Consideration of Seasonal Temperature Changes in The French Pavement Design Method

1
Consideration of Seasonal
Temperature Changes in the French
Pavement Design Method
The current French pavement design method assumes a

value of “ eq” for the equivalent temperature that is constant
throughout the year and for the entire lifespan of the road
structure. The goal of this study is to assess and quantify the
effects of seasonal variations in temperature on the
durability of road structures. In this work, three research
sites were chosen. The first is located near Bordeaux. It has
an oceanic climate typical of the Aquitaine region, which is
characterized by very mild winters and very warm summers.
The second site is located in Saint-Chély-d'Apcher, also in
France. Its climate is characterized by harsh winters and
sizeable temperature fluctuations occurring between day and
night, and winter and summer. The third site is located in
the experimental forest of Montmorency, north of Quebec
City in Canada. Its cold and humid climate is largely due to
its northerly and maritime location. The temperatures found
both on the surface and inside road structures in these three
sites were measured over several years. In this study, we
Chapter written by Rahma KTARI, Ferhat HAMMOUM, Pierre

HORNYCH, Denis ST-LAURENT, Paul MARSAC, Mai Lan NGUYEN and
Jean-Michel PIAU.
Risk Evaluation and Climate Change Adaptation of Civil Engineering Infrastructures and
Buildings: Project RI-ADAPTCLIM, First Edition. Edited by Panagiotis Kotronis.
© ISTE Ltd 2019. Published by ISTE Ltd and John Wiley & Sons, Inc.
2 Risk Evaluation and Climate Change Adaptation of Civil Engineering
have focused on three issues. The first was the effects of

spatial variations (site and climate) in temperature on
calculating the equivalent temperature eq. The second
relates to the effect of the structure type (inverted or thick
bituminous) on calculating the eq. Finally, the third issue is
the effect of eq on determining the thickness of the
underlying bituminous layer and on the lifespan of the thick
bituminous structure.
1.1. Introduction
Over the course of the last few decades, global warming

has fluctuated with a significant rise since the 1980s. From
1959 to 2009, the observed trend is an increase of around
0.3°C in the average temperature of the surface of the globe
each decade (from ONCERC1, http://onerc.org). During this
same period, the three warmest years were 2003, 2011, and
2015, respectively, thus all occurring in the 21st century.
According to Météo France (2017), the average temperature
during summer of 2017 was higher than normal with a 1.5 C
increase, making it the second warmest summer, at the
same rank as 2015, nevertheless far behind 2003 which saw
a 3.2 C increase.
A recent review of climate models has allowed

climatologists (Brown and Caldeira 2017) to confirm that
there is a more rapid increase in climate change than
predicted by the most pessimistic model created by GIEC2 in
2014. According to their research, at this pace, the range of
average temperature increase across the surface of the globe
predicted for 2081—2100 will be between +2.6 C and +4.8 C
when compared to preindustrial levels (1860).
1 Observatoire national sur les effets du réchauffement climatique,

created by the law of February 19, 2001.
2 Groupe d’experts intergouvernemental sur l’évolution du climat,
founded in 1988.
Seasonal Temperature Changes in the French Pavement Design Method 3
Furthermore, many countries are confronted with

significant differences in seasonal temperatures. As such,
the durability of road structures may need to be reassessed if
no precautions are taken on climate change, particularly in
areas that have a high temperature gradient for summer
and winter (Mauduit et al. 2013). Now more than ever, it is
crucial that the durability of road structures depending on
climatic conditions is studied in order to guarantee their
continuity (Hammoum et al. 2016). This study is anchored in
the framework of the RI-ADAPTCLIM project and financed
by the Pays de la Loire region. On taking these issues into
consideration, the main research question is as follows: how
we must improve our current design methods in order to
better mitigate the effects of climate on road infrastructure?
In practice, the designing of a road involves many

variables, thus making it a particularly complex task for
civil engineers (Doré et al. 2014). Indeed, over many seasons,
road structures are subject to the results of both heavy
vehicle traffic and climate.
Road structures consist of multiple layers (LCPC-SETRA

1998). Different types of road structures are as follows:
flexible structures, thick bituminous structures, semi-rigid
structures, mixed structures, rigid structures, and inverted
structures. Each of these will deteriorate differently (AIPCR
2014). Climatic phenomena contribute toward this
deterioration; during winter, the process of freezing and
thawing weakens the roads, particularly flexible structures
(Vu et al. 2018), by causing a loss of bearing capacity of the
soil during thawing (Rouquès and Caniaro 1975). During
summer, rutting caused by a viscoplastic creep of the
bituminous layers is the main source of potential road
degradation (Sohm et al. 2011). Finally, the effect of solar
rays upon the surface layers can cause hardening of the
bitumen, which can result in multi-directional thermal
fissuring (Mauduit et al. 2013).
Variations in temperature, whether daily or seasonally,

can strongly influence roadway performance as bituminous
materials are thermally susceptible, which can result in
thermal stress fields, in addition to the stresses caused by
road traffic (Hornych et al. 2013; Rouquès and Caniaro
1975). However, the effects of real variations in
temperatures on the design of bituminous roads have been
the subject of very little research work in France.
In France, the LCPC3 (currently IFSTTAR4) developed a

rational, mechanical, and empirical pavement design
method. This method takes into account:
— the load-bearing capacity of the roadbed, regarding the
geotechnical, mechanical, and hydric properties of its
composing materials;
— the mechanical performance of the treated materials
determined through standardized laboratory testing;
— the mechanical performance of unbound granular
materials defined empirically according to the composition of
these materials.
Traffic conditions are assessed by the load aggressiveness,

reflected in the notion of a number of equivalent axle loads
(the number of standard 130 kN dual wheel axles; Goacolou
et al. 2003), (NF P98-086; Balay et al. 2012). This approach
can predict the performance of the structure and thus allows
for structural design for a constant equivalent temperature.
For projects based in mainland France and in Saint Pierre et
Miquelon, this temperature is 15 C. For French overseas
territories, the temperature is 25 C and 28 C for Guyana.
This method could be improved by considering a more
precise reading of climatic conditions of the individual
locations of each particular project.
3 Laboratoire central des ponts et chaussées.

4 Institut Français des Sciences et Technologies des Transports, de
l’Aménagement et des Réseaux.
In this chapter, we consider that the hydric state stays

relatively constant, and as such, we have concentrated solely
on the influence of temperature on the pavement design.
Three research sites, with different climates, were

selected in order to highlight the effects of temperature
variations: two sites are located in France and one in
Canada. The first is near Bordeaux, France, on the A63
highway. It has an oceanic climate typical of the Aquitaine
region, which is characterized by very mild winters and very
warm summers.
The second site is located in Saint-Chély-d'Apcher,

France, in the H4 climate zone (defined in section 1.2), on
the A75 highway (in the Massif Central). Its climate is
characterized by harsh winters and sizeable temperature
fluctuations occurring between day and night and between
winter and summer. The third site is located in the
experimental forest of Montmorency, 70 km north of Quebec
City in Canada. Its cold and humid climate is largely a result
of its northerly and maritime location, which has four
extremely distinct seasons.
The objective of the study is to assess the effect of internal

road temperature on fatigue damage throughout the course
of a year for each of the three chosen sites. We wish to
estimate the fatigue damage of the road under various
temperatures and to calculate the equivalent temperature
for each site. Finally, we assessed the repercussions of these
specific equivalent temperature values on the thick
bituminous structure of the A63 highway using the French
design method, and those results are compared using the
standard temperature of 15 C.
This chapter has five sections. The first section introduces

the context of the study. The second presents the studied
structures and their corresponding climates. The third
describes the French design method for new pavements,
namely, the structural calculation method, as well as the

consideration of the temperature data for the calculation of
the equivalent temperature. The fourth section proposes an
incremental calculation procedure for the determination of
the equivalent temperature, by using real temperatures
measured on the different experimental sites. The fifth
section concerns the presentation of equivalent temperatures
calculated by this new procedure and the consequences upon
bituminous pavement design. This section summarizes the
discussion and the results for each of the questions under
exam. Finally, the sixth section presents conclusions and
perspectives.
1.2. The experimental road structures
Based on the cartography of the average temperatures in

