See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223946473

Evaluation of EPS application to embankment of Qinghai–Tibetan

railway

Article  in  Cold Regions Science and Technology · March 2005

DOI: 10.1016/j.coldregions.2004.11.001

CITATIONS READS
37 26

4 authors, including:

Zhi Wen Ma Wei

Chinese Academy of Sciences Chinese Academy of Sciences
65 PUBLICATIONS   1,361 CITATIONS    167 PUBLICATIONS   3,951 CITATIONS   

SEE PROFILE SEE PROFILE

Jilin Qi
Chinese Academy of Sciences
5 PUBLICATIONS   201 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Permafrost indices computing View project

Study on the instability process of slopes in permafrost regions by direct shear test of freezing-thawing interface View project

All content following this page was uploaded by Zhi Wen on 29 October 2021.

The user has requested enhancement of the downloaded file.

 Cold Regions Science and Technology 41 (2005) 235 – 247
www.elsevier.com/locate/coldregions

Evaluation of EPS application to embankment of

Qinghai–Tibetan railway
Wen Zhi, Sheng Yu*, Ma Wei, Qi Jilin
State Key Laboratory of Frozen Soil Engineering, CAREERI, CAS, Lanzhou, Gansu 730000, China
Received 9 March 2004; accepted 9 November 2004

Abstract

In many cases, protecting permafrost from further thaw due to human activities might be the first choice for embankment
design in permafrost regions. The observation data of an embankment with insulation at the Beiluhe site are analyzed. Two-
dimensional finite element analyses are conducted, in which phase change is taken into consideration to simulate the thermal
regime of the Qinghai–Tibetan railway with expanded polystyrene (EPS). Based on the predicted maximum thaw depth in the
following 50 years, the best position for insulation is presented and the relationship between the thickness of insulation and the
height of embankment is analyzed. Also, the applicable range of insulation in embankment engineering of the Qinghai–Tibetan
railway in terms of mean annual air temperature (MAAT) is suggested and the influence of the geothermal field of permafrost
on the applicable range of the insulation is discussed.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Permafrost; Insulation; Qinghai–Tibetan railway; Numerical simulation

1. Introduction thaw settlement in the Qinghai–Tibetan plateau (Wu

and Tong, 1995; Yu et al., 2002; Liu et al., 2002). As
The construction of embankments on permafrost the thickness of permafrost along the Qinghai–
may induce substantial disturbance on the heat and Tibetan railway varies from several to more than
mass transfer balance between the ground surface and 100 m (Li, 1983), the pre-thaw construction method
atmosphere, which results in more heat absorption in is hardly acceptable to most railway sections and the
the embankment, and the temperature of permafrost permafrost protection seems to be the only choice.
underneath increases even as the permafrost thaws, One of the ways to protect permafrost is the thermal
causing serious problems for embankments due to insulation method. The thermal insulation material
was used in roadbed engineering on permafrost in
1950s in Norway. From 1970s, EPS was used as
* Corresponding author. Tel.: +86 931 4967271. insulation in embankment engineering in America,
E-mail address: [email protected] (S. Yu). Japan and Canada, etc. (Gandahl, 1978; Johnson,
0165-232X/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.coldregions.2004.11.001
236 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247

