spots.
Spacer
Cryogenic alkali-free ultrafine
fibreglass paper
Compatible with oxygen;
Thickness = 0.06 mm;
Density of 12 ± 2.0 g/m
2
;
Longitudinal tensile strength = 0.04 kN/m;
Moisture content = 0.5 %;
Combustible matter content =1%;
Thermal conductivity = 0.037 W/(m·K).

tailed layout parameters of the UDMLI arrangements,
and the details of the VDMLI arrangements are pre-
sented in Table 4. The differences between UDMLI and
VDMLI are illustrated in Figure 5. UDMLI refers to the
alternating arrangement of one layer of the reflector and
one layer of the spacer from the cold boundary to the
warm boundary of the insulation structure. The VDMLIs
in this study were divided into three density segments:
low-, medium- and high-density segments, with spacer-
reflector ratios of 3:1, 2:1 and 1:1, respectively. The
measurement experiments were performed using the
standard static liquid-nitrogen boil-off calorimetry.
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226 221
Table 3: Wrapping details of the flame-retardant MLI materials with the UDMLI arrangement
No. of serial
experiments
Layer density
x (layers/mm)
Number of
reflectors
Number of
spacers
Wrapping thickness
(mm)
Types of seaming process
Exp. 1 4.47 60 60 13.4 Overlapped seaming process
Exp. 2 3.08 60 60 19.5 Overlapped seaming process
Exp. 3 2.50 60 60 24.0 Overlapped seaming process
Exp. 4 2.58 60 60 23.3 Fold-over seaming process
Exp. 5 2.98 70 70 23.5 Fold-over seaming process
Exp. 6 2.79 80 80 28.7 Fold-over seaming process
Table 4: Wrapping details of the flame-retardant MLI materials with the VDMLI arrangement
No. of serial experiments Exp. 7 Exp. 8
Low-density segment
Number of reflectors 10 20
Number of spacers 30 60
Thickness of the segment (mm) 5.52 9.89
Layer density x
L
(layers/mm) 1.81 2.02
Medium-density segment
Number of reflectors 10 20
Number of spacers 20 40
Thickness of the segment (mm) 4.57 6.76
Layer density x
M
(layers/mm) 2.19 2.96
High-density segment
Number of reflectors 40 20
Number of spacers 40 20
Thickness of the segment (mm) 11.14 4.84
Layer density x
H
(layers/mm) 3.59 4.13
Total number of layers of reflectors 60 60
defaultTotal number of layers of spacers 90 21.49120
Wrapping thickness of VDMLI (mm) 21.23
Seaming process Overlapped seaming process
Figure 5: Schematic illustration of the differences between UDMLI and VDMLI arrangements

3 MEASUREMENT UNCERTAINTIES
The heat-leakage values of the flame-retardant MLI
structures involved in the experiments were obtained us-
ing Equation (1), and the results for the eight experimen-
tal groups have been collated in Table 5. Heat leakage
can be considered a one-dimensional thermal transfer
from outside to inside, along the radial direction of the
calorimeter in the normal direction of the material sur-
face.
The heat leakage of MLI materials is given by:
22
QV L
T
T
P
P
=
⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ ⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟  g
0
1
1
0
(1)
where Q is the heat leakage through the MLI, W; V is
the boil-off gas flow rate collected by the flowmeter,
m
3
/s; L represents the latent heat of the vaporisation of
liquid nitrogen, J/kg;  g
is the density of gaseous nitro-
gen at 273.15 K, kg/m
3
; P
1
and T
1
are the pressure and
temperature at the outlet of the flowmeter, Pa and K, re-
spectively; P
0
is the standard atmospheric pressure of
1.01325 × 10
5
Pa; T
0
is the thermodynamic temperature
in the standard state, which is 273.15 K.
