25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
USING THE FREIGHT AXLE MASS LIMITS
INVESTIGATION TOOL (FAMLIT) TO ESTIMATE THE
MARGINAL COST OF ROAD WEAR
Will Hore-Lacy, Thorolf Thoresen and Tim Martin, ARRB Group,
Australia
ABSTRACT
Improved freight vehicle productivity can be obtained by increasing the allowable axle loads
above current load limits. In turn improved productivity can potentially reduce greenhouse gas
(GHG) emissions by reducing the number of freight vehicles for a given freight task. Increased
axle load pavement impacts can be managed using the marginal cost of road wear as the basis
of a price for increasing axle loads. These prices can provide a signal for targeting maintenance
and rehabilitation funding provided the revenue raised by the price is linked to funding.
FAMLIT is a pavement life-cycle costing model that has been used to estimate load-wear-cost
(LWC) relationships for a range of typical roads and pavement types subject to incremental
loading over current load limits by six axle groups. These load wear costs were estimated as
the present value (PV) of maintenance and rehabilitation costs incurred by managing road
within agreed conditions. These costs have been converted to equivalent annual uniform costs
(EAUC) and developed into LWC relationship against axle load (tonne-km) and standard axle
repetitions (SAR-km). The marginal cost of road wear was determined by the first derivative of
the LWC relationships. The estimated marginal road wear costs were found to vary across the
various road types and were highly dependent on the pavement/subgrade strength.
INTRODUCTION
Historically, improvements to Australia’s road freight productivity have mainly been brought
about by increasing freight vehicle payloads and allowing access to vehicles with longer, wider
and taller configurations that still comply with existing safety standards. However, most state
and local government road agencies are generally concerned about allowing greater heavy
vehicle access to the road infrastructure due to pavement wear. These agencies have the view
that additional road wear is quite significant once the current legal axle limits are exceeded. As
a consequence, two significant issues need to be resolved. Firstly, can the current road network
be used more productively and, secondly, can road network access be expanded in a way that
enables the additional cost of this improved access to be fully recovered.
In this context, following the Productivity Commission’s (2006) inquiry into road and rail
infrastructure pricing, Australian governments have been engaged in a process of exploring
alternative road infrastructure pricing models for heavy vehicles through the Council of
Australian Governments (COAG)’s road reform plan (CRRP). The common theme in this plan is
the belief that more efficient price signals to heavy vehicle road users regarding their use of
road infrastructure, taking into account the actual nature of the usage of the vehicle, in terms of
mass, distance travelled and the location of the road usage, has the capacity to improve the
utilisation of the road network by encouraging the right vehicle onto the right road. In addition,
this initiative has included investigation of pricing schemes that would enable access to be
opened up for vehicles to carry loads greater than the current mass limits provided that they
were associated with a charge to reflect the additional road wear cost.
This paper presents findings of recent Austroads and National Transport Commission (NTC)
research and other work in this area that has investigated the marginal costs associated with
higher axle loads on heavy vehicles, taking into consideration that the Australian road network
© ARRB Group Ltd and Authors 2012
1
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
has many different road and pavement types. The expectation is that these marginal costs can
provide input into the evaluation of road infrastructure pricing reform.
GENERAL APPROACH AND ASSUMPTIONS
Definition of Marginal Cost
Economic efficiency requires that prices are set equal to marginal costs. Total and marginal
costs of road usage, excluding congestion and other external costs, take into account two sets
of factors:
 the impact on road users in terms of vehicle operating costs
 the impact of heavy vehicles on the road infrastructure.
This paper sets out the approach used to estimate the marginal road infrastructure cost
resulting from increased axle loading on heavy vehicle axle groups beyond current legal or
agreed axle load limits. Marginal costs measure the additional whole of life agency cost
associated with either an additional axle group pass, or SAR unit. However, the marginal road
wear cost impact on road users has not been considered. The approach used in estimating the
marginal cost has been based on a pavement life-cycle costing analysis that ensured that the
level of service, in terms of the roughness and strength of the road pavement, was constrained
to fall within defined and well established bounds which means that the marginal cost impact on
road users is close to zero as illustrated in Newbery (1988). Since the focus is on the impact
that additional axle loading has on road wear costs, this paper is focused on estimating the
marginal cost of road wear which is defined as all of the relevant road costs that are impacted
by road usage. There are two types of marginal costs that can be considered:
 Short-run marginal costs (SRMC) take into account the cost of maintaining a road within its
defined roughness and strength level, but with the constraint that no pavement strengthening
will be allowed beyond its initial design strength.
