Professional Documents
Culture Documents
Vehicles
- an overview
Report nr. 08/2008
NVF committee Vehicles and Transports
Frfattare:
Titel:
Serie:
Upplaga:
Utgivningsort:
Mattias Hjort
Mattias Haraldsson
Jan M. Jansen
Road Wear from Heavy Vehicles an overview
NVF-rapporter
100
Borlnge, Sverige
Tryck:
Vgverket 2008
ISSN:
0347-2485
NVF-rapporterna kan bestllas via respektive lands sekretariat per telefon, fax,
e-post eller post. Se kontaktuppgifterna p nst sista sidan.
En uppdaterad rapportfrteckning finns p frbundets nordiska hemsida,
http://www.nvfnorden.org.
Sammanfattning
Yhteenveto
Summary
This report is an overview on how different studies of road wear have reported the effect of
various vehicle dependent parameters, for heavy vehicles, such as axle load, tyre properties etc.
The intention with this overview has only been to collect the different results, without any kind of
evaluation. The overview is mainly based on Cebons Handbook of Vehicle-Road Interaction, the
DIVINE project, and COST 334.
Table of contents
1. Introduction
10
12
2.1 Introduction
12
12
13
15
16
3. Vehicle Effects
17
3.1 Introduction
17
17
19
3.4 Effects of different axles: single axles, tandem axles and tri-axles 19
3.5 Tyre specific effects
22
25
26
4.1 Overview
26
27
27
33
36
36
39
6. Acknowledgements
43
7. References
44
1 Introduction
This report within NVF Fordon och Transporter is intended as an overview on how different
studies of road wear have reported the effect of various vehicle dependent parameters, such as
axle load, tyre properties etc. The intention with this overview is only to collect the different
results, without any kind of evaluation. Chapter 2 gives a short introduction to pavement
construction, and the various modes of damage that occurs. Chapter 3 describes the vehicle
dependent parameters and the models that are normally used for predicting vehicle induced road
wear. The range of reported values in the literature of these vehicle parameters are given in
chapter 4. Two examples of calculation models used in Denmark and Sweden are described in
some detail in chapter 5.
Some major works have been conducted in recent years, and form the basis of this report. The
primary sources are
DIVINE (Dynamic Interaction between Vehicle and INfrastructure Experiment) [2], OECD
project 1992-1998. An experimental study focusing on how dynamic loads are affecting
road deterioration.
COST 334 Effects of Wide Single tyres and dual Tyres [3] (1996-2000). Full scale
tests were carried out, in addition to literature studies, in order to describe all tyre
parameters that affect the road wear, and how their combined effect can be determined.
A general description of the field of Interaction between vehicle/climate and the road, including a
discussion on the further need of research, has recently been written in Swedish. :
That report has also been used as a basis for this report.
.
The present increase in deployment of mechanistic design, especially in USA, of course
enhances the world wide research and development within the issue of this report. Some very
interesting results have been presented recently from USA, but they have not been discussed in
context of this report. These new results are described very briefly below.
From implementation Mechanistic-empirical design in USA, the 3 major approaches about axle
load versus road wear are:
1. Development of a load damaging factor.
The dynamic effect of road unevenness adds a factor to the road wear calculated from static
axle loads. This effect is caused by an interaction between the dynamic of the vehicle and the
wave length and amplitude of the road surface roughness, and depending on the travel
speed. Thereby the damaging factor results, such that the effect on road wear will relate to
class of road, i.e. different effect on motorways and minor roads as well as rural and urban
roads.
2. Effect of axle loads, repetitions and spacing on dissipated creep strain energy, DCSE, in
relation to development of cracks in pavement materials.
10
11
12
Rigid pavements are mostly found in major highways and airports, and like flexible pavements
they are designed as all-weather, long lasting structures to serve modern day high speed traffic.
The rigid pavement has a top layer consisting of a concrete slab of thickness of 10-30 cm. The
load transmission mechanisms of these two pavement types are different. The rigid pavement
distributes the load over a large area, while flexible pavements, when the traffic load is applied
on top of the surface layer, a localized deformation occurs under the load, as shown in Fig. 2.1.
The load is distributed over a small area at the surface, but as the depth increases, the same
load is distributed over a larger area. Therefore, the highest stress occurs at the surface and the
stress decreases as the depth increases. Thus, the highest quality material needs to be at the
surface, and as the depth increases, lower quality materials can be used. When the load is
removed the pavement layers rebound. A very small amount of deformation, however, could
stay permanently which could accumulate over many load repetitions causing rutting in the
wheel path. [7]. The design life of a flexible pavement may be in the range of 15-20 years, while
it is common for a concrete pavement to be designed with a service life of 30-40 years.
fatigue cracking: being cracking in the bituminous or cement bound material originating
at the bottom of the respective layers, due to fatigue of the material by a great number of
repetitions of bending due to wheel loads (Fatigue defined in this way is used as a
parameter in pavement design. This does not include surface cracking and cracking due
to thermal cycling, although these are also due to fatigue because of repeated stress
cycles.)
thermal cracking: being cracking in the bituminous material due to tensile stresses
caused by temperature changes
surface cracking: being cracking in the bituminous material originating at the surface of
the pavement, due to fatigue of the material by a great number of shear loadings of the
pavement surface by the tyre (Ageing of bituminous materials plays an important role
here, too.)
reflective cracking: being cracking of the (top) bituminous layers (often in a composite
structure) as a result of cracks or joints in bound layers below.