1983; Olson, 1984). In China, a field test was employ the method of numerical simulation. Two-
conducted at Fenghuoshan experimental station in dimensional finite element analysis of a temperature
Tibet to examine the efficiency of this counter- field with phase change is conducted to study the
measure in 1970 (Zhang and Yao, 1994). In 1992, validity and applicable range of insulation (EPS),
another experimental section of thermal insulation considering the thickness and position of the insu-
was built in Qinghai–Tibetan highway (Sheng et al., lation, the height of embankment, MAAT and geo-
2002). The results of the tests showed that in both thermal condition.
cases the insulation functioned effectively. However,
further study was not made due to reconstruction of
the highway. Qinghai–Tibetan railway will cross 2. General situation of testing embankment
about 550 km of permafrost region and the insulation
method was selected as one of the countermeasures 2.1. Site conditions
for permafrost protection. To study the validity and
applicable range of insulation in Qinghai–Tibetan The Beiluhe testing embankment is located on the
railway, a testing section with a length of 375 m was Qinghai–Tibetan railway (DK1139+618–DK1139+
built in 2001. 975), which was constructed in 2001. The study site
The artificial permafrost table and the stability of is situated between Kekexili and Fenghuoshan, where
roadbed are affected greatly by the thickness and vegetation grows well and vegetation fraction is 10–
position of the insulation as well as geothermal 50%. It is an arid climate region in the Qinghai–
conditions of the permafrost. Besides, as a way to Tibetan plateau, where the freezing period is from
protect permafrost, the thermal insulation treatment September to April the next year which accounts for
has limitations, for example, while embankment with 7–8 months. The average mean annual air temperature
insulation maintains or raises permafrost table, the at the site is
3.8 8C with extremes of about 23.2 and
temperature of permafrost below the embankment
37.3 8C in short durations. Weather record shows
rises slowly and there is heat accumulation under the that the mean annual precipitation and mean annual
insulation. Therefore, not all permafrost regions can evaporation are about 290.9 and 1316.9 mm, respec-
use the thermal insulation method to keep the stability tively. Mean annual humidity is 57% and mean annual
of embankment. During the service life of 50 years, wind speed is 4.1 m/s. The maximum snow thickness
embankment with insulation may not maintain the is 14 cm.
stability of embankment in some regions and embank- Engineering geological investigations at the site
ments without insulation may produce adequate indicate that the top layer of a depth of about 1 m is a
stability of embankments in other regions depending sandy clay with gravels; the next 1.5 m is a yellow
on climate and permafrost conditions. sub-clay. Permafrost table is about 2.5 m. There is an
The observation data of embankments with ice-rich soil with a thickness of 0.5–1 m below the
insulation for the Beiluhe testing section on the permafrost table. The ice content by volume is about
Qinghai–Tibetan railway are analyzed in this paper. 50–80%. Under the ice-rich soil layer is mainly a
As the available observed data for this experimental dark-red, fully weathered mudstone with grey, heavily
embankment covers only 2 years, it is necessary to weathered sandstone.

Table 1
The location of testing sections and type of thermal insulation treatment
Insulation type Location Length (m) Thickness of insulation (m) Observation section Position in embankment
EPS-1 DK1139+618–700 82 0.10 DK1139+670 0.5 m above ground surface
EPS-2 DK1139+700–750 50 0.08 DK1139+740 0.8 m below the embankment surface
PU-1 DK1139+750–800 50 0.04 DK1139+780 0.8 m below the embankment surface
EPS-3 DK1139+800–850 50 0.08 DK1139+820 0.5 m above ground surface
PU-2 DK1139+850–925 75 0.04 DK1139+900 0.5 m above ground surface
Without insulation DK1139+925–975 50 DK1139+940
 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247 237

2.2. General design The temperature probes are made of thermal

resistors and their accuracy calibrated in the labo-
According to the type of insulation material (EPS ratory is 0.05 8C. When they are used in the field,
and polyurethane (PU)), the thickness and position of the values of resistance are taken with a Fluke
the insulation in embankment varied from 0.5 m multimeter and the temperatures are calculated
above ground surface to 0.8 m below the embankment according to the equations calibrated in the labo-
surface; five testing sections were constructed. A ratory. The observations started in October 2001 and
section without insulation was built at the site for the readings were taken three times a month. Up to
comparison. The height of testing embankment varies now, the data in the recent 2 years have been
about from 3.0 m to 4.0 m. The location of each obtained completely.
section and type of thermal insulation treatment are
presented in Table 1.
3. Analysis of geothermal field of testing
2.3. Observation section and instrumentation embankment