 The long-run marginal costs (LRMC) also take into account the cost of maintaining a road
within its defined roughness and strength level including allowance for pavement
strengthening of any nature (including reconstructions) to occur at any time during its lifecycle.
It should be noted the pavement strengthening under LRMC conditions is timed so that the road
agency life-cycle costs are reduced as much as possible, although not necessarily optimised
because road user costs have not been directly considered except by limiting road wear to
defined levels of service, as noted above.
FAMLIT Analysis Tool
The road wear costs were estimated using the Freight Axle Mass Limits Investigation Tool
(FAMLIT), a pavement life-cycle costing analysis model (Michel and Toole 2006), which was
applied to six different axle group types for a range of road types, representing Australia’s
sealed road network, including local roads, over a 50 year analysis period. Road wear costs
were based on the present value (PV) of the routine and periodic maintenance and rehabilitation
costs incurred by managing each road type within agreed condition (roughness and strength)
limits over the analysis period. The PVs of these aggregated costs were converted (Hudson et
al. 1997) to their equivalent annual uniform costs (EAUC, Australian dollars, AUD/lane-km/year)
to simplify the analysis.
The impacts of increasing axle load increments on road wear were quantified by developing
separate load-wear-cost (LWC) relationships for each of six axle groups for each of the
designated road types in the Australian sealed road network (Thoresen et al. 2012). Load-wear
costs, in terms of EAUC, were estimated using FAMLIT for 1 tonne increment increases for
each group axle load over a range of axle load increases ranging from the tare weight to well in
excess of those allowed under the general mass limits (GML) regulatory framework. For each
© ARRB Group Ltd and Authors 2012
2
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
axle group pass, apart from the target axle group, the axle loads of the other axle groups were
held constant at GML.
A simulated road network, using representative traffic and traffic load data, was developed as
follows to represent all three sealed pavement types (sprayed seal unbound granular, GN;
asphalt, AC; and, cement stabilised, CS) categorised in terms of traffic load capacity across a
range of different locations and climates in Australia:
 Some 17 road types defined by pavement type and road hierarchy were used to represent
the road network, which were further categorised in terms of design traffic levels, into urban
and rural locations under typical regional climates (three zones). They formed the basis of
an initial road network matrix of new and in-service pavements.
 The new pavements (N) were designed to suit assumed existing traffic loads (Austroads
2004) using an assumed Californian Bearing Ratio (CBR) of 5%.
 The existing in-service pavements (S) of each road type were examined at different points of
their life-cycle, so each road type was assumed to include sections covering a typical range
of pavement ages (distributed over 10, 20, 30 and 40 years, including new pavements). The
conditions (roughness and strength) of these sections were predicted by applying road
deterioration (RD) models to a newly constructed or rehabilitated pavement.
 Each road section was assumed to be one lane wide, with typical width, and one kilometre
long so road wear costs could be calculated in terms of AUD/lane-km. Potential biases
associated with initial pavement age and condition were controlled for by using a range of
ages and conditions for each road/pavement type.
The above assumptions regarding pavement conditions and pavement ages were necessary
because the relevant Australian wide network level pavement condition information was not
available. This means that the condition and age assumptions, although reasonable for a
network simulation, do not represent any particular part of the road network. However, the
simulation did cover the usual expected range of the variables impacting on the marginal cost of
road wear associated with higher axle loads on heavy vehicles, taking into consideration that
the Australian road network has many different road and pavement types.
The following additional assumptions were used in determining road wear costs:
 No traffic growth of heavy vehicles was considered during the life-cycle costing analyses of
road wear.
 The life-cycle costing analyses of road wear were conducted under an unconstrained
budget, that is, the road wear costs of all loading scenarios assumed that adequate funds
were available to perform maintenance and rehabilitation works when needed.
 The annual routine and periodic maintenance costs, me, were increased with increasing axle
loads. This was to simulate increases in these costs due to either the increased frequency of
resealing/resurfacing or the use of higher quality reseals/resurfacing along with additional
routine maintenance work to cope with the increased axle loads.
The annual pavement maintenance cost, me (AUD/lane-km), was previously found to be
related to road use (Martin et al. 2010) as follows:
me =
α + 0.00309 × ESA/lane/year
1
where
α=
ESA/lane/year =
routine maintenance cost (increased with traffic load range)
equivalent standard axle per lane per year.
Equation (1) quantified the impact of increased axle group loads on the pavement
maintenance expenditure.