Researchers and engineers usually use heat flux and
apparent thermal conductivity to characterise the thermal
insulation effects of MLI materials. The heat flux refers
to the heat leakage through the unit area of insulation
materials in the normal direction of the material surface
under a steady-state heat-transfer process. The apparent
thermal conductivity refers to the heat leakage through a
unit of insulation material per time unit under the
steady-state heat-transfer process with known cold and
warm boundary temperatures.
According to one-dimensional steady-state thermal
conduction theory, the heat flux and apparent thermal
conductivity of MLI materials can be expressed as
follows
23
:
q
QD / D
lD D
i
i
=
⋅
−
ln( )
()
0
0
π
(2)
 =
⋅
⋅
QD / D
lT
i
ln( )
0
2πΔ
(3)
where q is the heat flux, W/m
2
; D
0
is the outer diameter
of the wrapped MLI structures, m; D
i
is the inner diame-
ter of the wrapped MLI structures, m; l represents the
length of the test cylinder of the calorimeter inner bar-
rel, m;  represents the apparent thermal conductivity,
W/(m·K); and T is the temperature difference between
the warm and cold boundaries of the wrapped MLI
structures, K.
By comparing the heat fluxes and apparent thermal
conductivities, this study analyses and determines the ef-
fects of layer density, seaming process, number of reflec-
tor layers and VDMLI arrangement on the insulation
performance of flame-retardant MLI materials.
In addition, measurement uncertainties are related to
testing accuracy and are important for the credibility of
experimental results. A summary of the measurement
uncertainties of the LN
2
boil-off calorimeter is presented
in Table 6. Uncertainties Q, q and  can be obtained as
follows:
5,24
u
Q
Q
V
u
Q
V
=
⎛ ⎝ ⎜ ⎞ ⎠ ⎟ ∂
∂
ln
2
2
(4)
u
q
q
Q
u
q
D
u
q
D
q
QD
=
⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ +
⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ +
∂
∂
∂
∂
∂
∂
ln ln ln
2
2
0
2
2
0
i
Dl
u
q
l
u
i
⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ +
⎛ ⎝ ⎜ ⎞ ⎠ ⎟ 2
2
2
2
∂
∂
ln
(
5)
u
Q
u
D/ D
u
Q
i
D/ D
i
  
=
⎛ ⎝ ⎜ ⎞ ⎠⎟+
⎛ ⎝ ⎜ ⎜ ⎞ ⎠ ⎟ ⎟ ∂
∂
∂
∂
ln ln
()
()
2
2
0
2
0
2
2
2
2
2
+
⎛ ⎝ ⎜ ⎞ ⎠⎟+
⎛ ⎝ ⎜ ⎞ ⎠ ⎟ ∂
∂
∂
∂
ln ln 
l
u
T
u
lT
Δ
Δ
(6)
Based on Equations (4)-(6) and Table 6, uncertainties
Q, q and  were estimated to be 1.72 %, 1.74 %, and
1.96 %, respectively.
4 RESULTS AND DISCUSSION
4.1 Layer density
The performance of flame-retardant MLI materials
depends on the design and arrangement of the layer den-
sity. Theoretically, a lower layer density is conducive to
less heat flux and higher insulation efficiency. This is
caused by a reduction in the layer-to-layer contact area,
an increase in the contact thermal resistance, and im-
provement of vacuum conditions. Using Equations
(1)-(3), the heat fluxes and apparent thermal conductivi-
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
222 Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226
Table 5: Heat leakage measured for exps. 1 to 8
Serial experiments Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp. 6 Exp. 7 Exp. 8
Heat leakage Q (W) 0.390 0.311 0.205 0.153 0.150 0.136 0.208 0.323
Table 6: Measurement uncertainties of the LN
2
boil-off calorimeter
Symbols Description Units Uncertainties
V Boil-off gas flow rate m
3
/s ± (0.8 % of reading + 0.2 % of full scale)
Do
D
i
l
Outer/inner diameter
Length of the test cylinder
m ± 0.2 % of reading
 T Temperature difference K ± 0.5 % of reading

ties of Exps. 1, 2 and 3 were obtained, as shown in Fig-
ure 6, reflecting the influence of layer density on the in-
sulation performance.