France, as shown in Figure 1.1a, different types of climates
can be defined through spatial construction (Joly et al. 2010).
Furthermore, according to Mauduit et al. (2008), France is
divided into four different zones of winter severity, according
to the average amount of snow and ice days (with or without
rain) (cf. Figure 1.1b). Based on a map that shows the
variations in winter severity, which has been established at
a national level, two sites with temperature measurements
at different depths were selected. The first is located on
the A63 highway near Bordeaux, in a homogenous climate
zone, which is classed H1 (mild winter zone). The second
is on the A75 highway in Saint-Chély-d'Apcher near
Clermont-Ferrand, in the Lozère department. It was chosen
because it is exposed to significant climatic stress (harsh
winter H4).
The third site is the Site Expérimental Routier de

l’Université Laval (SERUL) situated in the forest of
Montmorency, approximately 70 km north of Quebec City
(Badiane 2016). It represents a Quebecois road network that
is exposed to particularly harsh climatic conditions, with
extremely cold temperatures in winter and hot temperatures

during the summer, which create, on an annual basis,
thermal variations of over 60 C. This situation is obviously
unfavorable for roads (MTQ 2016).
(a)
(b)
Figure 1.1. (a) Map of average annual temperatures (Joly et al. 2010).
(b) Different zones of winter severity (Mauduit et al. 2008). For a color
version of this figure, see www.iste.co.uk/kotronis/risk.zip
Table 1.1 shows the comparison the French and Quebecois

contexts. Two particular aspects become apparent: the
harshness of the climate and the length of the road network
in Quebec in relation to the number of inhabitants.
The chosen sites are exposed to sizeable differences in

seasonal temperatures; the meteorological temperatures in
the shade vary as follows:
— from 8 C in winter to +38 C in summer (Bègles station
near Bordeaux), from August 1, 2003, to August 1, 2013;
— from 18 C to 35 C (Saint-Chély-d’Apcher station at
1,000 m of altitude near Clermont-Ferrand), from August 1,
2003, to August 1, 2013;
— from 30 C in winter to +30 C in summer for Québec,
from November 2010 to July 2017.
Quebec France
Length of the road network in km
30,600 20,000
National roads and highways
Number of inhabitants (in
7.9 64.7
millions)
Average annual rainfall (mm) 1,000 800
Min-max (1981—2010) 292 to 1,5475 515 to 1,4516
Length of frost (days/year) 147 to 218 0 to 90
Frost penetration depth (m) 1.2 to 3 0 to 0.8
Average annual temperature (°C) 4.8 12.6
Monthly temperature (min-max)
18 to 257 5 to 20.8
(°C) (1981—2010)
Table 1.1. State-managed road networks: a comparison between

French and Quebecois networks (MTQ 2016)
5 http://www.mddelcc.gouv.qc.ca/changements/classification/index.
htm
6 http://education.meteofrance.fr/climatologie-et-records-de-precipi
tations
7 https://www.ville.quebec.qc.ca/touristes/meteo/index.aspx
1.2.1. The A63 highway (Bordeaux, France)
The first structure studied in France is located on the A63

highway near Bordeaux (see Figure 1.2a). It is a thick
bituminous structure with a bituminous road base.
It is composed of three asphalt layers: a semi-coarse

bituminous concrete (EB-BBSG3) wearing course of 8 cm
thickness; and two bituminous mix (GB3) base and subbase
layers, each in 13 cm thickness (see Figure 1.2b). The road
base bituminous mix is a GB3 (according to French
specifications), with 4.3 ppc of grade 35/50 pure bitumen.
This type of structure, with a total thickness of 34 cm, comes
from the French pavement structure catalog (LCPC-SETRA
1998). These layers rest on a PF3 roadbed, which is
characterized by an elastic modulus (E) of 120 MPa for
design calculations.
(a)
Surface layer in BBSG (8 cm)

Base layer in GB (13 cm)
Foundation layer in GB (13 cm)
Roadbed PF3
(b)
Figure 1.2. (a) Location of the A63 highway. (b) Pavement structure of
the A63. For a color version of this figure, see www.iste.co.uk/kotronis/risk.zip
The standard design parameters for this structure are

shown in Table 1.2.
E(MPa) ε6 (10 C,
Layer ν 1/b kc
(15 C, 10Hz) 25Hz)
BBSG3 7,000 0.35 100 5 1.1
GB3 9,000 0.35 90 5 1.3
Table 1.2. Standard design parameters of the materials according

to the NF P98-086 standard (ANNEX F)
6 is the strain leading to fatigue failure for 1 million loads

in a two-point bending fatigue test (NF EN 12697-24); 1/b is
the slope of the fatigue line, see equation [1.1]; and kc is the
calibration coefficient of the design method, which allows for
the recalibration of the predictions produced by laboratory
fatigue tests, with lifespans observed on actual roads.
Hourly temperature variations, as a function of depth, are

available for the 10 years between August 1, 2003, and
August 1, 2013, at the Bègles station near Bordeaux. An
example of temperature data for 2006, for seven probes
placed up to a depth of 1 m, is shown in Figure 1.3. From
this data, it is possible to calculate the equivalent
temperature with the French method (section 1.4). This site
is characterized by a seasonal variation in surface
temperature from 12 C in winter to +52 C in summer
(see Table 1.3) and from 3 C to +34 C for the subbase layer.
The average annual temperature for this site is 16 C.
Temp. ( C) Tair Tsurf T10 cm T20 cm T30 cm T50 cm T70 cm T1 m

Minimum 8.1 11.9 8.0 4.9 2.7 0.2 1.9 4.9
Maximum 38.7 51.6 42.5 37.3 34.3 31.3 29.8 27.5
Range 46.7 63.5 50.5 42.2 37.0 31.1 27.9 22.6
Average 13.6 16.1 16.0 16.0 16.0 16.0 16.0 16.0
Table 1.3. Recordings of extreme values of hourly temperatures at the

A63 site (from August 1, 2003 to August 1, 2013)
Figure 1.3. Hourly temperature changes for different depths in the

A63 structure for 2006. For a color version of this figure, see
www.iste.co.uk/kotronis/risk.zip
1.2.2. The A75 highway (Saint-Chély-d'Apcher, France)
The second site chosen for the project is the A75 highway
near Reference Point 124, direction 2 from Montpellier to
Clermont-Ferrand, in the Lozère department (Figure 1.4a).
It was chosen for its harsh climatic exposure. Traffic is in
class T0 with approximately 13,800 vehicles per day in both
directions with 13% heavy goods vehicles.
The structure is located at an altitude of about 1,000 m, in

the harshest climatic zone (H4). It has a south facing
exposure that provokes a large range in temperatures
between night and day. Its instrumentation has been
described by Mauduit and Florent (2015).
(a)
Surface layer BBMa 0/10 (4 cm)

Binder layer BBSG3 (5 cm)
Base layer in BBSG3 (9.5 cm)
GNT-B (12 cm)
Crushed stone, pozzolan, lime GPZC (25 cm)
Treated subgrade (35 cm)
Roadbed PF3
(b)
Figure 1.4. (a) Location of the A75 highway. (b) Pavement structure
of the A75. For a color version of this figure, see
It is a so-called “inverted structure” composed of a surface

layer, a binder layer and a base layer of asphalt concrete, a
layer of unbound granular material (UGM), and a foundation

layer of material treated with hydraulic binders. As
indicated in Figure 1.4b, the surface layer is 4 cm thick and
is made of type A thin asphalt concrete (BBMa), whereas the
binder layer is 5 cm thick and the base layer is 9.5 cm and
both are made of semi-coarse bituminous concrete (BBSG3).
These layers are placed on a layer, with a thickness of

12 cm, of Humidified Reconstituted crushed stone (GRH)
with an elastic modulus of 480 MPa, whose role is to prevent
crack propagation. The subbase layer is composed of
pozzolan, lime, and crushed stone (GPZC). It has a thickness
of 25 cm and an elastic modulus of 23,000 MPa. All these
layers rest on a 35 cm treated subgrade. The roadbed is of
class PF3.
Hourly temperature variations as a function of depth are

available for 10 years between August 1, 2003, and August 1,
2013, at the Saint-Chély-d'Apcher station. An example of
temperature data for 2006, for seven probes placed up to a
depth of 1 m, is shown in Figure 1.5. From these data, it is
possible to calculate the equivalent temperature with the
French method (section 1.4). This site is characterized by a
significant seasonal variation in surface temperature, with a
range of 65 C at surface level (see Table 1.4). The average
annual temperature is constant for the structure at 11 C,
thus lower than that for the A63 site (16 C).
Temp.
Tair Tsurf T10cm T20cm T30cm T50cm T70cm T1m
(°C)
Min. 17.5 17.0 13.5 10.6 8.6 5.9 4.0 0.0
Max. 34.8 48.1 38.5 32.6 29.6 26.8 25.4 23.2
Range 52.3 65.1 52.0 43.2 38.2 32.7 29.4 23.2
Average 8.0 11.1 11.0 11.0 11.0 11.0 11.0 11.1
Table 1.4. Recordings of extreme values of hourly temperatures at the