© ARRB Group Ltd and Authors 2012
3
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
FAMLIT Life-cycle Costing Analysis
The life-cycle costing analysis in FAMLIT involved the following:
 Rehabilitation works were triggered as a result of meeting pre specified intervention criteria.
Following rehabilitation works, the surface conditions were reset to new condition values
(Austroads 2007) and the pavement age was reset to zero. Road deterioration prediction
then recommenced from this point in the same manner as before.
 The road costs for the maintenance works (me, see equation (1)) were applied annually and
at a specific point in time for pavement rehabilitation.
 The weighted average EAUC for each pavement and road type and each climatic region was
then determined using the nominated pavement age distribution for each axle group load
increment to develop the LWC relationships with axle group load (tonne-km) and with
standard axle repetitions (SAR-km).
Pavement deterioration and interventions
One of the key aspects of the modelling is the way in which the road is predicted to deteriorate
over time and under what conditions the interventions are triggered. The rutting/roughness
deterioration model (RRD) was used (Austroads 2010a, 2010b, Martin 2009) where the
roughness deterioration is a function of cumulative SAR, climate, initial pavement strength and
the following variables:
1. The amount of annual pavement maintenance cost, me, undertaken affects the rate of
roughness deterioration where increasing the amount of ‘me’ reduces the rate of
deterioration and reducing the amount of ‘me’ increases the rate of deterioration and
influences the need for an intervention by rehabilitation.
2. The amount of rutting present also influences the rate of roughness deterioration and is also
taken into consideration in establishing whether an intervention by rehabilitation is required.
3. The initial design pavement strength (or new value after intervention).
Loading Scenarios
The loading scenarios considered covered the six main axle groups as follows: single axle
single tyre (SAST), single axle dual tyre (SADT), tandem axle single tyre (TAST), tandem axle
dual tyre (TADT), tri-axle dual tyre (TRDT) and quad axle dual tyre (QADT). The loading
scenarios used one tonne axle group load increments starting at the a mass associated with
vehicle tare weight and increasing well beyond GML. Table 1 summarises the loading scenarios
used.
As the analysis had assumed the distribution of load on the different axles groups for each
pavement and road type, it was therefore possible to make the above separate assessment of
the road wear cost of each axle group with incremental load increases.
Table 1: Axle group loading combinations used in the analysis
Loading
scenario
SAST
(tonne)
SADT
(tonne)
TAST
(tonne)
TADT
(tonne)
TRDT
(tonne)
QADT
(tonne)
Reference load
5.4
8.15
9.17
13.76
18.45
22.53
Load increment
1
1
1
1
1
1
Axle load offset
2
3
5
5.5
9
9.75
Tare weight
3
4
6
6.5
10
10.75
Base GML
6
9
11
16.5
20
24
Maximum load
12
15
20
26.5
35
45
© ARRB Group Ltd and Authors 2012
4
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Load increment
range
3 - 12
4 - 15
6 – 20
6.5 - 26.5
10 - 35
10.75 - 45
Traffic Loading
The annual traffic loading (base case) for each road type was based on annual weigh-in-motion
(WIM) data, which allowed assessment of the various heavy vehicle types and axle group load
distributions (Austroads 2008) together with typical annual average daily traffic (AADT) volumes.
Studies on the load equivalency and damage exponents for different pavement types
(Austroads 2010c) recommend the following damage exponents in estimating SAR to assess
the road wear impacts of different axle group loads on the different pavement types:
 granular pavements with a sprayed seal, GN, used SAR4, i.e., a damage exponent of 4
 asphalt pavements, AC, used SAR5, i.e., a damage exponent of 5
 cement stabilised pavements, CS, used SAR12, i.e., a damage exponent of 12.
Consideration of SRMC and LRMC
Initially, marginal costs were estimated using the standard SRMC approach with FAMLIT. At
each intervention, rehabilitation by either an AC overlay or a GN re-sheet was undertaken at a
thickness that ensured that the pavement strength was returned to no greater than the initial
design strength. However, the weakness of this approach was that the thickness was likely in
many cases to be substantially less than what was required to return roughness back to a
satisfactory achievable reset value of roughness and was usually a non-practical thickness.
This first approach was defined as SRMC1.