As mentioned above in Table 3, Exps. 1, 2 and 3
have the same number of reflectors – 60 layers of alu-
minium foil – and the overlapped seaming process is
adopted uniformly at the material junction; they differ
only in the layer density. The layer density in Exp. 1 is
4.47 layers/mm, while those of Exps. 2 and 3 are 3.08
and 2.50 layers/mm, respectively. When the MLI materi-
als are wrapped more loosely, the lower layer densities
result in thicker MLI structures and better thermal insu-
lation effects. Of these groups, Exp. 3 demonstrated the
best thermal insulation performance, with a minimum
heat flux of 1.151 W/m
2
and the smallest apparent ther-
mal conductivity of 1.319 × 10
–4
W/(m·K). A compari-
son of Exps. 1 and 2 indicates that the layer density in
Exp. 2 is 31.10 % lower than that observed in Exp. 1,
while the heat flux is 23.25 % lower. The layer density
and heat flux of Exp. 3 are 18.83 % and 35.77 % lower
than those of Exp. 2, respectively. These results indicate
that a layer density in a range of 2.5–3.0 layers/mm can
effectively diminish the heat flux and improve the ther-
mal insulation performance. The apparent thermal con-
ductivities of the three experimental groups were similar,
and the maximum fluctuation range was approximately
20 %, which is typical. Thus, reducing the layer density
can significantly reduce the heat flux; this observation
may be useful for engineering practice.
4.2 Seaming process
As illustrated in Figure 7, both Exps. 3 and 4 used 60
layers of aluminium foil as reflectors, and their layer
densities were similar, 2.50 layers/mm and 2.58 lay-
ers/mm, respectively. However, the groups differed in the
seaming technique employed during the material wrap-
ping process. Exp. 3 adopted an overlapped seaming pro-
cess, whereas Exp. 4 adopted a fold-over seaming pro-
cess. Exp. 4, using the fold-over seaming process,
demonstrated a better insulation performance, with less
heat flux and a smaller apparent thermal conductivity
than that of Exp. 3. The heat flux of Exp. 4 was
0.863 W/m
2
, which was 25.02 % lower than that of
Exp. 3. The apparent thermal conductivity of Exp. 4 was
9.248 × 10
–5
W/(m·K), which was 29.89 % lower than
that of Exp. 3. Thus, the fold-over seaming used at the
joints of materials is better for enhancing the insulation
performance of flame-retardant MLI structures than the
overlapping process.
4.3 Layer number of reflectors
As observed in Figure 8, the MLI structures in Exps.
4, 5 and 6 exhibited a good thermal insulation perfor-
mance, with heat fluxes of less than 1 W/m
2
. Given that
the layer densities were in the 2.5–3.0 layers/mm range
and the seaming process remained the same, the perfor-
mance of the MLI structures was effectively improved
with an increase in the number of reflector layers. In
Exp. 6, the heat flux of the MLI structure with 80 reflec-
tors was 0.742 W/m
2
, which was 14.02 % and 12.19 %
less than for Exps. 4 and 5, respectively. This is because
an increase in the number of reflectors can further
weaken the radiative heat transfer, thus reducing the total
environmental heat infiltration. In contrast, a larger num-
ber of reflective layers result in thicker MLI structures,
necessitating a more complex installation, a greater
outgassing amount (which reduces the vacuum life of the
cryogenic liquid containers) and a costlier insulation sys-
tem.
25,26
Therefore, the layer number of reflectors used
needs to be matched to the requirements of the actual ap-
plication.