A75 site (from August 1, 2003 to August 1, 2013)
Figure 1.5. Hourly temperature evolution for different depths in the

A75 structure for 2006. For a color version of this figure, see
1.2.3. Laval University Experimental Road Site (SERUL)

(Montmorency forest, Canada)
The third pavement structure is situated at the

experimental road site of the Université Laval or Site
Expérimental Routier de l’Université Laval (SERUL),
located in Montmorency forest (see Figure 1.6a),
approximately 70 km north of Quebec City, near kilometer
173 (Badiane 2016).
This thick bituminous structure (see Figure 1.6b) has the

asphalt layer with a total thickness of 20 cm. It is composed
of semicoarse surface and binder layers (two layers of ESG-
10 of 5 cm each) and of a 10 cm thick road base asphalt mix
(GB-20). Each of these layers is made with a PG 58-34
performance grade bitumen. These layers are placed on a
20 cm thick subbase layer, composed of 0/20 mm unbound
granular material (MG-20) and a 45 cm foundation, made
from sand (MG-112, with 72% passing at 5 mm and 5%
passing at 80 μm, respectively) and finally a subgrade

composed of silty sand (unified classification SM with 21%
passing at 80 μm). The standard parameters for the design of
this structure with the Alize-LCPC software are shown in
Table 1.5.
(a)
Bituminous concrete (20 cm)

Granular subbase (20 cm)
Foundation (45 cm)
Subgrade
(b)
Figure 1.6. (a) Location of SERUL. (b) Structure of the

SERUL pavement (Badiane 2016)
E(MPa) ε6
Layer 15 C and 10 ν (10 C, 25 1/b kc
Hz Hz)
GB-20
5,220 0.35 186 5 1
PG58-34
Table 1.5. Standard design parameters of the materials according to the NF

P98-086 standard (ANNEX F)
The results of complex modulus tests on the bituminous

mixes have been published by Doucet and Auger (2010). The
values of elastic moduli of the unbound materials are shown
in Table 1.6.
Type of layer Modulus (MPa)

Subbase (MG 20) 385
Foundation (MG 112) 109
Subgrade (SM) 54
Table 1.6. Elastic moduli of unbound materials
A road weather station with a 3 m long probe (18

thermistors) was installed at SERUL. The daily temperature
data have been recorded at the Jacques Quartier National
Park road weather station (see Figure 1.6a) from November
2010 to July 2017 for various levels of depth (air, 5 cm,
10—90 cm at 10 cm intervals, and 110—300 cm with 20 cm
intervals).
An example of the temperature data for 2012, for seven

probes placed up to a depth of 1 m, is shown in Figure 1.7.
This site is characterized by a significant seasonal variation
in temperature between winter and summer. The recordings
show a variation in air temperature that ranges from a
minimum of 30 C to a maximum of 23 C (Table 1.7). The
variation is reduced from 11 C to 21 C at a depth of 90 cm
and from 0 C to 12 C at a depth of 3 m.
Temperature ( C) Air 5 cm 10 cm 20 cm 40 cm 60 cm 90 cm 110 cm 300 cm
Minimum 0
Maximum 23 31.7 28.4 28 24.9 23.6 21.3 20 12.3
Range 53.8 55.8 51 48.7 41.8 38 32.6 29.7 12.3
Average 1.45 5.55 5 5.49 5.15 4.98 4.84 4.79 5.25
Table 1.7. Extreme values of daily temperatures at the SERUL site

(November 24, 2010 to July 1, 2017)
Seasonal Temperature Changes in the French Pavement Design Method
17
40 air surf 10 cm 20 cm
30 cm 40 cm 60 cm 90 cm
30
Temperature (°C)
20
10
0
-10
-20
-30
Figure 1.7. Daily temperature evolution for different depths at the

SERUL structure for 2012. For a color version of this figure, see
1.3. The French pavement design method
This section describes the main principles of the French

pavement design method, focusing on the criteria for
structural calculation and the calculation of the equivalent
temperature.
1.3.1. Structural calculation: design criteria
The French pavement design method combines theoretical

pavement mechanical calculations, the results of laboratory
modulus and fatigue tests on treated pavement materials,
and data from feedback from full-scale experiments: trial
sections on various networks, and experiments conducted on
the LCPC accelerated pavement testing facility (Corté and
Goux 1996, p. 2; LCPC-SETRA 1994). This method was
initially developed by LCPC and SETRA, and is currently in
line with the NF P98-086 standard (2011). The main input

parameters for pavement design calculations are as follows:
— Traffic: this expresses, for a traffic lane, the number of
vehicle passages for a given timeframe. It is composed of
different types of vehicles, with a wide variation in loads,
axles, and shapes. For design calculations, traffic is
converted into an equivalent number of reference axle loads,
defined as single axles with dual wheels, with a total load of
130 kN. In practice, for the calculation of the mechanical
response of the road structure, only a half-axle (two wheels,
loaded at 65 kN) is considered.
— Parameters linked to the management strategy: this is
the design life, and the risk coefficient, which defines the
probability of failure of the road, associated with an
economic objective, which varies according to the type of road
network.
— The mechanical parameters of the road materials
determined from laboratory test results. They include the
elastic moduli and Poisson’s ratios, which are used to
calculate the stress—strain response of the pavement, and
the fatigue parameters that are used to calculate the design
life.
The design method consists in calculating the maximum

stresses and strains within the pavement structure using a
mechanical model (multi-layer linear elastic model), and
comparing these calculated stresses or strains with
permissible design values. These design values depend on
the cumulative traffic, the calculation risk, the fatigue
resistance of the materials, and a calibration coefficient,
introduced to take into account the feedback from the in-use
pavement performance. The stresses and strains created by
the reference load are calculated at the bottom of each
treated layer and at the top of the unbound granular layers
and of the subgrade, by using the multi-layer linear elastic
and isotropic model (Burmister 1945). This model is
integrated with the calculation software Alize-LCPC (itech

2017), which is the baseline software of the French
pavement design method.
The Alize-LCPC software can also calculate the

permissible stress or strain values, function of the level of
traffic, and performance of the materials. The stress and
strain values that are calculated for the reference load are
compared to the permissible values of the materials that
make up the road. The design is considered satisfactory for
all underlying layers and the roadbed when the maximum
calculated stresses or strains are lower than the permissible
values.
This software has been used to develop the French

pavement design guide (LCPC-SETRA 1994) and the
pavement design catalog (LCPC-SETRA 1998). Alize-LCPC
can be used to design all pavement types (flexible, rigid,
concrete, etc.) on all types of road networks (urban,
interurban, public transport carriageways, industrial roads,
airstrips, port roads, etc.).
The permissible values are calculated with three general

formulae presented in the following:
— Permissible strain of bituminous materials:
ε, ε θ k k k [1.1]
— Permissible stress of materials treated with hydraulic

binders and cement concrete:
σ, σ k k k k [1.2]
where εt,ad is the permissible tensile strain at the base of

the bituminous layers; σt,ad is the permissible tensile stress
at the base of the layers treated with hydraulic binders and
cement concrete; NE is the reference number of equivalent
axles (in terms of material damage) that correspond to the

cumulative number (NPL) of heavy vehicles throughout the
period under calculation so that NE = CAM NPL, with
CAM being the average coefficient of aggressiveness of the
traffic; ε6 eq) is the fatigue strain parameter of the
bituminous material, representing the tensile strain leading
to a lifespan of 106 cycles with a probability of failure of 50%
for the equivalent temperature eq.
Due to temperature adjustments (detailed in the

following, see equation [1.5]), ε6 eq) is estimated with the
value ε6 10 C, 25 Hz), which is the fatigue law of the
material, determined by standardized two-point bending
fatigue tests at 10 C and 25 Hz (NF EN 12697-24); σ6 is the
parameter of the fatigue law, representing the tensile stress
leading to a fatigue lifespan of 106 cycles with a probability
of failure of 50%, for specimens with a curing time of 360
days (NF P 98-233-1); b is the slope of the fatigue law of the
material ( 1 < b < 0); kr is the coefficient that adjusts the
values of the permissible strains to the calculation risk,
taking into account the dispersion factors on the layer
thickness (standard deviation Sh) and on the fatigue results
(standard deviation SN). For a calculation risk of 50%, kr = 1;
kc is an empirical calibration factor, ensuring the adjustment
of the design model to the feedback from experimentations
done on road networks, test sections, and on the LCPC
accelerated pavement testing facility (kc 1) (see Table 1.8);
ks is a coefficient dependent on the elastic modulus of the
subgrade, NF P98-086 (see Table 1.9); and kd is the
coefficient needed to take into account the effects of
geometrical discontinuities (such as longitudinal edges,
joints, and transversal cracks) and vertical temperature
gradients in base layers made with either cement concrete or
materials treated with hydraulic binders.
Base course Bituminous High modulus