Rutting
Rutting is the development of depressions in the pavement surface along the wheel paths,
typically with a width of several decimetres and a length of tens to thousands of meters. The
three main categories are
13
Ravelling: being the loss of stones in the surface of the pavement as a result of failure of
the bond between the aggregate and the binder by a great number of shear loadings in
combination with ageing of the material. The primary case, is however insufficient quality
of the pavement material.
Potholes: resulting either from local collapse due to structural defects, or from frost
acting on water ingress (often through cracks). Potholes are not necessarily caused by
loading but mainly due to insufficient quality of the pavement
The more important of these distress modes are shown in Figure 2.2.
Figure 2.2. Various modes of pavement distress. (from Ref. [3]). For explanation on
nomenclature, see Sec.2.2. HMA is short for Hot Mix Asphalt.)
The main distress modes, especially from the point-of-view of traffic loading are:
14
1. Fatigue cracking. This occurs mainly on relatively weak / thin pavements (Visible
cracking in thick pavements is likely to originate (at least partly) at the surface.)
2. Primary rutting. This occurs mostly on main roads with thick bituminous layers.
3. Secondary rutting. This occurs mainly on relatively weak / thin pavements.
It is clear that these different distress modes react differently to changes in the influencing
factors. Take e.g. the influence of tyre type, where the stress and strain conditions near the
surface of the pavement are strongly influenced by the contact stresses and their distribution in
the tyre-pavement interface, whereas the stresses and strains deeper in the structure are mainly
influenced by the total load. Therefore, a change in contact stress distribution due to a change in
tyre type can generally have most influence on the upper layers2.
Most design methods for flexible pavements since the 1960s are based on the prevention of
fatigue cracking and secondary rutting. In relatively weak /thin pavements this does not always
succeed, but in thick pavements it generally succeeds. Then, primary rutting may become the
dominant distress.
Permanent deformation of bituminous layers is usually not considered as a part of structural
design. With proper bituminous mixture design the tendency for permanent deformation can be
decreased. However, the bituminous mixture design is always a compromise between many
properties (including price) and small changes in mixture composition during the manufacturing
may worsen the properties of bituminous mixtures. It is possible to manufacture mixtures, which
will not deform easily, by using modified bitumens but they are much more expensive and thus
their use is limited.
Ravelling and surface cracking, being the most superficial distress modes, may be influenced by
any differences in contact stress distributions between different tyre types.
However, exceptions to this rule may occur, depending on structure and material quality (e.g. a very
critical stress-sensitive granular layer below a rather thick AC layer may be the main cause of rutting
increase due to a slight increase in stresses).
15
16
3 Vehicle effects
3.1 Introduction
This section describes the influence of properties of different vehicle components on exposed
road wear. It is the intention of this chapter to explain the concepts, and to briefly describe what
kind of road damage the different factors leads to. Experimental findings concerning these factor
are reported in Chapter 4.
The basic parameter is the axle load. The effect of different axle load sizes is described in
section 3.2, explaining the concept of Load Equivalency Factors. It is important to realise that the
actual forces on the road are not equal to the static axle loads, but vary because of vehicle
dynamics. This is the topic of section 3.3, where dynamic loads are described. Also, the effect of
axle loads may be influenced by neighbouring axles, as explained in section 3.4. Another factor
to be taken into account is the lateral wander of the traffic, elaborated in section 3.5. In section
3.6, the load sharing between twinned tyres is discussed. The influence of different tyres are
described in section 3.7, introducing the concept of super single tyres. Finally, different kind of
suspensions are described in section 3.8.
N ref
Nx
W
= x
W
ref
(3.1)
where Wx and Wref are axle loads and Nx and Nref are the corresponding number of load
applications.
The exponent 4 in the fourth power law was found in the AASHO Road Test, carried out in USA
between 1958-1960.. However, it was not strictly constant in that test but varied from about 3.6 to
4.6. Later experimental and theoretical research has indicated greater variability in the exponent,
but has not been conclusive. As an example, it was found in the OECD FORCE project that the
exponent depends also on the extent of distress, the exponent being smaller in earlier phases
than in later phases of failure.
It must be understood that the fourth power law includes all distress modes. The most important
at the AASHO road test were rutting (caused by subgrade deformation) and roughness
(unevenness) of the road. Cracking had a minor effect and deformation of bituminous mixtures
was not important.
When individual distress modes are considered, different exponent values are found. E.g. COST
334 reports that cracking of bituminous layers has a value of 4 7, permanent deformation of the
subgrade has an exponent of perhaps 3 4 and permanent deformation of bituminous layers a
value of 1 2. As these values depend on many factors (a.o. material variations) and are not
fully known, the stated values should be regarded as best estimates [3].
17
For use in pavement design, where the actual spectrum of axle loads has to be converted to an
equivalent total number of standard axle loads, COST 334 concludes that the precise exponent
value is not very important. For exponent values between 2 and 6, most actual axle load spectra
were found to translate to roughly the same equivalent number of standard axles. (For low
exponents, the multitude of smaller axle loads contribute to the bulk to the total equivalent
number. For high exponents, the few overloaded axles contribute the most.) It was thus
concluded that the overall value of 4 is well suited [3].