In order to address the weakness of the SRMC1 approach, FAMLIT was altered to allow for a
realistic minimum overlay/re-sheet thickness to be applied at each rehabilitation intervention to
ensure that the roughness was returned back to a satisfactory achievable reset value. The
minimum thickness that was applied was different for an AC overlay and GN re-sheet. This
altered approach was defined as SRMC2. The unintended result of applying a minimum overlay
thickness was that if the strength did not deteriorate to a certain point by the time roughness
reached its intervention trigger point, the minimum overlay thickness would increase strength
beyond its initial design value. This somewhat contradicts the definition of SRMC and is close
to the definition of LRMC. However, it does deliver a reasonably consistent level of service for
road users, in terms of roughness within defined bounds and reflects the reality of maintaining
pavements.
In terms of LRMCs, the constraint that strength cannot be varied beyond its initial design
strength was breached with SRMC2. However, this was only done to ensure service levels
were maintained. A standard LRMC would allow strength to be varied at any point in the
pavement life-cycle. In this context LRMC1 was developed in FAMLIT such that strength is
improved at the first intervention point if this is considered optimal in the sense that higher loads
necessitate higher design strength. Clearly, there are some imperfections in this approach since
strength is only altered at the first intervention point and reconstruction at the start of the lifecycle was not considered as an option. It was also considered that reconstruction would not be
optimal given current pavement engineering practices except for the non-typical case of very
high loads on low strength pavements.
The analyses conducted using FAMLIT were under the condition of SRMC2 because under
most conditions SRMC2 is approximately equal to LRMC as found from a previous study (Martin
et al. 2010).
Representation of Analysis
To simplify the representation of the results, the EAUC values were regressed against the
tonne-km and SAR-km values separately for each one tonne load increment on the axle group
to form the LWC relationships. The two complementary sets of equations allow marginal costs
to be appreciated from both the vehicle operator and pavement provider perspectives. This
© ARRB Group Ltd and Authors 2012
5
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
resulted in a series of regression equations (LWC) for each axle group and road and pavement
type combination. The marginal cost (MC) of road usage was determined for each climate/axle
group/pavement type/road type combination. The MC was based on the LWC relationship, a
non-linear relationship in terms of EAUC as a function of axle group load (tonne-km), as shown
in Equation 2:
EAUC =
a0 + (TMI + TMIOFFSET)a1 + a2 × (Axle load – Offset)a3
2
The annual marginal cost (c/tonne-km/year), MCann, is the first derivative of Equation 2 as
shown by Equation 3:
MCann =
a2 × a3 × (Axle load – Offset)(a3 - 1)
3
where
TMI =
TMIOFFSET =
Offset =
Axle load =
a0, a1, a2 =
Thornthwaite Moisture Index (Thornthwaite 1948)
minimum TMI value used for each road type plus one
approximate group axle reference load base tonnes before incremental load
were varied
total load on axle group (tonne) per lane
non-linear regression coefficients
The MC was also based on the LWC relationship, a non-linear relationship expressed in terms
of EAUC as a function of standard axle repetitions (SAR-km) as shown in Equation 4:
EAUC =
a0I + (TMI + TMIOFFSET)a1l + a2I × (SAR-km – Offset)
4
The annual marginal cost (c/SAR-km/year), MCann, is the first derivative of Equation 4 as shown
by Equation 5:
MCann =
a2I
5
where
SAR-km =
Offset =
a0l, a1l, a2l =
annual pavement wear, SAR per lane -km
approximately reference base SAR-km before incremental loads increased
non-linear regression coefficients
all other variables are as defined previously.
The marginal costs determined from Equation 5 have the convenience of being a constant value
with increased SARs. This is in contrast to the marginal cost determined from the non-linear
Equation 3 which gives an increasing marginal cost with increased axle load.
Marginal axle load costs per axle pass can be calculated by dividing marginal axle load (tonnekm) costs per annum by the number of axle group passes per year. Similarly, the marginal
costs per SAR can be calculated by dividing the marginal SAR-km costs per annum by the
number of SAR-km per year.
LWC RELATIONSHIP ANALYSES AND RESULTS
Road and Pavement Types Considered
The following road types and pavements were analysed as part of the determination of the LWC
relationships (Equations 2 and 4) and the marginal cost estimation (Equations 3 and 5):

rural freeways with GN pavements
© ARRB Group Ltd and Authors 2012
6
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012

rural arterials with GN, AC and CS pavements in service (S); this road type represents
nearly 26% of the sealed road network with the GN pavement representing 96% of this
road type (Austroads 2005)

rural arterials with GN pavements new (N)

urban arterials with GN pavements; this road type represents 3.9% of the sealed road
network and the GN pavement represents 39% of this road type

rural collectors with GN pavements; this road type represents 10% of the sealed road
network

rural access roads with GN pavements; this road type represents 19% of the sealed road
network.