Material
bituminous mix concrete mix (EME)
kc 1.3 1.1 1
Table 1.8. Coefficient kc values for bituminous materials
Subgrade
modulus [20, 50] [50, 80] [80, 120] [120, 200] 200
(MPa)
ks 1/1.2 1/1.1 1/1.065 1 1
Table 1.9. Coefficient ks values (see NF P98-086)
The vertical strains in the unbound granular materials

and in the subgrade, which can potentially include a treated
or unbound capping layer, are verified in relation to the
permissible strain (εz,ad), which is calculated as:
.
ε , A NE [1.3]
where εz,ad is the permissible vertical strain in contraction

on the surface of the unbound materials and of the subgrade
and A is a parameter which takes a value of 0.012 for new
roads, with medium to heavy traffic (PL 250,000) and a
value of 0.016 for other types of new roads, with the
exception of the granular layers of inverted structures, for
which A shall be 0.0144.
1.3.2. Principle for calculating the equivalent temperature

of bituminous materials
The stiffness and fatigue performance of bituminous

materials vary with temperature (Bodin et al. 2010; Doucet
1999; Ktari et al. 2017). As a result, the strain levels and
rates of damage of a pavement composed of one or more
bituminous layers fluctuate throughout the year with the
temperature cycles.
According to Appendix 6 of the (LCPC-SETRA 1994)

guide, design calculations are carried out for a constant
temperature, named equivalent temperature θeq. This
temperature is calculated in such a way that the annual sum
of damage to each of the different materials with the real
temperature distribution on the site under consideration will
be equal to the damage that these materials would have
undergone under the same level of traffic but for a constant
temperature θeq.
For the French design method, it is important to know the

variations of the fatigue parameter (ε6) of the material under
study with temperature (Ktari et al. 2017). The French
method (LCPC-SETRA 1994) uses Wöhler curves to describe
the fatigue law of road materials. It is represented by the
following equation for bituminous materials:
Equation [1.4]:
ε NE, θ , f ε θ NE/10 [1.4]
with:
.
° ,
ε θ ε 10°C, 25 Hz [1.5]
,
It is interesting to note that the pavement design model

assumes that frequency F has no effect on the fatigue
performance of the materials, as given in equation [1.5], but
for road pavement design only. However, the effect of
frequency on 6 is taken into account for the design of airfield
pavements, industrial and port pavements for heavy loads,
tramway tracks, etc.
From this laboratory fatigue law, it is possible to derive

the law of material damage for structural design, by adding
adjustment and calibration factors (kr, ks, and kc), thus
leading to equation [1.1].
For positive temperatures, the evolution of fatigue

performance with temperature is represented, according to
NF P98-086 (equation [1.6]), as follows:
ε θ E θ constant [1.6]
In current practice, n is considered equal to 0.5, unless

fatigue test results for different temperatures are available.
This relation of thermal susceptibility was compared with

those obtained from fatigue tests at different temperatures,
which can be found in the literature (Ktari et al. 2018). The
results of this study show a significant influence of the
choice of the susceptibility function 6( ) on design results.
The calculation method for equivalent temperature in the

French method is detailed in Appendix 6 of the Pavement
Design Guide (Hornych et al. 2013; LCPC-SETRA 1994) and
Afnor standard NFP98 086—2018 revision, Appendix G. The
equivalent temperature is determined through the
application of Miner’s law (equation [1.7]), which introduces
the notion of long-term damage (Miner 1945).
∑ D ∑ 1 [1.7]
where Di is the partial damage resulting from repeated

loading during period i, following an incremental analysis;
ni is the number of loads applied to the pavement during
period i; and Ni is the number of permissible loads according
to the fatigue law of the material, under the conditions of
period i.
Following the French pavement design standard

(NF P 98-086) and considering the expression of the fatigue
law (equation [1.4]), the equivalent temperature is implicitly
defined by equation [1.8]:
∑ n θ 10 [1.8]
∑
where N( eq) is the total annual traffic applied at equivalent

temperature eq; 6( i) is defined earlier, see equation [1.1];
( i) is the maximum tensile design strain of the bituminous
pavement layers, under standard axle loading (the value
being determined from a structural calculation of the
projected pavement structure, with values of elastic moduli
of the bituminous layers corresponding to temperature i);
P i(Z) is the variable temperature for depth Z which
corresponds to the temperature profile of the bituminous
layers; and criti is the temperature at the level whereby the
fatigue criterion is applied, determined by calculating the
spatial average of temperatures in a “thin layer”, located at
that level, see the option in Alize-LCPC “Subdivision in thin
sub-layers” (section 5.2).
Only the fatigue damage of the asphalt layers is

considered in the above-mentioned case. However, the design
criterion can equally be the criterion zadm for the subgrade
and the granular layers, or even the criterion tadm for
hydraulic materials (see mixed structures, and especially the
GPZC material used in the inverted pavement structure of
Saint-Chély-d'Apcher).
1.4. Incremental method for calculating the equivalent

temperature using the “Alize-LCPC Recherche” software
A method of incremental damage calculation with fixed

time steps, which takes into account the real temperature
variations throughout the year obtained from instrumented
pavement sections, was developed and implemented in the
V.1.5.1 version of the pavement design software Alize-LCPC
(in the research options of the software).
The equivalent temperature depends on the type of road

structure, on the thickness of the structure, particularly the
total thickness of bituminous materials, on the daily traffic
distribution, on temperature variations in the pavement
(climate conditions), on the values of modulus (E), on the
materials’ fatigue strength (ε6), and on their variations with
temperature.
Calculating the equivalent temperature of a given road

structure means calculating the cumulative damage over an
entire year that is representative of the climatic conditions
for the entirety of the calculation period of the project,
during which we have access to a number of datasets: traffic,
temperature, and descriptive parameters of the mechanical
performance of the layers of the road.
Calculating the equivalent temperature is carried out

over nine steps that are detailed below:
1. Define the structure and the moduli of the bituminous

and non-bituminous layers. Later in the study, an example
of the calculation of the equivalent temperature for the thick
bituminous pavement of the A63 highway
(8BBSG/13GB/13GB/PF3) will be presented (Figure 1.2b).
2. Process real temperature data according to the selected
time step (hourly, daily, weekly, or monthly). The time step
generally used for measuring road temperature is 1 hour.
Daily, weekly, or monthly temperature profiles can be
obtained by calculating average values for the chosen time
interval. Figure 1.8 shows an example of how Alize-LCPC
represents measured vertical temperature profiles.
(a)
(b)
Figure 1.8. Vertical temperature profiles for the A63 highway for 2006 (Alize-
LCPC search): (a) 8,760 hourly profiles/year and (b) 53 weekly profiles/year.
For a color version of this figure, see www.iste.co.uk/kotronis/risk.zip
3. Describe the evolution of the bituminous layer moduli

(surface, base, and foundation layers), as a function of
temperature, by using reference material properties from
standard NF P98-086 (Appendix F), or else from the results
of laboratory testing. In Figure 1.9, the curves E = f( ) are
based on data from Appendix F.
Figure 1.9. Evolution of the complex modulus E of bituminous concrete