The apparent thermal conductivities of these three ex-
periments were successfully reduced to the order of
10
–5
W/(m·K). In addition, the decrease in the heat flux
between Exps. 6 and 5 was significantly greater than that
between Exps. 5 and 4. The reason for it may be that the
layer density of Exp. 5 was 2.98 layers/mm, which is
slightly higher than those of Exps. 4 and 6, which may
have negatively affected the insulation performance. The
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226 223
Figure 6: Comparison of the heat fluxes and apparent thermal conduc-
tivities of Exps. 1, 2, and 3
Figure 7: Comparison of heat fluxes and apparent thermal conductivi-
ties between Exps. 3 and 4

thermal insulation performance of UDMLI results from
multiple combined factors such as layer density, seaming
process, and the number of reflectors. Therefore, further
reducing the layer density of Exp. 5 may result in a
smaller heat flux.
4.4 VDMLI arrangement
As shown in Figure 9, two types of VDMLI are con-
sidered in this study: Exps. 7 and 8. However, the num-
bers of composite layers (the numbers of reflectors) allo-
cated to the low-, medium- and high-density segments
are different. The layout schemes of the VDMLI in Exps.
7 and 8 are described in detail in Table 4. The wrapping
thicknesses of Exps. 7 and 8 were similar, but the num-
ber of spacers in Exp. 7 was 25 % lower than that in Exp.
8. As demonstrated in Figure 9, the insulation perfor-
mance of Exp. 7 was better than that of Exp. 8. The heat
flux of Exp. 7 was 35.42 % lower than that of Exp. 8, and
the apparent thermal conductivity was 37.98 % lower,
with values of 1.187 W/m
2
and 1.158 × 10
–4
W/(m·K),
respectively. This shows that, under the condition of a
certain wrapping thickness of the VDMLI structure, an
appropriate reduction of the number of reflectors in the
low- and medium-density segments and a decrease in
their thicknesses can not only reduce the material usage
and save the cost but also contribute to improving the in-
sulation performance.
Exps. 2 and 7 inckuded UDMLI and VDMLI with
the same number of 60 reflector layers, and the wrapping
thicknesses were similar, 19.48 mm and 21.23 mm, re-
spectively. The heat flux of Exp. 7 was 33.76 % lower
than that of Exp. 2, and the apparent thermal conductiv-
ity was 29.95 % lower therein. Thus, given the same
number of reflector layers and a similar wrapping thick-
ness, the VDMLI arrangement exhibits a better thermal
insulation performance than the UDMLI arrangement.
Notably, the results of Exps. 3 and 7 were relatively
close, with the same number of reflector layers but dif-
ferent thicknesses. Theoretically, the performance of
VDMLI in Exp. 7 was expected be better than that of
UDMLI in Exp. 3. Analysis showed the layer density to
be the cause of the observed phenomenon. The layer
density of Exp. 3 was 2.50 layers/mm, which is appropri-
ate and suitable. The layer density of each density seg-
ment in Exp. 7 was excessive, and even the layer density
of the high-density segment was greater than 3.5 lay-
ers/mm. This observation also explains why Exp. 8 ex-
hibited an insulation performance inferior to Exps. 2 and
3. When the high-density segment of Exp. 8 was com-
pared with the densities of Exps. 2 and 3, the layer den-
sity of the high-density segment was 4.13 layers/mm,
which was 1.34 times that of Exp. 2 and 1.65 times that
of Exp. 3. These phenomena further indicate that the in-
sulation performance of flame-retardant MLI materials is
the result of a combination of various factors such as the
number of reflectors, layer density and UDMLI or
VDMLI arrangements. For VDMLI, the handling and
control of layer densities of each density segment must
be addressed. The various influential factors must be
comprehensively optimised to enhance the thermal insu-
lation effect.
Table 7 summarises the results of the thermal-insula-
tion-performance tests of the flame-retardant MLI mate-
rials obtained with Exps. 1 to 8, sharing the data and pro-
viding a reference for the thermal insulation design of
cryogenic liquid containers storing flammable and explo-
sive energy sources like LH
2
or LNG.