For detailed studies into pavement wear effects the exponent values for the individual distress
modes should be distinguished. This is especially the case when conclusions should be drawn
from accelerated tests at high load values.
Besides being used within road design, the fourth power law has also been used for evaluating
the road damaging potential of abnormal vehicles. E.g., in Denmark, a formula based on the
fourth power law is currently used for classifying the road damage of Heavy Abnormal
Transports, and is the base for taking decisions on allowed routes for heavy transports. This
formula is described in some detail in section 5.1.
However, according to Cebon [1], the validity of the fourth power law is questionable,
particularly for current axle loads and axle group configurations; tyre sizes and pressures; road
construction; and traffic volumes: all of which are significantly different from the conditions of the
AASHO road test.
The Divine project comes to a similar conclusion, and write that the use of the fourth power law
may not be appropriate in all situations unless the environment, traffic, pavement type and
pavement construction methods are the same as, or very similar to, those in the AASHO Road
Test [2].
Note that the fourth power law was derived from measurements on heavy vehicles. To apply the
law also on light vehicles, such as passenger cars, would imply a vast extrapolation outside the
range of vehicle loads used in the AASHO experiments. The Swedish Road Administration
advice against comparisons between heavy vehicles and passenger cars, with respect to the
concept of an equivalent number of vehicles [8].
An example on how to apply the fourth power law on measured axle loads is taken from
Denmark. From weighing of actual axle loads at weighing stations or weigh-in-motion bridges,
the load equivalent of vehicle classes can be obtained for ordinary traffic using the fourth power
law. Results from these measurements, compared with the ESAL resulting from maximum
allowed weights of these vehicles, are shown in table below:
Table 3.1: Load equivalent of vehicle classes obtained from weighing the actual axle loads
Road wear in 10 T ESAL by
Typical permissible Gross
Typical Vehicle classes
4th-power law
weight
From
From max.
measured
load
loads
2-axle lorry
0.3
1,4
18000 kg
Lorry with trailer
1,1
3,4
38000 kg
Semitrailer
0.9
2,6
42000 kg
Busses
0,4
1,4
18000 kg
For more details regarding this study, we refer to Jan M. Jansen, Vejdirektoratet, Denmark.
18
5 10% for well-damped air suspensions and soft, well-damped steel leaf suspensions.
Dynamic loading increases pavement wear. Because of the power-law dependency of pavement
distress on axle loads, the loads above the static load increase the pavement wear more than the
decrease in wear due to the loads below the static load.
Besides load magnitude, also frequency content is important for pavement wear. Most heavy
vehicles have dynamic wheel loads either in the 1.5 4 Hz range, associated with bounce
(up/down) and pitch (rotating forward/backward) motions of the vehicle body, or in the 8 15 Hz
range, associated with axle-hop vibration. Axle hop vibrations are more significant if the
pavement is rough and the vehicle speed is higher than approximately 40 km/h.
As stated before, the tyre characteristics (vertical spring compliance and damping) influence the
dynamic vehicle loads. Therefore, these should be considered when establishing pavement wear
effects of different tyres.
3.4 Effects of different axles: single axles, tandem axles and tri-axles
According to COST 334, tandem axles and tri-axles (see the definitions in Table 3.1) generally
cannot be treated by summation of the effects of their constituting individual axles, because of
two reasons:
The load spreading of thick pavements may be such that the responses (stresses and
strains) due to neighbouring axles in a tandem or tri-axle configuration may substantially
increase the responses under the axle considered. Due to the non-linearity of the
performance relations, such increased responses will lead to much more pavement wear
than the summed responses of individual axles.
Due to the visco-elastic nature of bituminous materials, stresses and strains caused by an
axle load need some time to relax after the axle has passed. When another axle arrives
within that period, some residual stresses and strain will still be present, which may
19
compound with the stresses and strains caused by the new axle, resulting in higher total
values. The effects of this mechanism are not well understood.
For axle load limitations, this is reflected in maximum allowed tandem axle (and tri-axle) loads
which are less than twice (or three times) the allowed single axle load. (Two axles at more than
1.8 m spacing are not considered a tandem axle but a double axle and are treated as two
single axles.)
Cebon concludes the opposite, and writes that It is generally concluded that for equal damage to
flexible and rigid pavements, tandem and triaxle groups can carry more weight than the same
number of widely spaced single axles, because the primary response fields of nearby axles
overlap.
The complicated relation between inter-axle spacing and relative road damage is illustrated in
Figure 3.1, taken from Cebon, where the effect of tandem axle group spacing on theoretical
fatigue damage in a rigid pavement is shown. The figure shows that there is an optimum axle
spacing, and using either a wider or more narrow axle spacing will result in increased road
damage. Moreover, this optimum axle spacing depends to a large extent on the thickness of the
concrete slab.
Figure 3.1: The effect of tandem axle group spacing on theoretical fatigue damage in a rigid
pavement (From Cebon, p.301)
For pavement design purposes, however, the loads of tandem axles and tri-axles are mostly
converted to a number of equivalent standard axle loads (Nesal) by summing the contributions of
the individual axles [3]. These individual contributions are then calculated using the Load
Equivalency Factor described in Section 3.2., resulting in:
N esal =
nr of axles
20
Waxle
Wstandard axle
(3.2)
a single axle with more than 1.8 m spacing from other axles.