(BBSG3) and road base bituminous mix (GB3) as a function of temperature
according to Appendix F of standard NF P98-086
4. Calculate the maximum tensile strains εt (θ(Z)) at the

base of the bituminous subbase layer of the structure and
the maximum vertical strains εz (θ(Z)) at the top of the
subgrade, for the vertical temperature profiles corresponding
to each selected time step. In Figure 1.10, strains εt and εz
are presented for a time step of one week, as a function of
the temperature at the bottom of the subbase. These
calculations were carried out with a standard dual wheel
axle load of 65 kN, for the A63 and A75 highways in France
(see Table 1.10). For the SERUL structure, the description of

the reference axle load in Quebec is also shown in Table 1.10.
A63 highway, Bordeaux

Location SERUL, Quebec
A75 highway, Bordeaux
Single axle with dual Single axle with dual
wheels loaded at 130 kN wheels loaded at 80 kN
Standard
dual
wheel
load
375 mm 305 mm
r =125mm r = 87,7mm
Table 1.10. Description of reference axle loads
Figure 1.10. Seasonal variations of tensile strains at the base of the road
bituminous mix, for 2006 (weekly interval) for the A63 highway. For a color
5. Choose the thermal susceptibility function ε6( ). For

this study, three functions were chosen for ε6( ):
— The simplified function from the Design Guide
(LCPC-SETRA 1994) defined by equation [1.6], see Figures
1.11a and 1.12.
— Functions determined from experimental results. Two

functions from the literature were selected: for the two
French structures, a function based on two-point bending
fatigue tests on a GB3 mix (Bodin et al. 2010) (Figure 1.11b);
for the structure located in Quebec, a function based on the
results of fatigue tests on a GB-20 mix, based on the work of
Doucet (1999), who used four-point bending tests (Figures
1.11c and 1.12).
(a) (b)
(c)
Figure 1.11. Example of Alize-LCPC’s interface for filling in the parameters

of ε6 = f(θ), represented as a third-degree polynomial. (a) Design Guide
(LCPC-SETRA 1994); (b) (Bodin et al. 2010); and (c) (Doucet 1999)
Figure 1.12. Evolution of the design parameter ε6(θ)/ε6(10 C) as a function of

temperature (Ktari et al. 2017). For a color version of this figure, see
6. Define the typical hourly breakdown of traffic, based on

data from traffic counting stations. In this study, the traffic
data were not available. In this case, five “typical’
histograms of hourly traffic proposed by the research version
of the Alize-LCPC software can be used. Figure 1.13 presents
two examples of traffic distributions available in Alize-LCPC
that have been used in this study. The first histogram
defines seven time periods, with heavier traffic during the
periods 6—12 hours and 14—19 hours. The second histogram
supposes a uniform distribution of traffic over an entire day
(24 hours).
(a) (b)
Figure 1.13. Examples of histograms of hourly traffic distribution available in

Alize-LCPC
7. Calculate the elementary damage (equation [1.9]) for

one load, then the cumulative damage, through an
incremental process, for measured variations in temperature
with the chosen time step (hourly, daily, weekly, or
monthly), and taking into account one of the three relations
of thermal susceptibility ε6( ) (Figure 1.11).
/
d θ 10 . [1.9]
In this chapter, the calculation is carried out solely for the

road base, and using only the fatigue criterion for
bituminous materials. The complete study would require the
generalization of equation [1.8], and to calculate eq for the
other design criteria, εz for UGMs or subgrade soils, and σt

for hydraulically bound materials.
We underline that basic damage as expressed by equation

[1.9] is not the material damage in the road structure in use.
Indeed, adjustments and calibration coefficients kr, ks, and kc
should logically intervene in the expression of this structural
damage (see equation [1.1]). In fact, equation [1.9] expresses
the value of the elementary damage corresponding to
laboratory test conditions. When calculating the equivalent
temperature eq, this simplification is of no consequence, as
the product of kr ks kc will intervene later in the
denominator of the two sides of the implicit equation [1.8].
The reader must nonetheless keep in mind the fictitious
nature of the damage values presented later in the article
(elementary and cumulative damages). In particular, the
inverse of the annual damage values does not in any case
represent the lifespan of the studied structure.
8. Calculate the cumulative damage over the chosen year
with Miner’s law, taking into consideration the traffic
histogram (equation [1.10]):
∑ [1.10]
where n corresponds to the number of time intervals over

a year.
Figure 1.14 presents the weekly increments of fatigue

damage associated with traffic )
calculated for 2006 on the A63 structure, with the following
data:
— traffic distribution histogram from Figure 1.13a;
— weekly temperature data from Figure 1.8b;
— the LCPC-SETRA thermal susceptibility relation
(Figure1.11a).
The temperature shown on the secondary axis

corresponds to criti, at the bottom of the GB3 foundation
layer. These results show that the highest damage is
obtained at high temperatures, according to the calculation
model adopted for the present study.
The summation of the 52 values of damage contained in

this figure leads to a total damage value of 0.04. This value
of damage cumulated over a year is needed for the
calculation of eq.
Figure 1.14. Weekly increments of fatigue damage and temperatures

of the GB3 subbase layer on the A63, relation ε6(θ) according
to the LCPC-SETRA 1994 model
9. Determine the equivalent temperature. According to

the curve of the evolution of annual damage as a function of
temperature (Figure 1.15), the equivalent temperature for
the A63 structure is thus approximately 19.3 C for 2006.
Figure 1.15. Method of determining equivalent temperature with data from the
A63 structure for 2006, using weekly temperatures and the thermal
susceptibility model from the 1994 LCPC guide (LCPC-SETRA 1994)
These nine sequences leading to the calculation of eq for a

given road structure and for temperature values measured
over 1 year are managed automatically by the Alize-LCPC
software. The equivalent temperature is, therefore,
calculated taking into account the design criteria of the
structure under study: fatigue damage of bituminous
materials, hydraulic materials, or cement concrete, and
damage by the permanent strain of granular materials and
of the subgrade.
1.5. Results and discussion
1.5.1. A63 Highway: effects of the time intervals, the law of

thermal susceptibility 6( ), and the calculation year on
equivalent temperature
For the A63 site in Bordeaux, in an H1 climate zone (a

mild winter zone) and for the structure shown in Figure 1.2b,
calculations of θeq were carried out over 3 years (2004, 2006,

and 2008), by calculating incrementally the fatigue damage
of the bituminous material, and considering different values
of time intervals. The objective was to assess the influence of
different calculation parameters on the calculation of θeq,
namely:
— the length of intervals of temperature data: hourly
(Figure 1.16a), daily (Figure 1.16b), weekly (Figure 1.16c), or
monthly (Figure 1.1d);
— the relation of thermal susceptibility 6( eq) (the LCPC-
SETRA law of 1994 or the experimental law proposed by
Bodin et al. [2010]).
T (°) T (°)
55 55
45 45
35 35
25 25
15 15
5 5
-5
-5
01/01/2006 16/06/2006 30/11/2006
1 /1 11 /4 20 /7 28 /10
00:00 16:00 08:00
/2006 /2006 /2006 /2006
(a) (b)
T (°) T (°)
55 55
45 45
35 35
25 25
15 15
5 5
-5 -5
7 /1 27 /5 14 /10 31 /1 30 /4 31 /7 31 /10
/2006 /2006 /2006 /2006 /2006 /2006 /2006
(c) (d)
Figure 1.16. Evolution of temperatures at the A63 structure for 2006 for the
following time parameters: (a) hourly; (b) daily; (c) weekly; and (d) monthly.
Hourly Daily Weekly Monthly
Susceptibility 6( ) Year (8,760 hours/ (365 days/ (53 weeks/ (12 months/
year) year) year) year)
2004 19.1 18.6 18.4 19.5