Climates considered
The TMI extremes considered were from -50 to +80. Table 2 shows the climate zones for TMI
variations that were applied to the various road types.
Table 2: TMI climate classification
Climate zones
TMI
Alpine/wet coastal
+80 to + 40
Wet temperate
+ 10 to + 40
Temperate
+ 10 to - 5
Dry temperate
- 5 to - 25
Semi-arid
- 25 to -50
LWC Relationships as a Function of Axle Load
Impact of pavement type and axle group on a given road type
For the three most common heavy vehicle axle groups, SADT, TADT and TRDT, the LWC
relationships using EAUC as a function of tonne-km (Equation 2) were determined on an inservice (S) rural arterial whose pavement types were: GN, AC and CS under mid-range TMI
values (temperate climate).
Figures 1(a) to 1(c) show the distinctly different LWC relationships for the different pavement
types, while the variations in LWC relationships due to the different axle groups are relatively
minor. From these figures it is apparent that the CS pavements, which have a load damage
exponent of 12, are the most cost sensitive to increases in axle load and therefore will have the
highest marginal cost. The next most cost sensitive to axle load increases are the AC
pavements with a load damage exponent of 5, followed by the GN pavements with a load
damage exponent of 4.
Figure (1b) for the TADT axle group appears to have the highest LWC relationships for all
pavement types. This means that this axle group is likeliest to produce the highest marginal
cost.
© ARRB Group Ltd and Authors 2012
7
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Tandem Axle Dual Tyre - Rutting/roughness model
Single Axle Dual Tyre - Rutting/roughness model
25000
25000
Rural Arterial GN (S)
Rural Arterial GN (S)
20000
Rural Arterial AC (S)
Rural Arterial AC (S)
Rural Arterial CS
Rural Arterial CS
EAUC ($)
EAUC ($)
20000
15000
10000
15000
10000
5000
5000
0
0
4
6
8
10
12
14
16
6
11
Axle Group Mass (t)
16
21
26
Axle Group Mass (t)
1(a) SADT axle group
1(b) TADT axle group
Tri Axle Dual Tyre - Rutting/roughness model
25000
Rural Arterial GN (S)
EAUC ($)
20000
Rural Arterial AC (S)
Rural Arterial CS
15000
10000
5000
0
9
14
19
24
29
34
Axle Group Mass (t)
1(c) TRDT axle group
Figure 1: EAUC axle group loads variations with load for pavement types on a rural arterial
(mid-range TMI values)
Impact by road hierarchy type
For the most common heavy vehicle axle group, SADT, the LWC relationships using EAUC as a
function of tonne-km (Equation 2) were determined for various typical road types with a GN
pavement under mid-range TMI values (temperate climate).
Rutting/roughness model
1400
Rural Fw y GN
1200
Rural Art GN (N)
Rural Collector GN
1000
EAUC ($)
Rural Access GN
800
600
400
200
0
3
5
7
9
11
13
15
Axle m ass (t)
Figure 2: EAUC road types variations with load on GN pavements (mid-range TMI values,
SADT axle group)
© ARRB Group Ltd and Authors 2012
8
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Figure 2 shows that although the rural collectors and rural access roads have relatively low
magnitudes of EAUC (road wear cost), they are the most cost sensitive to axle load increases.
This suggests that these road types will have a high marginal cost. On the other hand, while the
rural freeway has the highest magnitude of EAUC, it is the least cost sensitive to axle load
increases and therefore will have a low marginal cost.
Impact of axle groups on various road types
For all the common heavy vehicle axle groups, SAST, SADT, TADT, TAST, TRDT and QADT, the
LWC relationships using EAUC as a function of tonne-km (Equation 2) were determined for
various typical road types with a GN pavement under mid-range TMI values (temperate
climate).
Figures 3(c) and 3(d) show that the rural collectors and rural access roads are the most cost
sensitive roads to axle load increases. Figures 3(a) and 3(b) show that the rural freeways and
rural arterials are much less cost sensitive. Figure 3(a) to 3(d) all show that the SAST axle
group is the most cost sensitive on the road types examined, followed by the SADT axle group.
The least cost sensitive to axle load increases are the TRDT and QADT axle groups.