The plane view of observation sections and the 3.1. Analysis of geothermal field under embankment
locations of boreholes to observe ground temperature
are shown in Fig. 1. At each observation section, four Ground temperatures at different depths below
boreholes were made, one at the centerline of the embankment in DK1139+820 section after embank-
embankment, two at each shoulder of embankment ment construction are illustrated in Fig. 2. Due to the
and one on the natural ground for ground temperature embankment construction, the temperature at depths
detecting. In each borehole, temperature transducers of 5–10 m increased significantly. Permafrost temper-
were set up along the depth at intervals of 0.5 m, ature at 5 m increased greatly, while the permafrost
starting from the ground surface. The depths of the temperature at 15 m kept constant. Ground temper-
boreholes, at the centerline and shoulders of the ature at 2.5 m decreased slowly and went to a constant
embankment are 15, 18 and 10–15 m, respectively. state, while ground temperature at 3.0 m did not
Moreover, single temperature probes were placed on change.
top and bottom of the insulation at the center and the After construction of the embankment with
shoulders. insulation in October 2001, permafrost table under

Fig. 1. Plane view of observation sections and the location of boreholes to observe ground temperature.
238 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247

Fig. 2. Ground temperature at different depths below embankment in DK1139+820 section.

the embankment usually rose by 1–1.5 m and the embankment worked efficiently and achieved
ground temperatures in the initial active layer expected function. The temperature of gravel above
decreased. Original active layer became a zero the insulation responded well to the change of air
curtain about 2 months after construction of the temperature and had a big range of temperature, while
embankment. the ground temperature below insulation had a delay
in response, at the same time it had a small range of
3.2. Temperature difference between top and below of temperature. Owing to the thermal insulation, the
the insulation annual range of ground temperature below the
insulation decreased greatly and the temperature
Obvious temperature differences between the top difference between top and bottom of the insulation
and bottom of the insulation in DK1139+820 section reached 6–8 8C. Besides, due to the insulation, heat
(Fig. 3) indicated that the thermal isolative material in exchange was reduced and there was less heat

Fig. 3. Temperature difference between top and bottom of the insulation in DK1139+820 section.
 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247 239

absorption in the permafrost. Therefore, the perma- 1.0, 1.5, 2.0, 2.5, 3.0, 3.5 and 4.0 m. The subgrade
frost below the thermal isolative material was width is 7.4 m and the side slope is 1:1.5. The width
protected from thawing and the stability of embank- of insulation is changed with the embankment during
ment was maintained. the computation since insulation is the full width of
Observation data showed that the permafrost table the embankment. The embankment is regarded as
under the embankment went up while shallow ground unlimitedly long, with a constant cross-section; thus,
temperatures also rose after embankment construc- the heat conduction process can be modeled in two
tion, which implies that the rise of the permafrost table dimensions.
was accompanied by a temperature rise in the under-
neath layer. 4.2. Soil properties in computational domain

The density and thermal properties of soils in the

4. Mathematical formulations computational domain were based on the data from
field tests. Soil properties in the computational
As available data from the experimental embank- domain are shown in Table 2.
ment covers only 2 years, it is necessary to carry out
numerical simulations in order to predict the long- 4.3. Latent heat
term efficiency of insulation.
Latent heat of soil can be calculated by the
4.1. Computational domain following equation:
Q ¼ Lqd ðW
Wu Þ ð1Þ
Computational domain is shown in Fig. 4. Part A is
gravel with fine inclusion. Part B is sub-clay. Part C is where Q and L are latent heats of soil and water,
an ice-rich soil. Part D is mudstone. Seven embank- respectively. W and W u are the water content and
ment heights were selected during the computation: unfrozen water content, respectively.

Fig. 4. Computational model.