Double axle
a configuration of two axles, with more than 1.8 m spacing. (This is not a
tandem axle, and the individual axles of a double axle are considered
separately.)
Tandem axle
a configuration of two axles, with less than 1.8 m spacing between the axles.
(Often the suspension of a tandem axle is such that the load on the tandem
axle is shared rather equally between the constituent axles.) The maximum
load is dependent on axle spacing and suspension, and is different for motor
vehicles or for trailers and semitrailers. (The EC also distinguishes a bogie,
being two axles with shared suspension and less than 1.3 m spacing).
Tri axle
As an example on how different axle configurations are used in regulations, the maximum
allowed loads in Sweden are shown in the table below:
Table 3.3: The maximum allowed loads (in metric tons) in Sweden for different axle
configurations and road classifications (BK1, BK2 and BK3). From [4].
BK 1
BK 2
BK3
10
10
b) Driven axle
11,5
10
11,5
11,5
11,5
16
16
12
18
16
12
19
16
12
20
16
12
21
20
13
24
22
13
1. Axle load
and the driven axle has dual tyre fitment and air suspension, or
equal suspension, or if the driven axles has dual tyre fitment
and the load on each axle does not surpass 9,5 tons
e) Inter-axle distance 1,8 m
3. Tri axle load
21
Figure 3.2: Wide-base Single Tire (445/50R22.5) compared to standard dual tyres
(275/80R22.5). (Picture from Ref. [10], with permission from the author)
However, although wide base single tyres might increase road damage through its smaller
footprint and higher inflation pressure, it improves roll-over stability of the vehicle. This in turn has
consequences for suspension design. As is discussed below, suspension stiffness is necessary
for roll-over stability but increases road damage. It is therefore at least a theoretical possibility
that the increased road damage induced by the change from dual tyres to super single tyres can
be compensated for by reducing the spring stiffness. Also, the damping needed to reach the
22
minimum road damage is about 25 percent lower for wide single tyres than for dual tyres, mainly
because of a lower unsprung mass. This improves the possibility to design an optimal
suspension. (Cebon p. 453).
Another important aspect to take into account when comparing single and dual tyre assemblies is
the concept of unequal load sharing of the dual tyres. When comparing dual and single tyre
assemblies at equal wheel load, generally the assumption is made that the wheel load is shared
equally between both tyres of the dual assembly. However, in practice this might not be true. A
number of reasons could cause an unequal load division (load imbalance) between both tyres:
1. differences in vertical stiffness between both tyres, because of
Figure 3.3: Causes unequal load sharing between tyres in a dual tyre assembly (load
imbalance). From Ref. [3].
Axle- and tyre configuration also affect road damage in an indirect way, an effect that is denoted
lateral wander.
In practice, not all wheels will pass at the same lateral position in a road section. Vehicles
generally follow a slightly zigzagging course between the bounds of the traffic lane, which is
called lateral wander. Therefore, the wheel positions of consecutive vehicles will be transversely
distributed over the pavement.
23
Detailed measurements and analysis of this distribution are reported by Blab [5]. He showed that
the probability distribution of the vehicle positions is a Laplace distribution, in stead of the normal
distribution that is often assumed. For a certain vehicle width and lateral wheel spacing, the
probability distribution of the wheel (centre) positions is a Laplace distribution, too. However, the
number of hits by a tyre per cm pavement width is approximately normally distributed, due to the
summation over various vehicle widths, wheel spacings, dual and single wheels, and various tyre
widths. The difference is shown in Figure 3.4.
(a)
(b)
Figure 3.4: The difference between probability distribution of the wheel positions (a), and the
number of hits by a tyre per cm pavement width (b). From Ref. [5].
Lateral wander distributes pavement loading, and hence pavement wear, over a larger area of
the pavement. This prolongs the pavement service life. The effects of lateral wander are different
for the different distress modes. They also may differ between dual tyres and wide base singles.
COST 334 reported that the distress reduction factor of lateral wander on primary rutting
compared to non-wandering loading were in the range 0.67 0.87, using different road structures
and a range of tyres. Generally, the effects of lateral wander for the dual tyres that were tested
were all very similar and close to those of a 385 mm wide single tyre. The effect of lateral wander
for a 495 mm wide single tyre were smaller (about 30%) than for the other tyres. They also
concluded that the beneficial effects of lateral wander for all tyres increase with decreasing
pavement thickness.
24
25
Figure 4.1: Summary of literature on the effects of various vehicle features on road damage.
From Ref. [1].
From the chart in Fig. 4.1, Cebon draws the following conclusions:
1. Applying a tandem suspension load to a single axle can be expected to increase road
damage by a factor of up to 25 (first bar of chart)3.
2. Replacing dual tyres with wide-base single tyres may increase road damage by a factor of
up to 10 (second bar of chart).
3
Cebon points out that this is not a particularly realistic scenario, but the included it for comparison purposes.
26
3. Unequal static load sharing between axles in a tandem suspension may increase road
damage by a factor of up to 3 (third part of chart).