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
224 Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226
Figure 8: Comparison of the heat fluxes and apparent thermal conduc-
tivities of Exps. 4, 5 and 6
Figure 9: Comparison of the heat fluxes and apparent thermal conduc-
tivities from Exps. 2, 3, 7 and 8
Table 7: Thermal-insulation-performance indexes obtained with Exps.
1to8
Serial Experi-
ments
Heat leakage
Q
Heat flux q
Apparent thermal
conductivity ë
(W) (W/m
2
) (W/(m·K))
Exp. 1 0.390 2.335 1.457 × 10
–4
Exp. 2 0.311 1.792 1.653 × 10
–4
Exp. 3 0.205 1.151 1.319 × 10
–4
Exp. 4 0.153 0.863 9.248 × 10
–5
Exp. 5 0.150 0.845 9.184 × 10
–5
Exp. 6 0.136 0.742 9.839 × 10
–5
Exp. 7 0.208 1.187 1.158 × 10
–4
Exp. 8 0.323 1.838 1.867 × 10
–4

5 CONCLUSIONS
In this study, the thermal insulation properties of
eight groups of flame-retardant MLI materials were
tested, and the effects of four factors – layer density,
seaming process, number of reflector layers and VDMLI
arrangement – on the insulation performance of MLI
structures were analysed. As a result, the following con-
clusions are drawn:
A moderate reduction in the wrapping tightness of
flame-retardant MLI structures, i.e., a decrease in their
layer densities, can effectively reduce the heat leakage
and flux and improve the insulation performance. In this
study, when the layer density was reduced by 44.07 %,
the heat leakage and heat flux were reduced by 47.44 %
and 50.71 %, respectively. In the 2.5–3.0 layers/mm
range, the heat flux decreased significantly with the layer
density. However, in practical engineering applications,
the size of the vacuum-insulated annular space of a cryo-
genic liquid container must also be considered. There-
fore, the layer density of materials should be appropri-
ately decreased within a limited space.
The fold-over seaming process better enhances the
thermal insulation performance of flame-retardant MLI
materials than the overlapped seaming process. The
fold-over seaming process leads to less heat leakage and
heat flux, and better apparent thermal conductivity. As a
result, the choice of the seaming process plays a signifi-
cant role in improving the insulation performance of
flame-retardant MLI structures.
Adding more reflector layers can reduce the radiative
heat transfer and further decrease the heat leakage and
flux. In this study, when the number of reflector layers
was increased by 33.33 %, the heat leakage decreased by
11.11 % and the heat flux was reduced by 14.02 %. In
engineering practice, the number of reflectors can be de-
termined by the type of cryogenic liquid stored. For
cryogenic LH
2
containers, the number of reflectors in the
MLI system ranges from 80 to 120 layers. For cryogenic
LNG containers, this number ranges from 40 to 60 lay-
ers. When the number of reflectors is fixed, the MLI sys-
tem can achieve an outstanding insulation performance
with an appropriate layer density and proper wrapping
technique. However, a higher number of reflectors can
result in increased production cost and increased mate-
rial outgassing volume.
For the 60-layer-reflector VDMLI arrangement, the
number of reflectors in low- and medium-density seg-
ments is suggested to be limited to 10 layers or fewer,
and the layer density should be paid attention to and
carefully controlled in each density area.
The minimum heat flux was achieved in Exp. 6
(0.742 W/m
2
). The minimum apparent thermal conduc-
tivity was obtained in Exp. 5 (9.184×10
–5
W/(m·K)). The
data and wrapping parameters for flame-retardant MLI
materials can help improve the insulation design for
cryogenic LH
2
or LNG containers and provide meaning-
ful reference values for practical engineering applica-
tions.
Acknowledgement
This study was financially supported by the Natural
Science Foundation of the Gansu Province, China
(21JR7RA269).