240 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247

Table 2
Soil properties in computational domain
Soil type
20 8C
10 8C
5 8C
2 8C
1 8C
0.5 8C 0 8C 20 8C
Gravel with fine inclusion, q (kg m
3 )=2100
k (J m
1 8C h
1) 5400 5400 5400 5400 5400 5400 5040 5040
C (J kg
1 8C
1) 830 840 850 860 870 900 1060 1060

Sub-clay, q (kg m
3 )=1920

k (J m
1 8C h
1) 6480 6480 6480 6480 6480 6480 5400 5400
C (J kg
1 8C
1) 910 1040 1090 1110 1140 1200 1270 1270

Ice-rich soil, q (kg m
3 )=1500

k (J m
1 8C h
1) 7920 7920 7920 7920 7920 7920 3600 3600
C (J kg
1 8C
1) 1180 1290 1310 1350 1370 1420 1910 1910

Mudstone, q (kg m
3 )=2200

k (J m
1 8C h
1) 9000 9000 9000 9000 9000 9000 7200 7200
C (J kg
1 8C
1) 950 1080 1100 1140 1200 1250 1350 1350

The unfrozen water content can be obtained by the layer, the 2-D heat conduction with phase change in
following equations (Xu and Wang, 2001): roadbed and base can be described as follows:




b BT B BT B BT
Wu ¼ aT ð2Þ qc ¼ k þ k ð5Þ
Bt Bx Bx By Bx

where q=soil density, c=specific heat capacity of soil,

lnW0
lnWu t=time, k=coefficient of heat conductivity and T=
b¼ ð3Þ
lnT
lnTf temperature.

BTf BTu Bf
kf
ku ¼ LqW0 ð6Þ
a ¼ W0 T b ð4Þ Bn Bn Bt

where W u=unfrozen water content, W 0=initial water

content and T f =freezing point. Tf ðfðt Þ; t Þ ¼ Tu ðfðT Þ; T Þ ¼ Tm ð7Þ

where f and u denote frozen and the unfrozen states,

4.4. Insulation material properties
respectively. L, W 0, T m and f are latent heat of water,
the initial water content, the frozen point of media and
The thermal conductivity of EPS is 0.03 W
the frozen and thawed interface, respectively.
m
1d 8C
1.The specific heat capacity and density
are 1400 J kg
1 8C
1 and 40 kg m
3, respectively.
5.2. Boundary condition and initial condition

On the basis of observed ground temperature at the

5. Governing equations and their finite element
depth of 0.5 m and use of Fourier theorem (Ku, 1992),
formulate
the top boundary temperatures were assumed as
follows:
5.1. Governing equations
T ð x; y; t Þ ¼ T0 þ g ðt Þ þ 13sinð2pt=8760 þ 5p=4Þ
By ignoring water infiltration into the embankment P P P
and assuming water movement is absent in the active ð x; yÞa BC; CD; DE ð8Þ
 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247 241

where
T ð x; y; t Þ ¼ T0 þ g ðt Þ þ 7:3sinð2pt=8760 þ 5p=4Þ
Z Z 

P P
ð x; yÞa AB; EF ð9Þ BNi BNj BNi BNj
Kij ¼ k þ dxdy ð15Þ
X Bx Bx By By

T0 ¼ Ta þ DT ð10Þ Z Z
KCij ¼ qcNi Nj dxdy ð16Þ
X
where T 0 and Ta are the embankment or ground
surface temperature and mean annual air temperature, Z
respectively. Ta was assumed to be
5.0,
4.5,
4.0, qi ¼ qNi dR ð17Þ