4. The fourth bar summarises the literature on the road-damaging effects of dynamic tyre
forces. The average increase in damage caused by dynamic forces, compared to static
forces alone is approximately 10% 40% (mean damage on the fourth bar of the chart,
as calculated by the road stress factor and/or by neglecting spatial repeatability). This is
small compared with the effects of tyre type and unequal static load sharing shown in the
second and third bars. If instead a high degree of spatial repeatability is assumed, the
relative increase in peak road damage caused by dynamic forces is in the range 2 14
(peak damage in the fourth bar of the chart), which is comparable with the effects of
tyres and unequal static load sharing.
The rest of this chapter compares results from the main references used in this report, with
respect to vehicle influence on road deterioration.
N ref
Nx
W
= x
W
ref
(4.1)
where Wx and Wref are axle loads and Nx and Nref are the corresponding number of load
applications.
Cebon concludes that for flexible pavements, values of 1.3 6 has been reported in the
literature, while for composite and rigid pavements, values are thought to be as high as 8 33.
COST 334 makes a distinction between individual distress modes, and reports that cracking of
bituminous layers has a value of 4 7, permanent deformation of the subgrade has an exponent
of perhaps 3 4 and permanent deformation of bituminous layers a value of 1 2.
Primary roads: Defined as the network of principal roads in a country or state, generally
comprising motorways (autoroutes, autostrade, etc) and other principal roads, state
owned or otherwise. This network provides the major links between large urban areas
and key national long-distance routes.
27
The road damage from a specific tyre and assembly is described with respect to a reference tyre
in terms of Tyre Configuration Factor (TCF). This concept is introduced in that project, and the
interpretation is straightforward: if a tyre has a TCF value equal to three, that means that for the
same axle load, this tyre is three times as aggressive as the reference tyre. Different formulas for
the TCF were derived for different pavement thicknesses and different distress modes. A tyre
specific part taking the tread width, contact length and the tyre pressure ratio (with respect to the
recommended pressure) into account is combined with a factor taking the tyre fitment into
account. This last factor comprises factors for tyre characteristics regarding dynamic force
transmissibility, and for potential load imbalance between the tyres of a dual assembly. The TCF
formulae are shown below in table 4.1.
Table 4.1: The TCF formulae from COST 334.
Total factor for translation to
real world conditions
Pavement Tyre specifications
thickness
Dual
tyres
Wide base
single tyres
Single
tyre
1.01
0.97
1.00
Thick
0.99
1.00
0.97
0.97
1.00
Medium
0.98
0.97
1.00
Thick
about equal to 1
0.94
0.97
1.00
Medium
0.96
0.97
1.00
Thick
about equal to 1
Below TCF values for different tyres and assemblies reported in COST 334 are presented, using
a reference tyre of width 235 mm put in dual fitment. In Fig. 4.2 4.4 the graphical overviews of
the TCF values of the individual tyres are presented. Dual tyres are represented in green,
standard single tyres in red and wide base single tyres in blue. The TCF values for primary
roads are represented by solid bars, the TCF values for secondary roads are striped. Note that
for primary roads, only primary rutting is considered, while for secondary roads, it is the average
of primary rutting, secondary rutting and fatigue cracking that is used. In most cases the TCF-
28
values for each of the distress modes considered for secondary roads, are of the same order.
The only exception concerns steering axles, where secondary rutting accounts for roughly 60%
of the total TCF-value.
Figure 4.2: TCF of common current and possible future tyres for towed axles.
Figure 4.3: TCF of common current and possible future tyres for driven axles.
29
Figure 4.4: TCF of common current and possible future tyres for steering axles.
Based on the presented results, COST 334 makes the following conclusions:
The values in the graphs show that there is not one unique answer to the question
whether the common current and (possible) future wide base singles are better or worse
with respect to pavement damage.
Replacement of duals by wide base singles, both on towed or driven axles, generally
results in more pavement damage, for the observed range of common current and
possible future tyres. This effect is more pronounced on secondary roads.
Replacement of single tyres on steering axles by wide base singles, however, results in a
reduction of pavement damage.
Cebon cites several sources, exhibiting a wide span of results (see Cebon p. 321). The earliest
reports, from 1965 and 1978, indicate that the pavement damage from wide single tyres are up to
7-10 greater compared to dual tyres. Later work, from 1991 and 1992, report an increased road
damage between 1.1 4 for wide based single tyres. Another report (by M. Huhtala, VTT), from
1988, concludes that wide based single tyres are likely to cause 3.5 7 times more damage than
dual tyres, and that the worst conditions are for thinner asphalt layers. That study also reports
that a wide-based single tyre is only 1.5 times more damaging than an unevenly inflated dual pair
with 500 kPa in one tyre and 1000 kPa in the other.
Regarding these studies, Cebon notes that the large pavement damage factors in the studies
that he cites comes from the use of the fourth power law applied on measured (or calculated)
strains in the road under dual and wide single tyres. Quoting Cebon, This raises the important
question of whether a fourth power is appropriate, or, whether it may bias the results
excessively.
Cebon concludes that various experimental and theoretical studies have indicated that single and
wide based single tyres can cause up to 10 times more fatigue damage on thin flexible
pavements, compared to dual tyres carrying the same static load. Moreover, tyre contact
conditions are less important for rutting of thicker flexible pavements for which wide single tyres
30
are only 1.5 2 times more damaging than dual tyres, and that the tyre type has little influence
on fatigue damage of rigid pavements.
A TFK report from 1989 [9] presents equivalence factors for different tyre configurations, shown
in Fig. 4.5.