LCPC-SETRA
2006 19.9 19.5 19.3 20.4
(1994; GB3)
2008 18.5 18 17.9 18.2
2004 19.6 19.3 19.1 19.8

Experimental
(GB3; Bodin et al. 2006 20.8 20.4 20.1 20.2
2010)
2008 19.1 18.7 18.6 18.8
Table 1.11. The calculation of equivalent temperature for different time intervals,
for the 6( ) LCPC-SETRA law of 1994 and the LCPC-Experimental law of 2010,
for the structure and climate of the A63
Seasonal Temperature Changes in the French Pavement Design Method
37
From the hourly measurements, daily, weekly, and

monthly temperatures were calculated by calculating the
average over the chosen time interval. The results, presented
in Table 1.11, show that the four time intervals lead to
relatively low differences in equivalent temperature, which
are lower than 1 C.
A weekly time interval, therefore, seems sufficient for the

calculation of eq. For this climate, the effects of the chosen
thermal susceptibility curve 6( ) and of the chosen years
(2004, 2006, and 2008) are greater (the differences are
around 2 ). The equivalent temperature calculated for the
climate in Bordeaux is on average 19.5°C.
1.5.2. A75 Highway: effects of the type of structure, the

type of climate, the width of vertical discretization of
measured temperatures, and the calculation year
(2004–2012) on equivalent temperature
For the A75 highway, we studied the effects of two

parameters on equivalent temperature:
— The type of road structure (while maintaining the
climate of the A75);
— The vertical discretization of the measured temperature
profiles.
In order to do so, calculations were carried out for nine

consecutive years, from 2004 to 2012, for the thick
bituminous structure of the A63 (Figure 1.2b) and the
inverted structure of the A75 (Figure 1.4a) for the same
harsh climate as found in Saint-Chély-d'Apcher, France. The
thick structure was divided into eight sublayers of 4 cm
each. The inverted structure was also divided into eight
sublayers. The binder layer, 5 cm thick, was divided into two
sublayers of 2.5 cm each, and the base layer of 9.5 cm into
two sublayers of 4.75 cm each. These calculations of
equivalent temperature were carried out with a weekly time
interval, with the relation of thermal susceptibility of

LCPC-SETRA 1994, and with the traffic histogram in
Figure 1.13a.
The results are shown in Figure 1.17. For identical

calculation hypotheses, the equivalent temperature of the
thick bituminous structure of the A63 is approximately 2 C
higher than the one of the inverted structures of the A75.
The values of equivalent temperature are not intrinsic to the
climate of the location studied, but also depend on the type of
structure, as well as to a lesser extent, on the width of the
discretization of vertical temperature profiles. The choice of
the year to be used for temperature data also had a
significant influence (of around 2 C), and thus, it is
preferable to calculate the value of eq over several years, in
order to obtain a truly representative value.
Figure 1.17. Effects of the type of structure (thick bituminous for the A63 and
inverted for the A75), and the vertical subdivision of the structure, for the
climate of Saint-Chély-d'Apcher for the calculation of equivalent temperature,
with the relation 6( ) from LCPC-SETRA 1994. For a color version of this
figure, see www.iste.co.uk/kotronis/risk.zip
The weekly increments of fatigue damage during a

relatively warm year (such as 2006) are shown in Figure
1.18 for the two climates in Bordeaux (A63) and in Saint-
Chély-d’Apcher (A75), for the thick structure
(8BBSG/13GB/13GB/PF3) of the A63 highway. In this
comparison, for the A75, the temperatures applied to the
thick bituminous structure corresponded to the internal
temperature measured in the inverted structure.
Figure 1.18. Increments of weekly fatigue damage, temperatures in the

GB3 subbase on the A63, and temperature in the GB3 subbase for the A75,
with the relation ε6(θ) from LCPC-SETRA 1994…. For a color version of this
figure, see www.iste.co.uk/kotronis/risk.zip
We observed for both climates that the damage rate is

lower during winter and that the maximum value
corresponds to the peak of temperature in summer. This
result is important, as it indicates that structures in warmer
climates are subject to higher rates of fatigue damage than

those in colder climates, provided that all other parameters
are the same, especially the level of rainfall, and winter
severity, in terms of freezing and thawing periods.
For a thick bituminous structure, the climate at the A63

in Bordeaux, which has a θeq value of 19 C, leads to higher
fatigue damage of the bituminous materials than the climate
at Saint-Chély-d’Apcher, at 1,000 m in altitude and with an
equivalent temperature of θeq of 16 C. Indeed, the annual
cumulative damage for the A63 is of 0.04, as opposed to
0.026 for the A75. The ratio between these two values of
annual damage represents the ratio of the lifespans of these
two structures. The impact of the difference in equivalent
temperature shall be evaluated in section 5.4, in terms of
both lifespan and design thickness of the base layer.
1.5.3. SERUL: effects of the daily distribution of traffic and

thermal susceptibility ε6(θ) on equivalent temperature
In this section, equivalent temperature, as understood in

the French pavement design method, is calculated for the
SERUL site, using daily temperature data measured in the
pavement (Figure 1.19). Calculations are carried out again
with Alize-LCPC (research version). For this structure, the
effect of traffic distribution was studied, for the two
distributions shown in Figure 1.13. In Quebec, during
thawing periods, heavyweight vehicles undergo restrictions.
As such, the aggressiveness of traffic, and the equivalent
traffic are lower during this period. This leads to a
non-uniform distribution of yearly traffic. These load
restrictions were not taken into account in this study for the
calculation of eq for the SERUL structure.
42
Risk Evaluation and Climate Change Adaptation of Civil Engineering
Figure 1.19. Evolution of temperature at different depths for the SERUL structure for 2013 and
2014 (Alize-LCPC). For a color version of this figure, see www.iste.co.uk/kotronis/risk.zip
Seasonal variations in horizontal strains (at the bottom of

the base layer in GB-20) and vertical strains (at the top of
the PF3 subgrade), calculated with Alize-LCPC, are shown
in Figure 1.20a. The calculation was done in an incremental
manner, with a daily temperature interval and under the
Quebecois reference load (see Table 1.10).
According to Table 1.5, the value of 6(10 C, 25 Hz) of

GB-20 is 186 μ strains, measured by four-point bending
fatigue tests (Doucet 1999). This value is quite high in
comparison to the usual French materials. The French
model, and particularly its calibration, is based on a two-
point bending fatigue test. Therefore, to be able to use the
French design method, adjustments were made. These
adjustments were done by modifying the value of the
calibration coefficient kc, from 1.3 to 1, considering the very
high value of 6.
Due to the thermosusceptible behavior of the asphalt

layers, strains in the base layer increase with temperature,
due to the parallel reduction of elastic modulus. In Figure
1.20b, the daily increment of fatigue damage calculated with
the relation of thermal susceptibility (Doucet 1999) is plotted
on a logarithmic scale, as it is negligible during winter. It is
more than a thousand times greater in spring, summer, and
fall.
Overall in Quebec, it is recognized that frozen water in

the unbound layers reduces strains and damage in winter,
whereas the increase in water content during thawing
increases strains and damage during spring. Despite the fact
that these variations have not been taken into account here,
the results are consistent with the results from the previous
research (St-Laurent and Prophète 2002; St-Laurent and
Roy 1995).
(a)
(b)
Figure 1.20. The effects of seasonal variations in temperature: a) on strains;

b) on the daily increment of fatigue damage of the base layer for the climate
and structure at SERUL, relation 6( ) according to Doucet (1999). For a color
An example of the calculation of equivalent temperature

is shown in Figure 1.21. In this example, the damage is
calculated with the experimental thermal susceptibility
function 6( eq) for base course bituminous mix (GB-20 with
conventional bitumen) determined by Doucet (1999). The
annual cumulative damage is calculated with the traffic
distribution shown in Figure 1.13a.
The thermal susceptibility function of the GB-20 mix from

Doucet (1999) is defined between 0 C and 17 C (see Figure
1.12). Temperatures during winter are mostly below 10 C
and even 15 C at the bottom of the base layer. The
adjustment of the third-degree polynomial function on these
experimental results can lead to 6 values that are too high
when extrapolated for low temperatures, and similarly for
high summer temperatures, clearly exceeding 20 C.
However, these extreme temperatures being infrequent,
their influence on the value of eq is likely to be relatively
low.
Figure 1.21. Example of equivalent temperature calculation for 2014 at