Nomenclature
D
o
: Outer diameter of MLI (m)
D
i
: Inner diameter of MLI (m)
l: Length of the test cylinder of the calorimeter inner bar-
rel (m)
L: Latent heat of the vaporisation of liquid nitrogen
(J/kg)
P
0
: Standard atmospheric pressure (Pa)
P
1
: Pressure at the outlet of the flowmeter (Pa)
q: Heat flux through MLI (W/m
2
)
Q: Heat leakage through MLI (W)
T
0
: Standard thermodynamic temperature (K)
T
1
: Temperature at the outlet of the flowmeter (K)
V: Boil-off gas flow rate (m
3
/s)
x: Layer density of UDMLI arrangements (layers/mm)
x
L
: Layer density of the low-density segment in VDMLI
(layers/mm)
x
M
: Layer density of the medium-density segment in
VDMLI (layers/mm)
x
H
: Layer density of the high-density segment in VDMLI
(layers/mm)
 T: Temperature difference between warm and cold
boundaries of MLI (K)
 : Apparent thermal conductivity of MLI (W/(m·K))
 g
: Density of gaseous nitrogen at 273.15 K (kg/m³)
6 REFERENCES
1
BP Statistical Review of World Energy 2022, https://www.bp.com/
content/dam/bp/business-sites/en/global/corporate/pdfs/energy-eco-
nomics/statistical-review/bp-stats-review-2022-full-report.pdf
2
S. A. Sherif, N. Zeytinoglu, T. N. Veziroglu, Liquid hydrogen: Poten-
tial, problems, and a proposed research program, Int. J. Hydrogen
Energy, 22 (1997) 7, 683–688, doi:10.1016/s0360-3199(96)00201-7
3
M. Aziz, Liquid hydrogen: A review on liquefaction, storage, trans-
portation, and safety, Energies, 14 (2021) 18, 1–29, doi:10.3390/
en14185917
4
C. Winnefeld, T. Kadyk, B. Bensmann, U. Krewer, R. Hanke-
Rauschenbach, Modelling and designing cryogenic hydrogen tanks
for future aircraft applications, Energies, 11 (2018) 1, doi:10.3390/
en11010105
5
B. Wang, R. Luo, H. Chen, C. Zheng, Y. Gao, H. Wang, A. R.
Hashmi, Q. Zhao, Z. Gan, Characterization and monitoring of vac-
uum pressure of tank containers with multilayer insulation for cryo-
genic clean fuels storage and transportation, Appl. Therm. Eng., 187
(2021), doi:10.1016/j.applthermaleng.2021.116569
6
P. M. Sutheesh, A. Chollackal, Thermal performance of multilayer
insulation: A review, IOP Conf. Series: Materials Science and Engi-
neering, Kerala State 2018, doi:10.1088/1757-899X/396/1/012061
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226 225

7
D. Singh, A. Pandey, M. K. Singh, L. Singh, V. Singh, Heat radiation
reduction in cryostats with multilayer insulation technique, J.