Figure 4.5: The load equivalents for different tyre configurations and tyre dimensions as a
function of axle load on two different asphalt thicknesses: 80 mm (left diagram) and 150 mm
(right diagram). A dual fitment with 12R22,5 tyres with axle load 10 tons has been used as a
reference and has the value 1. For all of the 5 configurations, three different tyre air pressures
are indicated, where the bold line corresponds to recommended air pressure, and the upper and
lower line corresponds to +20% and -20% of recommended air pressure respectively. From Ref.
[9].
A recent study from 2005 [10] concluded that a wide-base single tire (445/50R22.5) caused
similar, if not identical, pavement response as a conventional dual tire assembly. In fact, the two
configurations for the three measured responses (asphalt strain, base stress, subgrade stress)
produced, statistically, the same results.
Regarding the influence from using different tyre pressures: COST 334 reports that the ratio of
actual to recommended inflation pressure was shown to be influential for the cases of primary
rutting on thick (and probably medium) pavements and secondary rutting on thin and medium
pavements. An inflation pressure 10% higher than that recommended for the actual tyre load
results in about 15% increase in pavement wear.
Cebon reports (see Cebon p. 304-305) that several studies has indicated that fatigue damage
due to tensile strain at the bottom a thin asphalt pavements is likely to increase rapidly with
average contact pressure, while the inflation pressure has little effect on subgrade rutting. Based
on asphalt pavement strain measurements, it has been reported that a 40% increase in tyre
pressure would increase fatigue damage by 26%. Also, laboratory measurements on a 225 mm
31
thick asphalt road surface model have shown that rut depth development was approximately
linearly related to the average contact pressure (independent of load). Another study on high
type pavements found that overinflation of conventional dual tyres by 170 kPa nearly doubles
flexible pavement fatigue. Similar overinflation of wide base single tyres was even more critical,
increasing fatigue by a factor of four. In contrast, the tyre inflation on rigid pavements had a
moderate influence on fatigue.
Figure 4.6: Comparison of three suspension for a principal road (from Cebon p. 561)
The DIVINE project reports on a comparison between air and steel suspensions conducted as an
accelerated test on an indoor pavement. In both cases the static load was 49 kN and wide single
tyres were used. In brief the test showed that the steel suspension produced a 15 percent
increase in road roughness (IRI) and 10 percent more cracking. Rutting was generally low during
the test and no difference in mean rut depth between air and steel suspensions could be
observed. Looking at the maximum rut depths though, it was concluded that if the maximum rut
depth in the range 11-12 mm is taken as a critical level then the air suspension would achieve
45-65 percent more load cycles than the steel suspension to produce the same rutting distress.
Cebon summarises the general conclusion about the effects of suspension types on dynamic tyre
forces (Cebon, p. 123-124):
32
All studies found dynamic tyre forces to increase with speed and road roughness
It has been noted that reducing suspension stiffness generally reduces tyre forces
Centrally pivoted tandem axle suspensions such as walking beams and single-point
suspensions were always found to generate the highest dynamic loads because of their
lightly damped pitching modes at around 8-10 Hz. It has been noted, however, that these
suspensions can be improved considerably by suitable use of hydraulic dampers.
Four spring tandem suspensions were generally found to generate smaller dynamic
loads than walking beams. Torsion-bar and air suspensions generated the lowest loads.
It has been noted that modern single-spring parabolic suspensions with good hydraulic
damping are not significantly worse than stiff air suspensions. Moreover, air
suspensions without hydraulic dampers could generate significantly higher dynamic loads
than leaf spring suspensions.
Triaxle suspensions were found to generate smaller dynamic loads than tandem
suspensions in several studies.
It has been reported that varying the axle spacing of an air spring tandem suspension had
negligible effect on the dynamic loads, whereas the Dynamic Load Coefficient generated
by a four-spring tandem suspension varied considerably with axle spacing, depending on
the speed and road roughness.
= ( I II Pstat )
where
(4.2)
= the dynamic road stress factor (takes suspension type into account)
I = parameter to account for the tyre configuration (single or dual tyres)
II = parameter to account for the tyre contact pressure
Pstat = the static (average) tyre force
Sometimes an additional factor is included to also account for the type of axle group (single,
tandem or triaxle). Cebon concludes that considerable research effort has gone into quantifying
these parameters for a variety of suspensions and tyre contact conditions.
The road stress factor has also been the subject of considerable criticism (also from VTI [11]).
According to Cebon, The Road Stress Factor approach incorporates all of the uncertainties
inherent in the fourth power law, which has itself been the subject of considerable criticism. It has
three other highly questionable features:
I. It assumes that the strain in the road surface is directly proportional to the instantaneous
wheel force and neglects the sensitivity of road surface response to the speed and
frequency of the applied loads and to the structural response characteristics of the road.
33
II. In using the mean value of the damage criterion, it implicitly assumes that road damage is
spread randomly over the surface and does not account for any concentration of damage
which may occur in the vicinity of particular roughness features.
III. It assumes that each suspension system on a vehicle is dynamically independent and
does not influence the tyre forces, and hence road damage, generated by other axles.
Thus suspensions are compared through analysis of the wheel loads generated by
individual axles of axles groups, rather than through analysis of road damage done by the
whole vehicle.