Rural Arterial GN (N) - Rutting/roughness model
Rural Freeway GN - Rutting/roughness model
25
6
20
SAST
SADT
TAST
TADT
TRDT
QADT
4
EAUC (cents)
EAUC (cents)
5
3
SAST
SADT
TAST
TADT
TRDT
QADT
15
10
2
5
1
0
0
0
5
10
15
20
25
0
30
5
10
15
20
3(a) Rural freeway
30
3(b) Rural arterial
Rural Access GN Rutting/roughness model
Rural Collector GN - Rutting/roughness model
800
300
700
SAST
SADT
TAST
TADT
TRDT
QADT
250
SAST
SADT
TAST
TADT
TRDT
QADT
200
150
600
EAUC (cents)
EAUC (cents)
25
Axle mass in excess of tare (t)
Axle mass in excess of tare (t)
500
400
300
100
200
50
100
0
0
0
5
10
15
20
25
30
Axle mass in excess of tare (t)
3(c) Rural collector
0
5
10
15
20
25
30
Axle mass in excess of tare (t)
3(d) Rural access
Figure 3: EAUC variation with load by axle group and road type (rural GN road types,
mid-range TMI values)
Sensitivity analyses
For a given in-service (S) rural arterial with a GN pavement and axle group, three main
parameters influencing EAUC with increased axle load were varied as follows under a constant
TMI (= 0):
© ARRB Group Ltd and Authors 2012
9
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012

the pavement/subgrade strength was varied ± 30% from the original value used in the
FAMLIT analysis

roughness progression for the pavement was reduced by lowering the calibration factor in
the RD model

the traffic level on the road type was varied ± 15% from the original value used in the
FAMLIT analysis.
Figure 4 shows the resulting impact on the EAUC estimates with increasing axle load due to
changes in the above variables. Figure 4 shows the most significant influence on EAUC with
increased axle load was the reduction in strength. Under these conditions EAUC increased
exponentially with increased axle load, suggesting increasing marginal cost. An increase in
strength reduced EAUC to a lesser extent than the decrease in strength increased EAUC.
Figure 4 tends to show under these conditions only minor increases in marginal cost occur with
increased axle load. Changes in traffic had much less influence on EAUC and the reduction in
roughness progression uniformly reduced EAUC with increased axle load. These results show
that the most pronounced influence on marginal cost was a reduction in strength.
37000
Base
Rough Progression
32000
Strength (low)
EAUC ($)
27000
Strength (high)
Traffic (low)
22000
Traffic (high)
17000
12000
7000
2000
15
17
19
21
23
25
axle mass (t)
Figure 4: Effect of varying pavement and traffic characteristics on EAUC for TRDT axle
group load on in-service rural arterial GN pavements (TMI = 0)
LWC Relationships as a Function of SARs
Impact of axle group on a given road type
For the all the axle groups, the LWC relationships using EAUC as a function of SAR-km
(Equation 4) were determined on an urban arterial with pavement types GN under a given TMI.
50000
45000
SAST
SADT
35000
TAST
TADT
30000
TRDT
QADT
EAUC ($)
40000
25000
20000
15000
10000
5000
0
200000
300000
400000
500000
600000
700000
800000
SARS
Figure 5: EAUC vs. SAR for all axle groups, urban arterial GN pavement
© ARRB Group Ltd and Authors 2012
10
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
As Figure 5 shows that approximately the same linear EAUC relationship with SAR-km occurs
for all the axle groups on a given road type. This is because when the axle loads (tonne-km)
are expressed as SAR-km there is no need to present the results by different axle groups
because of the axle load equivalency nature of the specific reference load for each axle group
(in converting axle load to SARs) that gives equal wear for each axle group relative to the
standard single axle.
Impact of road type on SAR relationships
For the TADT axle group, the LWC relationships using EAUC as a function of SAR-km (Equation
4) were determined for various typical road types with a GN pavement a constant TMI value.
Figure 6 shows the five distinct LWC relationships with EAUC as a function of SAR-km for the
five road types. The slopes of these relationships give the marginal cost with the lowest
marginal costs on the rural freeway and rural arterial (N), while the highest marginal costs occur
on the rural collector and rural access roads. This is the same outcome found when using the
LWC relationship with EAUC as a function of tonne-km for the SADT axle group.
20000
Rural Freeway GN
18000
Rural Arterial GN (S)
EAUC ($)
16000
Rural Arterial GN (N)
14000
Rural Collector GN
12000
Rural Access GN
10000
8000
6000
4000
2000
0
0
100000
200000
300000
400000
500000
600000
SARS
Figure 6: EAUC vs. SAR relationships for GN pavements (TADT axle group)
Climate
Although different climatic conditions, represented by the variable TMI, were tested for each
road type and while they influenced the magnitude of EAUC they did not have an impact on
marginal costs. This is confirmed by Equation 5 where the coefficient, a2l, in the original LWC
relationship as a function of SAR-km, is the coefficient for load, in terms of SAR only (see
Equation 4).