3.5,
3.3 and
3.0 8C during the computation. R

Based on observation data, DT are equivalent to 3 8C where N i and N j are the shape functions.
for ground surface and 4.5 8C for embankment
surface. g(t)=At denotes the climate warming effect.
where A is taken as 0.02 8C/year and t is time (year). 6. Numerical results and discussion
P
Below boundary KL held heat fluxes constant and
the temperature gradient at 30 m depth is 0.038 8C/m: Fig. 5(a) and (b) shows the temperature distribu-
tions of embankment with insulation on October 30 in
BT the 10th year and 50th year after its construction,
k ¼
q ð11Þ
Bn respectively. MAAT was taken as
3.5 8C, the
P P thickness of EPS for analysis was 10 cm, and the
The conditions at AL and FK are:
height of embankment is 2.0 m during computation. It
can be seen from Fig. 5(a) that ground temperature
BT under insulation rises in the 10th year after its
k ¼0 ð12Þ
Bn construction. The predicted maximum thaw depth
under side slope is same as that of the center of
Initial condition is as follows:
embankment if the insulation is placed at 0.5 m above
ground surface. It is shown in Fig. 5(b) that
T ð x; y; 0Þ ¼ T0 ð13Þ permafrost temperature rises greatly due to heat
accumulation and the predicted maximum thaw depth
where the initial temperature distribution T 0 was under side slope is higher than that of the center of
obtained through a long-term transient solution of embankment. So the predicted maximum thaw depth
Eq. (5) without consideration of the climate warming below the center of embankment was selected as a
effect by letting g(t)=0. way of evaluating rule for use of insulation.
5.3. Finite element formulation 6.1. The influence of insulation location in the
embankment
Because the specific heat and thermal conductivity
of media versus temperature and the frost front Referring to the local condition, MAAT was taken
position is to be determined, it becomes apparent that as
4.0 8C, the thickness of EPS for analysis was 10
the condition of energy conservation at the frost front cm. Fig. 6(a) and (b) shows the calculated maximum
position is nonlinear. Therefore, this is a nonlinear thaw depth for different insulation positions in over 50
problem and its analytical solution is not available. By years, with embankment heights of 3.5 and 2.5 m,
using the Galerkin residual weighted method, the respectively. Fig. 6(a) shows that the deeper the
following finite element formulae were obtained. insulation was placed, the better the thermal insulation
  efficiency is. Firstly, taking into account the time
BT when permafrost table begins to decline, it is
½ K fT g þ ½KC  ¼
½q ð14Þ
Bt approximately in the 9th year after construction of
242 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247

Fig. 5. The temperature distributions of embankment with insulation (MAAT:
3.5 8C) (including a and b).

embankment for the insulation at 2.0 and 3.0 m above maximum thaw depth is deeper than 3.0 m, they are
ground surface, while it is about the 30th year for the the 25th year, 23rd year and 19th year for insulation
insulation placed at 0.5 m above the ground surface. position at 0.5, 2.0 and 3.0 m above ground surface,
Secondly, considering the time when the predicted respectively. Thirdly, as for the predicted maximum
 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247 243

Fig. 6. The predicted maximum thawing depth for different positions of insulation (including a and b).

thawing depth of embankment with insulation placed attributed to the fact that the insulation with lower
at 0.5 m above ground surface in the 50th year, it is position cuts down the entry of heat absorption from
0.6 and 1.0 m shallower than that of insulation placed the slopes.
at 2.0 and 3.0 m above ground surface, respectively.
Besides, computational results of 2.5 m embankment 6.2. The relationship between the thickness of
produced similar results (Fig. 6(b)). insulation layer and the height of embankment
The computational result of insulation location
shows that the higher the insulation is placed, the In Fig. 7(a)–(c) are the predicted maximum thaw
shorter time permafrost table maintains. For the depth for 1.0, 2.5 and 3.0 m high embankment with
embankment with 50 years service life, lower position different thickness EPS, respectively. EPS was placed
for insulation is a better choice, which may be at 0.5 m above ground surface and the mean annual
244 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247

Fig. 7. The predicted maximum thawing depth for different insulation thickness (including a, b and c).
 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247 245