SERUL, Quebec, with the thermal susceptibility function by Doucet (1999).
The results of calculations carried out for six consecutive

years are shown in Table 1.12 and show equivalent
temperature values between 6 C and 9 C for the thermal
susceptibility functions of (LCPC-SETRA) and from the
experimental results on GB-20 from Doucet (1999). These
results show that the thermal susceptibility function 6( ) is
an important parameter in determining equivalent
temperature.
On the other hand, the daily distribution of traffic had a

negligible effect on the calculation of θeq, for both distribution
hypotheses.
Fatigue Traffic
2011 2012 2013 2014 2015 2016
law dist.
(Doucet
1999), 0—24 hrs
8.7 8.5 8.9 8.4 8.7 8.4
GB 20 100%
conv
(Doucet 6—12 hrs
1999), 35%
8.8 8.5 8.8 8.2 8.8 8.4
GB 20 14—19 hrs
conv 35%
(LCPC-
0—24 hrs
SETRA 6.9 7.1 7.2 6.2 7 5.9
100%
1994)
6—12 hrs
(LCPC-
35%
SETRA 7.1 7.1 7.5 6 7.4 6
14—19 hrs
1994)
35%
Table 1.12. Results of equivalent temperature

calculations for the SERUL site
1.5.4. Consequences for the design of bituminous

pavements
Table 1.13 summarizes the results of the calculations of

equivalent temperature for the three studied climates, and
the different structures selected (two thick bituminous
structures and one inverted structure), and the average
yearly temperatures for the air and at different depths (from
10 cm to 1 m), for the years for which the data are available
(10 years for the A63 and A75 and 6 years for SERUL). For
the different types of structures, the results show that
equivalent temperatures are often different from the value of
15 C commonly used for mainland France, and clearly
higher than air temperatures, and temperatures at different
depths provided by meteorological data.
The thickness of the GB3 base and subbase layers of the

initial structure (8BBSG/13GB/13GB/PF3) of the A63 has
been adjusted, for different values of θeq (from 0 C to 35 C),
in order to study the incidence of equivalent temperature on
the design thickness of the asphalt layers. For this
calculation, an annual average daily traffic of 2,000 heavy
vehicles per day and per lane, with an arithmetic growth
rate of 3%, and an aggressiveness coefficient CAM of 0.8
(according to NF P98-086 Appendix C) have been chosen.
Over 20 years, this corresponds to cumulative heavy vehicle
traffic of 18.8 million vehicles, or cumulative traffic of 15
million French equivalent standard axle loads of 130 kN.
According to the NF P98-086 standard (Appendix E
(informative)), a risk of failure of 1% is used, which is the
value specified for highways with a traffic level Ts (between
2,000 and 5,000 heavy vehicles/day).
Montmorency
Bordeaux, Saint-Chély-d'Apcher,
Climate forest,
France France
Canada
Thick Thick Thick
Type of Inverted
bituminous bituminous bituminous
structure A75
A63 A63 SERUL
Average
annual air 13.6 8 8 1.45
temp. ( C)
Average
annual
temp. at a 16 11 11 5
1 m depth
( C)
Equivalent
temp. θeq 18—20.8 13.5—16.5 16.2—18.4 6—8.9
( C)
Table 1.13. Summary of results for the calculation of θeq for the
three climates studied
For the same values of eq and the same values of design

parameters, the potential lifespans of the A63 structure
(8BBSG/13GB/13GB/PF3) were also calculated.
The results are shown in Table 1.14 and Figure 1.22. The
increase in θeq leads to a significant reduction in lifespan (for
the same layer thickness), or else an increase in the
thickness of the base layer, in comparison to the thickness at
15 C, for the same lifespan. To consider realistic
temperatures in pavement design appears essential, as, for
the thick bituminous pavement studied, an increase of 1 C of
eq leads to a reduction of 8% of the lifespan of the pavement
or an increase of 0.25—0.5 cm of the GB3 thickness to keep
the same lifespan.
Equivalent
temperature 0 10 15 18 20 21 25 30 35
( C)
Lifespan
51 32 24 18 15 14 10 6 5
(years)
Thickness of 22 24 25 27 27.5 28 30 33 35
GB3 (base +
subbase) (cm)
for 20 years 4 2 1 +1 +1.5 +2 +4 +7 +9
Table 1.14. Influence of temperature on design
Figure 1.22. Effect of temperature on the thickness and lifespan of the thick
bituminous structure of the A63. For a color version of this figure, see
1.6. Conclusion
The objective of this study was to better understand and

predict the effect of seasonal temperature variations on the
calculation of fatigue damage for three sites with very
different climates. A comprehensive methodology was used
in this research, comprising: 1) the acquisition, exploitation,
and analysis of temperature datasets and 2) detailed
comparative modeling, which aimed to evaluate the
influence of different calculation parameters on the value
of eq.
This study thus demonstrated that temperature is one of

the main climatic factors affecting road performance, along
with moisture, and freeze/thaw cycles. To assess the
durability of the structure, it is essential to consider the
realistic seasonal temperature variations in the calculation
of equivalent temperature. Indeed, many countries, such as
Canada and some French regions, are exposed to large
seasonal temperature variations. Due to the thermal
susceptibility of bituminous material, warm periods will
accentuate the development of fatigue damage in pavement
structures.
The study has also helped evaluate the effects of different

hypotheses on the calculation of the equivalent temperature.
We have thus shown that the value of equivalent
temperature depends strongly on:
— The period over which the equivalent temperature is
calculated (one or several years), several years being
preferable in order to obtain representative values;
— The law of evolution of fatigue parameters with
temperature.
However, the equivalent temperature appears to be less

affected by the length of the time intervals chosen for
calculation (weekly intervals appear sufficient) and by the

distribution of daily traffic.
The current French pavement design method does not

take into consideration the project’s local climate data. Only
a constant “equivalent temperature” eq (15 C in mainland
France) for the entire year is used in the calculation. This
study proposes a new procedure, based on an incremental
calculation of eq, which allows us to take into account the
more realistic seasonal temperature variations, specific to
the studied project.
In France, the analysis for two locations with different

climates has shown that equivalent temperatures can vary
significantly (between approximately 15 C and 20 C). This
analysis could now be extended to the entire country. For the
location in Quebec, a value of eq of about 7 C was obtained.
It would also be beneficial to better take into account the
effect of freeze/thaw cycles on the stiffness of unbound
materials in design calculations.
To conclude, therefore, we suggest improving the accuracy

of the French mechanical-empirical design method
concerning climate parameters to better anticipate the
performance of road structures and to optimize their design.
With regard to future research, it would be interesting to:

— increase the number of on-site measurements of
meteorological data (air temperature, nebulosity, sunshine,
wind speed, hygrometry, etc.), internal temperature data, as
well as measurements of the moisture content of the
subgrade and strains in the bituminous layer. Such data
would be useful not only for mainland France but also for the
overseas territories, where temperature conditions can be
very different;
— develop a thermal model which could calculate, from
meteorological data, the internal and surface temperature
values for all types of road structures. This could facilitate

the use of project-specific temperature values for design;
— establish correlations (most probably experimental) for
the French context between the project’s climatic
characteristics, the types of structures, and the moisture
conditions in treated and untreated materials, and in the
subgrade, during the lifespan of the road, in order to better
consider moisture conditions in pavement design;
— better take into account layer interfaces in the design
method, by developing models describing the behavior of
pavement layer interfaces in relation to temperature.
1.7. Acknowledgments
The authors would like to thank:

— Météo France for providing climate data;
— Prof. Guy Doré from the Université Laval, a kind host
who provided the data on the SERUL structure;
— Mr. Jean-Maurice Balay for diligent proofreading and
whose constructive discussions improved the overall content
of the paper;
— The Pays de Loire region for financing the RI-
ADAPTCLIM project.
1.8. References
AIPCR (2014). Impacts des changements dans la rigueur de l’hiver

sur la viabilité hivernale. Technical report B5. Association
mondiale de la route (AIPCR).
Badiane, M. (2016). Effet des charges sur les chaussées en période
de restriction des charges-volet terrain. Master’s thesis, Laval
University, Quebec, Canada.
Balay, J.M., Brosseaud, Y., Bara, B., and Castaneda, E. (2012).

Adaptation of the French Pavement Design to Countries in
South America. Congrès 8ème Jornadas International Des
Asfalto, France.
Bodin, D., Terrier, J.-P., Perroteau, C., Hornych, P., and Marsac, P.
(2010). Effect of Temperature on Fatigue Performances of
Asphalt Mixes. 11th International Conference on Asphalt
Pavements, Nagoya, Japan.
Brown, P.T. and Caldeira, K. (2017). Greater Future Global
Warming Inferred from Earth’s Recent Energy Budget. Nature,
552, 45—50. Available: https://doi.org/10.1038/nature24672.
Burmister, D.M. (1945). The General Theory of Stresses and
Displacements in Layered Soil Systems. II. Journal of Applied
Physics, 16, 126—127. Available: https://doi.org/10.1063/
1.1707562.
Corté, J.-F. and Goux, M.-T. (1996). Design of Pavement
Structures: The French Technical Guide. Transportation
Research Record Journal of the Transportation Research Board,
1539, 116—124. Available: https://doi.org/10.3141/1539-16.
Doré, G., Bilodeau, J.-P., Thiam, P.M., and Drolet, F.P. (2014).
Impact des changements climatiques sur les chaussées des
réseaux routiers Québécois. Thesis, Laval University, Quebec,
Canada.
Doucet, F. (1999). Caractérisation de la rigidité et de la résistance
à la fatigue des enrobés. Thesis, Université de Sherbrooke,
Canada.
Doucet, F. and Auger, B. (2010). Détermination du module
complexe des enrobés au ministère des Transports du Québec.
Report, Service des matériaux d’infrastructures, Direction du
laboratoire des chaussées, Ministère des Transports du Québec,
Canada.
Goacolou, H., Piau, J.M., Balay, J.M., Odeon, H., Petitgrand, J.C.,
Kobisch, R., Paillard, M., Freneat, E., and Lerat, P. (2003). La
méthode française de dimensionnement. Revue Générale des
Routes et de l'Aménagement, 823, 48—62.
Hammoum, F., Mauduit, C., Bouilloud, L., and Mathon, D. (2016).