The opinion on the use of the road stress factor seems to be split among the researches. Cebon
reports that Ullidtz notes that accounting for dynamic loads using a number of equivalent
standard axles calculated using a fourth power law (as per the Road Stress Factor), would result
in completely erroneous results. He also quotes Morris, and according to him the road stress
factor is a plausible rule of thumb that can serve as a bench-mark for comparison with more
analytical approaches.
COST 334 used the road stress factor approach to propose a formula for the damage
contribution of a single passage of an axle on primary roads. The so called Axle Wear Factor
(AWF) is a dimensionless factor relating the damage contribution of a specific tyre at a given axle
load and axle configuration to the damage contribution of a single passage of a reference tyre
with a reference axle load (10 tons).
P
AWF = TCF
10
where
(4.3)
TCF = the Tyre Configuration Factor developed in COST 334. The factor depends
on the total tread width (2 x single tyre tread width for dual fitment) and the diameter
of the tyre, and thus takes the tyre fitment (single or dual) into account. The
following formula is recommended: TCF = (tread width/470)-1.65 * (diameter/1059) -1.12
P = Axle load in tons.
The following remarks regarding this formula are made in COST 334:
34
Only primary roads are considered, so that only primary rutting is taken into account,
which implies a power of 2 in the load equivalency factor.
The value of the factor for Axle Configuration is assumed to be equal to unity (1). It is
generally accepted (OECD, 1983) that tandem or triple axles with axle spacings below 1.4
m, cause (slightly) more damage that two, or three passages, respectively, compared to a
single axle of equivalent loading. For primary rutting, however, no specific information is
available, and a factor of unity appears to be reasonable.
The value of the factor for Suspension Configuration is also assumed to be unity (1).
Strictly, this value is valid only for those axles having air suspension, but since this is the
case for most of the heavy goods vehicles under consideration, the assumption is again
reasonable.
The traction effects of the drive axle on the pavement are ignored.
Correction factors for load imbalance and dynamic effects in the TCF formula should be
ignored, since they add only about 1% of additional precision to the calculation of TCF.
35
Number of axle groups (axles that are separated by more than 1,8m are divided into
different groups)
Tyre configuration
Tyre width
Distance between tyres on the same axle (single or dual fitment, or something in
between)
Suspension type
The formula is
ESAL10 =
mi
5
i =1 j =1
mi
(5.1)
where
n = number of axle groups
mi = number of axles within group i
Ai = Constant that accounts for both the number of axles within group i, and the distance
between the axles within group i. There is a distinction between roads with a strong base
(where only primary rutting is considered), and those with a weaker base (where also
36
secondary rutting, a permanent deformation below the pavement, occurs). For the weaker
roads, Ai = the number of axles within the group, independent of the inter-axle distance. For
stronger roads, it is assumed that a group of axles with a short inter-axle distance is
producing less damage to the road, compared to the same group of axles with a larger interaxle distance. Thus, on this kind of roads, Ai is lower than the number of axles. A smaller
inter-axle distance leads to a smaller value of the constant.
Table 5-1
Ai
Strong road
Weak road
Number of Inter-axle
axles
distances
distance (m)
1.0
1.2
1.4
1.6
1.8
Bij = Constant that takes the tyre fitment on each axle into account. If dual fitment is used (or if
the distance between tyres on an axle is less than the tyre width + 8 cm), Bij = 1. If the tyre
fitment is single (if the distance between tyres on an axle is more than 2,5x the tyre width),
Bij is given a number that is larger than one, and which depends on the tyre width. A wider
tire leads to a lower value on the constant. If the distance between the tyres falls in between
the definitions of single and dual fitment, Bij is supposed to be determined by interpolation
between the single and dual values of the constant, taking the distance between tyres into
account. The values of this constant for single fitment are:
Table 5-2
Tyre width (mm) Bij
256
3,99
315
2,91
365
2,19
425
1,68
445
1,25
Ci = Contant that takes the type of suspension into account. The following suspension types are
distinguished between:
37
Table 5-3
Suspension type
Ci
Leaf spring
1,0
Parabolic spring ?
Coil spring
Rubber
0,95
Hydraulics
Air
Oil
No suspension
1,4
Dij = Constant that accounts for the air pressure of the tyres on a specific axle. A differentaion is
made between single and dual fitments, and whether primary rutting or deformations below
the pavement is considered. Dij is almost linear with respect to the air pressure, as shown in
the figure below.
The constant Dij
1.8
Primary rutting
Subpavement deformation: Single fitment
Subpavement deformation: Dual fitment
1.6
1.4
Dij
1.2
0.8
0.6
0.4
7
8
Tyre air pressure [bar]
10
38
Historically, since 1994 a factor 1.3 has been used. However, in ATB vg 2005, section C6,
methods for determining the B-factors are given:
To determine the B-factor measurements of gross vehicle weights and axle loads shall be
carried out. The measurements must be performed during a period of at least 7 days. The
B-factor shall then be calculated from these measurements. These measurements could
be performed using the B-WIM (Bridge Weigh In Motion) technique described in VV
Publication 2003:165. In that report a number of measurement results are presented
which could serve as a support for choosing the B-factor.