MARGINAL COST ESTIMATES
Table 3 summarises the marginal cost estimates for the road types considered by this paper
using the LWC relationships based on EAUC as a function of SAR-km.
Table 3: Summary of marginal cost estimates
Road type
Marginal cost
(c/SAR-km)
Rural freeway GN
0.8
Rural arterial GN (S)
1.8
Rural arterial GN (N)
0.9
© ARRB Group Ltd and Authors 2012
11
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Rural collector GN
21.6
Rural access GN
55.7
Urban arterial GN
0.8
Table 3 shows and confirms the earlier observations that the rural collectors and rural access
roads have the highest marginal cost while the rural freeways and rural arterials have the lowest
marginal cost. Table 3 also shows the impact of a new pavement (N) compared to an in-service
pavement (S) for rural arterials on the marginal cost, with the rural arterial (S) having a much
larger marginal cost than the rural arterial (N) as expected.
SUMMARY AND CONCLUSIONS
The findings from this research study have illustrated some important outcomes regarding the
estimation of the marginal costs of road wear as follows:
1.
Some investigation into the definition and measurement of the SRMC and LRMC was
needed to achieve a realistic assessment of these measures in this context. The
SRMC was defined by its aim to restore pavement strength its original design value
when intervention is needed to maintain set levels of service, using a minimum
rehabilitation treatment thickness. This definition often resulted in the restored
pavement having a higher strength than the original design strength because of the
magnitude of the minimum thickness of the rehabilitation treatment. A previous study
found that the SRMC and LRMC were approximately equal under most conditions.
2.
The marginal road wear costs can be simplified to a cost per SAR-km measure, which
can be applied to all axle groups on a given road type. This has the advantage of being
a constant cost per road type while the marginal cost, using tonne-km, varies as the
axle load varies
3.
Variation in climatic conditions, as represented by TMI, while they can influence the
magnitude of the annual road wear costs, EAUC, they do not influence the marginal
costs based on LWC relationships using Equations 2 and 4.
4.
A sensitivity analysis confirmed that the pavement/subgrade strength has the greatest
impact on the estimated marginal cost. This is because of the sensitivity of RD models
used in FAMLIT to increased axle load where the pavement strength is inadequate for
the load.
5.
The LWC relationships of EAUC as a function of tonne-km and SAR-km both confirmed
that the rural collectors and rural access roads have the highest marginal cost while the
rural freeways and rural arterials have the lowest marginal cost.
6.
The impact of age of the pavement, either a new (N) or an in-service (S) pavement, has
a significant impact on the marginal cost.
REFERENCES
Austroads. (2007), Interim Works Effects Models, AP-R 300/07, Austroads, Sydney, Australia.
Austroads. (2008), High Productivity Vehicles and Pavement Economic Impacts: Network Level
Assessment Approach, Draft Austroads Report, Austroads, Sydney, Australia.
Byrne, M. and Martin, T. (2008), Establishing a New National Pavement Maintenance Database,
Report for Austroads Project AS1337, ARRB, Vermont South, Victoria, Australia.
Austroads 2010a, ‘Interim network level functional road deterioration models’, by T Martin, & L
Choummanivong, AP-T158/10, Austroads, Sydney, NSW.
© ARRB Group Ltd and Authors 2012
12
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Austroads 2010b, ‘Predicting structural deterioration of pavements at a network level: interim
models’, by L Choummanivong &T Martin, AP-T159/10, Austroads, Sydney, NSW.
Austroads, (2010), Guide to the Pavement Technology: Part 2: Pavement Structural Design, by
G Jameson, AGPT02/10, Austroads, Sydney, Australia.
Gomez-Ibanez J., Tye W.B., Winston C. (1999), ‘Pricing’ in Essays in Transportation Economics
and Policy.
Hudson, W.R., Haas, R. and Waheed, U. (1997), Infrastructure Management, McGraw-Hill, New
York, USA.
Martin, T. (2009), “New Deterioration Models for Sealed Granular Pavements,” Proceedings
Institution of Civil Engineers - Transport, 162, 4, pp 215-226, Thomas Telford, London, UK.