air temperature was assumed to be
4.0 8C during the condition of permafrost and climate, etc. when
computation. considering the thickness of insulation layer.
Fig. 7(a) shows that as far as high roadbed in the It can be seen from Fig. 7(a)–(c) that for the
next 20 years is concerned, the predicted maximum roadbeds with heights of 1.0, 2.5 and 3.5 m with
thaw depth of embankment with 10 cm thick EPS has gravelly soil cover, there exists substantial difference
little difference from that of embankment with 6 cm in maximum thaw depth between the roadbed with
thick EPS. The predicted maximum thaw depth of insulation and that without insulation. The simulated
embankment with 10 cm EPS in the 50th year is only results of embankment thermal regime show positive
20 cm less deep than that with 6 cm thick EPS. effect of the insulation on protecting permafrost.
For low roadbed, there is an obvious difference
(Fig. 7(c)). Firstly, the rise of permafrost table under 6.3. Applicable range of insulation in the embankment
the center of embankment with 10 cm EPS is about 50 engineering design of the Qinghai–Tibetan railway
cm more than that of embankment with 6 cm EPS.
Moreover, The predicted maximum thaw depth below As a countermeasure to protect permafrost, the
the center of embankment with 6 cm thick EPS in the insulation treatment has limitations, namely, while
50th year is also 60 cm deeper than that with 10 cm embankment with insulation maintains or raises the
thick EPS. Besides, taking into account the time when permafrost table, the temperature of permafrost under
the permafrost table begins to decline, it is about 3–4 the embankment rises and there is an extra heat
years later for embankment with 10 cm thick EPS than accumulation under insulation. Therefore, not all
that with 6 cm thick EPS. For 2.5 m high embankment permafrost regions can use thermal insulation method
(Fig. 7(b)), the change predicted in the permafrost to keep the stability of embankment. During the
table lies between that of the 3.5 m high embankment service time of 50 years, when MAAT is higher than a
and 1.0 m high embankment. certain value, the use of insulation cannot maintain the
The results from numerical simulation indicate that stability of embankments. On the other hand, there is
the thickness of thermal isolative material affects the a lower limit of an MAAT below which embankments
maximum thaw depth below the center of embank- without insulation may maintain the stability. The
ment slightly if the roadbed is high, while it has an predicted maximum thaw depth below the center of
obvious effect on the maximum thawing depth under embankment was selected as a way of evaluating rule
the center of the embankment for low embankments, for use of insulation.
which means the fill of the embankment has also a Fig. 8 is the applicable range of insulation in the
thermal insulation function. So the height of embank- embankment engineering of Qinghai–Tibetan railway
ment should be taken into account in addition to the in terms of MAAT, which can be used to estimate the

Fig. 8. The applicable range of insulation in Qinghai–Tibetan railway in terms of MAAT.

246 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247

required thickness of insulation or fill to maintain the slowly to the change of MAAT and the condition of
0 8C isotherm at or above original permafrost table ground surface. Moreover, for permafrost regions
during the service time of 50 years. The insulation with the same MAAT, mean annual ground temper-
was placed at 0.5 m above ground surface in ature (MAGT) differs greatly due to the condition
computation. As will be noted, there are several of ground surface and the direction of slope, etc.
different thickness of gravel or insulation that will However, MAGT is related to MAAT and the
provide the same level of thermal protection and the difference between ground temperatures can reach 1
choice is a matter of economy, combined with the 8C in permafrost areas with the same MAAT.
minimum thickness of gravel cover for structural Therefore, geothermal field of permafrost has
concerns. visible influence on applicable range of the insu-
As is shown in Fig. 8, the efficiency of insulation lation. For permafrost with low temperature,
improves slightly if the insulation is thicker than 10 cm. embankment with insulation may maintain perma-
So 10 cm thick EPS is a proper choice considering frost table and keep its stability, whereas heat
expense and intensity. It is impossible to guarantee the accumulation can result in the thawing of perma-
embankment stability of Qinghai–Tibetan railway frost with high temperature.
during the service time of 50 years if MAAT is higher Fig. 9 shows the influence of geothermal field of
than
3.84 8C, while 1 m high embankments without permafrost on applicable range of insulation in the
insulation may maintain the permafrost table and embankment of Qinghai–Tibetan railway. The thick-
control settlement if mean annual air temperature is ness of EPS is 10 cm and it is placed at 0.5 m above
lower than
5.20 8C. Taking expense and security into ground surface during computation. As for perma-
consideration, thermal insulation method may maintain frost regions with the same yearly average air
thermal stability of railway embankment if MAAT is temperature, the necessary height of embankment
between
4.0 and
5.20 8C. may decrease if MAGT is low. Similarly, the roadbed
structure can fit permafrost regions with higher
6.4. The influence of geothermal field of permafrost MAAT for the same height of embankment. More-
on applicable insulation range over, the influence of MAGT varies with the height
of embankment. MAGT has an obvious influence on
Ground temperature field is the result of the low fill while there is less influence on high
formation history of the permafrost, which responds embankment.