Conditions climatiques: Impact sur les infrastructures
routières. Revue Générale des Routes et de l'Aménagement, 933,
36—39.
Hornych, P., Balay, J.M., Mauduit, C., and Bodin, D. (2013).
Evaluation of the Concept of Equivalent Temperature for
Pavement Design. 9th International Conference on the Bearing
Capacity of Roads, Railways and Airfields (BCRRA 2013),
Trondheim, Norway.
itech (2017). Alize-LCPC — The Professional Software for Design of
Pavement Structures. Available: http://www.alize-lcpc.com/fr/
[Accessed 24 September 2018].
Joly, D., Brossard, T., Cardot, H., Cavailhes, J., Hilal, M., and
Wavresky, P. (2010). Les types de climats en France, une
construction spatiale. Cybergeo: European Journal of
Geography. Available: https://doi.org/10.4000/cybergeo.23155.
Ktari, R., Hornych, P., Hammoum, F., Marsac, P., Nguyen, M.L.,
and Piau, J.M. (2017). Dépendance thermique des performances
en fatigue des enrobés bitumineux. 35èmes Rencontres de
l’AUGC, ECN/UN, Nantes, France.
Ktari, R., Hornych, P., St-Laurent, D., Hammoum, F., and Marsac,
P. (2018). Effect of the Temperature and Climatic Conditions on
the Asphalt Pavement Performances by the Mechanistic-
Empirical Methods. Building and Environment.
LCPC-SETRA (1994). Conception et dimensionnement des
structures de chaussée. Report, Ministère de l’Équipement des
Transports et du Tourisme, Paris.
LCPC-SETRA (1998). Catalogue des structures types de chaussées
neuves. Report, Ministère de l’Equipement, des Transports et
du Logement, Paris.
Mauduit, C. and Florent, L. (2015). Mise en oeuvre d’un site
instrumenté sur une section autoroutière de l’A75 -
Configuration du site et matériel utilisé. CCLEAR seminar,
IFSTTAR, February 3.
Mauduit, C., Livet, J., Martin, E., Duval, R., and Dupont, O. (2007).
Impacts of Climate Change on Frost Design and Winter
Maintenance Activities in France. 8th International Symposium on
Cold Region Development, Tampere, Finland, 25—27 September.
Mauduit, V., Mauduit, C., Greullet, N.V., Coulon, N., Hammoum,
F., David, H., Kerzreho, J.P., Piau, J.-M., and Chabot, A. (2013).
Dégradation subite des enrobés bitumineux par période de
gel/dégel: Analyse de cas de terrain et recherche exploratoire en
laboratoire. Bull Liaison Lab Ponts Chaussées, 279, 47—63.
Météo France (2017). Bilan climatique de l’été 2017. Available:
http://www.meteofrance.fr/climat-passe-et-futur/bilans-climatiq
ues/bilan-2017/bilan-climatique-de-l-ete-2017 [Accessed 16 April
2018].
Miner, M.A. (1945). Cumulative Damage in Fatigue. Journal of
Applied Mechanics, 12, 159—164.
MTQ (2016). Fiche 4 : L’effet du climat sur les chaussées - La
protection du réseau routier: une priorité. Report, Transports
Québec.
Rouquès, G. and Caniaro, L. (1975). Gel et dégel des chaussées.
Technical note, Laboratoire central des ponts et chaussées.
Sohm, J., Gabet, T., Hornych, P., and Piau., J.-M. (2011). Essais de
fluage confinés et thermo-régulés sur enrobés bitumineux:
expérimentation et modélisation. 20ème Congrès Français de
Mécanique, France.
St-Laurent, D. and Prophète, F. (2002). Évolution mécaniste de
l’impact des restrictions de charge en période de dégel sur les
chaussées souples. Congrès annuel de 2002 de l’association des
transports du Canada, Winnipeg, Canada.
St-Laurent, D. and Roy, M. (1995). Evaluation structurale de
chaussées souples dans un contexte climatique nordique.
Master’s thesis, Laval University, Quebec.
Vu, V.T., Chupin, O., Piau, J.-M., Hammoum, F., and Bouron, S.
(2018). Experimental study and modeling of the behavior of
partially saturated asphalt concrete under freezing condition.
Construction and Building Materials, 163, 169—178; doi:
10.1016/j.conbuildmat.2017.12.070

Consideration of Seasonal Temperature Changes in The French Pavement Design Method

Uploaded by

Copyright:

Available Formats

Consideration of Seasonal Temperature Changes in The French Pavement Design Method

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Consideration of Seasonal Temperature Changes in The French Pavement Design Method

Uploaded by

Copyright:

Available Formats

1

The current French pavement design method assumes a

Chapter written by Rahma KTARI, Ferhat HAMMOUM, Pierre

have focused on three issues. The first was the effects of

Over the course of the last few decades, global warming

A recent review of climate models has allowed

1 Observatoire national sur les effets du réchauffement climatique,

Furthermore, many countries are confronted with

In practice, the designing of a road involves many

Road structures consist of multiple layers (LCPC-SETRA

Variations in temperature, whether daily or seasonally,

In France, the LCPC3 (currently IFSTTAR4) developed a

Traffic conditions are assessed by the load aggressiveness,

3 Laboratoire central des ponts et chaussées.

In this chapter, we consider that the hydric state stays

Three research sites, with different climates, were

The second site is located in Saint-Chély-d'Apcher,

The objective of the study is to assess the effect of internal

This chapter has five sections. The first section introduces

namely, the structural calculation method, as well as the

1.2. The experimental road structures

Based on the cartography of the average temperatures in

The third site is the Site Expérimental Routier de

extremely cold temperatures in winter and hot temperatures

Table 1.1 shows the comparison the French and Quebecois

The chosen sites are exposed to sizeable differences in

Table 1.1. State-managed road networks: a comparison between

1.2.1. The A63 highway (Bordeaux, France)

The first structure studied in France is located on the A63

It is composed of three asphalt layers: a semi-coarse

Surface layer in BBSG (8 cm)

The standard design parameters for this structure are

Table 1.2. Standard design parameters of the materials according

6 is the strain leading to fatigue failure for 1 million loads

Hourly temperature variations, as a function of depth, are

Temp. ( C) Tair Tsurf T10 cm T20 cm T30 cm T50 cm T70 cm T1 m

Table 1.3. Recordings of extreme values of hourly temperatures at the

Figure 1.3. Hourly temperature changes for different depths in the

1.2.2. The A75 highway (Saint-Chély-d'Apcher, France)

The structure is located at an altitude of about 1,000 m, in

Surface layer BBMa 0/10 (4 cm)

It is a so-called “inverted structure” composed of a surface

layer of unbound granular material (UGM), and a foundation

These layers are placed on a layer, with a thickness of

Hourly temperature variations as a function of depth are

Table 1.4. Recordings of extreme values of hourly temperatures at the

Figure 1.5. Hourly temperature evolution for different depths in the

1.2.3. Laval University Experimental Road Site (SERUL)

The third pavement structure is situated at the

This thick bituminous structure (see Figure 1.6b) has the

passing at 80 μm, respectively) and finally a subgrade

Bituminous concrete (20 cm)

Figure 1.6. (a) Location of SERUL. (b) Structure of the

Table 1.5. Standard design parameters of the materials according to the NF

The results of complex modulus tests on the bituminous

Type of layer Modulus (MPa)

Table 1.6. Elastic moduli of unbound materials

A road weather station with a 3 m long probe (18

An example of the temperature data for 2012, for seven