Martin, T, Byrne, M & Aguiar, G 2010, ‘Establishment of a New Pavement Maintenance
Database – Stage 1 & 2 Analysis’, contract report, Austroads project AS1337, ARRB Group,
Vermont South, Victoria, Australia.
Martin, T, Thoresen, T, Clarke, M. & Hore-Lacy, W. 2010, Estimating the marginal cost of road
wear on Australia’s sealed road network, Proceedings HVTT11 International Heavy Vehicle
Symposium, Melbourne, Australia.
Michel, N. and Toole, T. (2006) Freight Analysis and Maintenance Costs Estimation Tool: Part 1
- Overview, ARRB Report VC71393, ARRB, Vermont South, Victoria, Australia.
Newbery, David M. (1988), Road Damage Externalities and Road User Charges, Econometrica
56: 295-316
Patterson, WDO and Attoh-Okine, B (1992) Summary Models of Paved Road Deterioration
Based on HDM-III, Transportation Research Record, No: 1344, pp 99-105.
Productivity Commission. (2006), Productivity Commission Inquiry Report, Road and Rail
Freight Infrastructure Pricing Report, Melbourne, Australia.
Small, K.A., Winston C. and Evans C.A. (1989), Road Work a New Highway Pricing and
Investment Policy, Brookings Institute, Washington DC, US.
Thoresen, T., Martin, T., Hassan, R., Byrne, M., Hore-Lacy, W. and Jameson, G. (2012),
Preliminary Methodology for Estimating Cost Implications of Incremental Loads on Road
Pavements, Report for Austroads Project AT1394, ARRB, Vermont South, Victoria, Australia.
Thornthwaite, C.W. (1948), “An Approach Toward a Rational Classification of Climate,”
Geographical Review, 38(55), American Geographical Society, New York, USA.
AUTHOR BIOGRAPHIES
Will Hore-Lacy
Will Hore-Lacy commenced working with ARRB Group in September 2005 as an engineer while
completing his final semester at university. Since then he has worked across a number of areas
including asset management, transport economics, road safety, pavements research, transport
operations and transport planning. This work has include; road safety assessments (including
field testing), involvement with the Accelerated Load Facility, onsite heavy vehicle testing,
economics research and crash database analysis. More recently his work at ARRB has
increasingly involved programming, databases manipulation, GIS and statistical analysis as a
way of managing data and generating outputs, for asset management, pavement condition and
road safety data. Recent work has involved historical pavement data analysis and well as future
scenario modelling, both at a national level. He has also been involved in developing pavement
maintenance treatment and prioritisation tools for local government.
© ARRB Group Ltd and Authors 2012
13
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Thorolf Thoresen
Thorolf Thoresen is one of the leading Australian experts on road user costs and road
expenditure evaluation methods. He is an economist of over 30 years experience in roads and
road transport, with a specific interest in statistics and quantitative methods. He holds an
honours economics degree and post graduate qualifications in statistics, and has had
considerable experience working at the interface between engineering and economics. At
ARRB Group, Thorolf has undertaken or managed a wide range of projects in the transport
economics area for both local and overseas clients. He also has previous experience with the
Commonwealth Government in shipping and land transport. He is a member of a number of
key Austroads committees and working groups and an author of some 150 research reports and
papers.
Tim Martin
Dr Tim Martin is a Chief Scientist, Asset Management, and has graduate and post graduate
qualifications in civil engineering and a PhD in predicting network level sealed granular
pavement deterioration from Monash University. Tim's research at ARRB has involved leading
an extensive road track cost attribution study, the design and implementation of a major
observational study using long term pavement performance (LTPP) and long term pavement
maintenance (LTPPM) sites, experimental studies with accelerated load testing, and the
development of a range of pavement deterioration and works effects models for arterial road
sealed granular pavements in a life-cycle costing approach to asset management. More
recently this research has extended to the development of deterioration models for unsealed
and sealed local roads as part of a national study across all Australian states.
Biographies to be no more than 150 words for each author. Normal style text.
Copyright Licence Agreement
The Author allows ARRB Group Ltd to publish the work/s submitted for the 25th ARRB Conference,
granting ARRB the non-exclusive right to:
• publish the work in printed format
• publish the work in electronic format
• publish the work online.
The Author retains the right to use their work, illustrations (line art, photographs, figures, plates) and
research data in their own future works
The Author warrants that they are entitled to deal with the Intellectual Property Rights in the works
submitted, including clearing all third party intellectual property rights and obtaining formal permission from
their respective institutions or employers before submission, where necessary.
© ARRB Group Ltd and Authors 2012
14