Fig. 9. The influence of geothermal field of permafrost on applicable range of insulation.

 W. Zhi et al. / Cold Regions Science and Technology 41 (2005) 235–247 247

7. Conclusions References

(1) Observed data showed that permafrost table Gandahl, R., 1978. Some aspects of the design of roads with boards
under insulated embankments rose and shallow of plastic foam [A]. Proceedings of 3rd International Conference
on Permafrost [C]. Edmonton, Canada, pp. 792 – 797.
ground temperatures also rose after embankment Johnson, G.H., 1983. Performance of an insulated roadway on
construction. The range of annual ground permafrost [A]. Proceedings of the 4th International Conference
temperature and heat exchange below the on Permafrost [C]. Fairbanks, USA, pp. 548 – 553.
insulation was all reduced greatly. Ku, B.A., 1992. Fundamentals of Frost Forecasting in Geological
(2) Numerical results show that 10 cm EPS is an Engineering Investigation. Lanzhou Univ. Press, Lanzhou.
(in Chinese).
appropriate choice considering economy and Li, S., 1983. The permafrost temperature and thickness along the
intensity. Moreover, the closer the insulation is Qinghai–Tibetan highway. Proceedings of the Symposium on
placed to the ground surface, the better the Glaciology and Cryopedology Held by Geographical Society of
thermal insulation efficiency. This is because China, Science Press, Beijing, pp. 7 – 10 (in Chinese).
less heat from the embankment slopes is allowed Liu, Y., Wu, Q., Zhang, J., Sheng, Y., 2002. Deformation of
highway roadbed in permafrost regions of Tibetan plateau.
to penetrate and warm the foundation soils. Journal of Glaciology and Geocryology 24 (1), 10 – 15.
(3) When considering the thickness of insulation, the Olson, M.E., 1984. Synthetic insulation in arctic roadway embank-
height of embankment should be taken into ment [A]. Proceedings of the 3rd International Cold Regions
account in addition to the condition of permafrost Engineering Specialty Conference [C]. Canadian Society of
and climate, etc. Taking expense and security into Civil Engineering, pp. 739 – 752.
Sheng, Y., Zhang, L., Yang, C., et al., 2002. Application of
consideration, thermal insulation method may thermal-insulation treatment to roadbed engineering in perma-
maintain thermal stability of railway embankment frost regions. Journal of Glaciology and Geocryology 24 (5),
if MAAT is between
4.0 and
5.20 8C. 618 – 622.
(4) Geothermal field of permafrost has influence on Wu, Q., Tong, C., 1995. Permafrost change and stability of
Qinghai–Tibet highway. Journal of Glaciology and Geocryology
applicable range of insulation and the influence
17 (4), 350 – 355.
of MAGT varies with the height of embankment. Xu, X., Wang, J., 2001. Frozen Soil Physics Science. Science
Press, Beijing (in Chinese).
Yu, W., Lai, Y., Niu, F., 2002. Temperature field features in the
Acknowledgements laboratory experiment of the ventilated railway embankment in
permafrost regions. Journal of Glaciology and Geocryology 22
(5), 601 – 607.
The research is supported by the grant of the Zhang, J., Yao, C., 1994. Applications of the industrial insulation
Knowledge Innovation Program of the Chinese material in roadbed engineering on permafrost. Proceeding of
Academy of Sciences, No. KZCX1-SW-04, and by the 1st Conference of Young Researchers in Cold Region
the National Key Basic Research and Development Environment and Engineering, Lanzhou Univ. Press, Lanzhou,
Foundation of the Ministry of Sciences and Tech- pp. 78 – 83 (in Chinese).
nology of China (Grant No. 2002CB412704) and by
the National Natural Science Foundation of China
(Grant No. 90102006).

View publication stats