Instrum., 15 (2020), doi:10.1088/1748-0221/15/07/P07032
8
X. Xu, H. Xu, B. Yang, L. Chen, J. Wang, A novel composite insula-
tion system of hollow glass microspheres and multilayer insulation
with self-evaporating vapor cooled shield for liquid hydrogen stor-
age, Energy Technol., 8 (2020) 9, doi:10.1002/ente.202000591
9
J. Demko, J. E. Fesmire, Q. Shu, Cryogenic Heat Management:
Technology and Applications for Science and Industry, 1st ed., CRC
Press, Boca Raton 2022, 69
10
W. L. Johnson, J. E. Fesmire, Cryogenic testing of different seam
concepts for multilayer insulation systems, Advances in Cryogenic
Engineering: Transactions of the Cryogenic Engineering Conference
– CEC, 2010, 905–912, doi:10.1063/1.3422447
11
A. Hedayat, T. M. Brown, L. J. Hastings, J. Martin, Variable density
multilayer insulation for cryogenic storage, 36th AIAA/ASME/
SAE/ASEE Joint Propulsion Conference & Exhibition, Huntsville
2000, doi:10.2514/6.2000-3790
12
J. J. Martin, L. Hastings, Large-scale liquid hydrogen testing of a
variable density multilayer insulation with a foam substrate,
NASA/TM-2001-211089, 2001
13
L. J. Hastings, A. Hedayat, T. M. Brown, Analytical modeling and
test correlation of variable density multilayer insulation for cryo-
genic storage, NASA TM-2004-213175, 2004
14
J. Zheng, L. Chen, C. Cui, J. Guo, W. Zhu, Y. Zhou, J. Wang, Experi-
mental study on composite insulation system of spray on foam insu-
lation and variable density multilayer insulation, Appl. Therm. Eng.,
130 (2018), 161–168, doi:10.1016/j.applthermaleng.2017.11.050
15
GB/T 34510-2017 – Liquefied natural gas cylinders for vehicles,
Standards Press of China, Beijing (in simplified Chinese)
16
ASTM C740/C740M-13 – Standard guide for evacuated reflective
insulation. In: Cryogenic Service, ASTM International, West
Conshohocken, Pennsylvania
17
Y. H. Huang, B. Wang, S. H. Zhou, J. Y. Wu, G. Lei, P. Li, P. J. Sun,
Modeling and experimental study on combination of foam and vari-
able density multilayer insulation for cryogen storage, Energy, 123
(2017), 487–498, doi:10.1016/j.energy.2017.01.147
18
Z. Liu, Y. Li, F. Xie, K. Zhou, Thermal performance of foam/MLI
for cryogenic liquid hydrogen tank during the ascent and on orbit pe-
riod, Appl. Therm. Eng., 98 (2016), 430–439, doi:10.1016/j.appl-
thermaleng.2015.12.084
19
J. Zheng, L. Chen, J. Wang, X. Xi, H. Zhu, Y. Zhou, J. Wang, Ther-
modynamic analysis and comparison of four insulation schemes for
liquid hydrogen storage tank, Energ. Convers. Manage., 186 (2019),
526–534, doi:10.1016/j.enconman.2019.02.073
20
GB/T 31480-2015 – Materials for high vacuum multilayer insulation
of cryogenic vessel, Standards Press of China, Beijing (in simplified
Chinese)
21
S. A. Dye, A. B. Kopelove, G. L. Mills, Wrapped multilayer insula-
tion for cryogenic piping, Advances in Cryogenic Engineering:
Transactions of the Cryogenic Engineering Conference – CEC, 2012,
1293–1298, doi:10.1063/1.4707053
22
X. Shen, Study on an insulation performance testing system for high
vacuum MLI, Master’s Thesis, Zhejiang University, Hangzhou,
China, 2013
23
J. E. Fesmire, Standardization in cryogenic insulation systems testing
and performance data, Phys. Procedia, 67 (2015), 1089–1097,
doi:10.1016/j.phpro.2015.06.205
24
J. H. Kim, T. W. Simon, R. Viskanta, Journal of heat transfer policy
on reporting uncertainties in experimental measurements and results,
J. Heat Transfer, 115 (2018) 1, 5–6, doi:10.1115/1.2910670
25
Y. G. Shi, Research on outgassing characteristics of cryogenics insu-
lation materials under vacuum condition, Master’s Thesis, Lanzhou
University of Technology, Lanzhou, China, 2022, doi:10.27206/
d.cnki.ggsgu.2022.000462
26
C. Zheng, Research on vacuum outgassing characteristics and testing
methods of key materials of an LNG tank container, Master’s Thesis,
China Jiliang University, Hangzhou, China, 2021, doi:10.27819/
d.cnki.gzgjl.2021.000360
X. MA et al.: EXPERIMENTAL STUDY ON THE FACTORS AFFECTING THE INSULATION PERFORMANCE ...
226 Materiali in tehnologije / Materials and technology 57 (2023) 3, 217–226