For calculation method of the ESAL (equivalent number of standard axles) for each heavy
vehicle type , ATB VG 2003 is referred to, which shows an example on how an axle of
16 tons (apparently not making any difference between single, tandem or triaxles) results
in 6.7 standard axles of 10 tons, using the fourth power rule. The corresponding formula
of the ESAL for each heavy vehicle type that is recommended by ATB VG 2003 can
then be written
39
W
ESAL =
100
(5.2)
If measurements cannot be carried out, ATB VG 2005 suggests that four or five of the
most common heavy vehicle classes are used, with weights estimated from local
experience. Then the fourth power rule is applied on each axle.
In a recent report from the Swedish Road Administration [14], concerning B-WIM measurements
during 2004-2005, another way of using the fourth power rule for calculating ESAL is
demonstrated. In section 4.5 of that report, as an illustration on the effects of overloading, four
examples on how ESAL values should be calculated are shown. No formula is given, but from
private communication with Tomas Winnerholt of the Swedish Road Administration, it was
clarified that the fourth power rule together with factors taking the axle type into account has
been used. More specifially:
i
W
ESAL = i k i
n =1 10
i=
Wi =
ki =
(5.3)
40
6
18
18
18
ESAL = * 0.0952 + * 0.0952 + * 0.0952 + = 3.1
10
10
10
10
8.7
20.9
21.2
15.4
ESAL =
= 4.85
* 0.0952 +
* 0.0952 +
* 0.0952 +
10
10
10
10
41
10 10
ESAL= + = 2
10 10
For a configuration as in Figure 5.2.4, the calculation is:
4
13 7
ESAL = + = 3.1
10 10
42
6. Acknowledgements
The authors would like to thank Leif Sjgren for contributing with material to this report. The
members of NVF Fordon och Transporter are also thanked for valuable comments and
suggenstions.
43
7. References
[1] D. Cebon, Handbook of Vehicle-Road Interaction. Swets &Zeilinger Publishers, Lisse, The
Netherlands (1999).
[2] Dynamic Interaction between Vehicle and Infrastructure Experiment, DIVINE Technical
Report, DSTI/DOT/RTR/IR6(98)1/FINAL, OECD (1998).
[3] COST 334 Effects of Wide Single tyres and dual Tyres Report, Chapter 4, version 29,
November 2001
[4] U. Isacsson, Interaktion mellan fordon/milj och vg, KTH, (2004)
[5] R. Blab Die Fahrspurverteilung als Einflussgre bei die bemessung des
Straenoberbaus; Mitteilungen des Institutes fr Strassenbau und Strassenerhaltung TU Wien,
ISTU, Heft 5; Institut fr Strassenbau und Strassenerhaltung TU Wien; 228 pp.; Wien, AT (1995)
[6] Jan M. Jansen, Srtransporters vejslid Klassificering av kretjer. Rapport nr 269,
Vejdirektoratet. 2002
[7] The Handbook of Highway Engineering, CRC Press, Taylor & Francis Group. 2006
[8] Private communication with Leif Sjgren, VTI.
[9] TFK rapport 1989:5 Optimalt Dckval fr Tunga Fordon, Fltmtningar av Vgpknning,
Transportekonomi och Vgkostnader. L. Djrf, M. Huhtala, M. Johansson, E. Samuelsson.
[10] A. L. Priest, D. H. Timm, and W. E: Barrett, NCAT report 05-03 Auburn University:
Mechanistic comparison of wide-base singe vs. standard dual tyre configurations. June 2005
[11] G. Magnusson, H.E. Carlsson, and E. Ohlsson, The influence of heavy vehicles springing
characteristics and tyre equipment on the deterioration of hte road, VTI (Translated by TRRL as
WP/V&ED/86/16, 1986), Report Number 270, 1984.
[12] ATB VG 2005. VV Publication 2005:112.
[13] BWIM-mtningar 2002 och 2003 Slutrapport. VV Publication 2003:165
[14] BWIM-mtningar 2004 och 2005 Projektrapport. VV Publication 2006:136
44
NVF
Vejdirektoratet
Niels Juels Gade 13
Postboks 9018
DK-1022 Kbenhavn K
Danmark
Telefon +45 7244 33 33 telefax +45 33 32 98 30
E-post: nvf@vd.dk
NVF
c/o Vgfrvaltningen
Postbox 33
FIN-00521 Helsingfors
Finland
Telefon +358 204 22 2575 telefax +358 204 22 2471
E-post: nvf@finnra.fi
NVF
c/o Landsverk
Box 78
FO-110 Torshavn
Frerne
Telefon +298 340 800 telefax +298 340 801
E-post: lv@lv.fo
NVF
c/o Vegagerdin
Borgartun 7
IS-105 Reykjavik
Island
Telefon +354 522 1000 telefax +354 522 1009
E-post: hreinn.haraldsson@vegagerdin.is
NVF
c/o Vegdirektoratet
Postboks 8142 Dep
NO-0033 Oslo
Norge
Telefon +47 22 07 38 37 telefax +47 22 07 37 68
E-post: publvd@vegvesen.no
NVF
c/o Vgverket
SE-781 87 Borlnge
Sverige
Telefon +46 243 757 27 telefax +46 243 757 73
E-post: nvf@vv.se
NVF-rapporterna kan bestllas via respektive lands sekretariat per telefon, fax,
e-post eller post. Se kontaktuppgifterna p nst sista sidan.
En uppdaterad rapportfrteckning finns p frbundets nordiska hemsida,
http://www.nvfnorden.org.
45