You are on page 1of 6

DETERIORATION COST DUE TO CAMBER FOR CHIPSEALED PAVEMENTS OVER

GRANULAR BASES
Jacobus Daniel van der Walt *
PhD. Candidate, Department of Civil & Natural Resources Engineering, University of Canterbury.
* 20 Kirkwood Ave, Upper Riccarton, Christchurch,8041, New Zealand jdv18@uclive.ac.nz

Dr. Eric Scheepbouwer


Director of Contraction Management, Department of Civil & Natural Resources Engineering, University of
Canterbury, Christchurch, New Zealand

Dr. Bryan Pidwerbesky


General Manager – Technical, Fulton Hogan Ltd, Christchurch, New Zealand

Dr. Brian H.W. Guo


Lecturer, Department of Civil & Natural Resources Engineering, University of Canterbury, Christchurch, New
Zealand

ABSTRACT: Pavement design assumptions suggest that traffic loading on pavements is evenly distributed
between the wheel paths. The practice of including a slight camber, commonly up to 4% to assist runoff is
assumed to have negligible effects concerning vehicle loading. However, it has been shown that a slight camber
or crown can cause unequal loading in the wheel paths. The New Zealand Long Term Pavement Performance
data shows that there is a significant difference between the amount of rutting occurring in the outside compared
to the inside side wheel path. This research will compare the pavement deterioration models of the outside and
inside wheel paths individually. It will conclude that this difference in rutting is due to cumulative damage as a
result of the difference in loading due to camber. The results show that due to camber there is a higher likelihood
that the outside wheel path will experience accelerated rutting before the inside wheel path over time. This result
forces a re-think of anecdotal homogeneous pavement design assumptions and maintenance procedures.
KEY WORDS: Camber, crown, rutting, load distribution, rut progression, difference in rutting

1. INTRODUCTION

A rut is a progression of longitudinal depressions along the wheel paths which is mainly caused by progressive
movement or consolidation of materials due to repeated loading [1]. Rutting can occur in the surface, subsurface
and subgrade layers or in a combination of these layers depending on their strength and magnitude of the load
applied. Because chip-seal surfaces are thin, rutting in chip-sealed roads is commonly caused by mechanical
deformation in the underlying pavement layers [2].Rutting is one of the most widely used pavement performance
indicators for thin flexible pavements. It is important for a number of reasons. Firstly, it is important from a
safety point of view. For safety reasons the recommended rut depth should not exceed 11mm [3] to 25mm [4]
based on road geometry, road class, surface type and vehicle speed. This could lead to an increased risk in water
ponding which could lead to hydroplaning, loss of control and insufficient breaking distances. To help drain
water off the road, the use of camber, cross-fall or crown of approximately 1-3% is common in pavement
designs. Secondly rutting is an important performance indicator and trigger mechanism in Pavement
Management Systems (PMS). In most PMS systems rut depth or some other form of rutting index is used as part
of a trigger criterion for future intervention and maintenance[5]. Rutting is measured using the worst affected
area, or the deepest rut of the two wheel paths. [6, 7]. Lastly, rutting is commonly used as design criteria in
pavement design processes. In pavement design, the recognised terminal rut depth seems to be 20-25 mm [8].
According to Austroads design handbook the critical trigger depth of 20 mm is assumed to be the failure point in
the subgrade for thin flexible pavements [9, 10].

Page 1
In most pavement design methods, it is assumed that each wheel carries 50/50 of the axle load and there are no
allowances put forward for a difference in load in the wheel paths. Municipalities and highway agencies set out
maximum axle and truck loads. The wheel loads are assumed to be 50% of the maximum axle load. From
previous research [11] it was found that due to camber a significant amount of load is shifted to the outside
wheel path. This can substantially increase the difference in the number of ESALs experienced by inside and
outside wheel path as will be disused in the next section.
1.1 Calculating and converting the damaging effect of camber and rutting
Observations of pavements in Canterbury, New Zealand indicate that rutting occurs more in the outside wheel
path than in the wheel path closest to the middle or crown of the road. Prior research [11] shows that this effect
is not as minimal as is normally assumed. Using the free body diagram (Figure 1) it can be shown that the effect
of camber would cause significantly more load in the outer than the inner wheel path. In equation 1, the load
ratio between inner and outer wheels, is shown to be a function of camber (θ), the height of the centre of mass of
the vehicle (H) and the width of the wheelbase (C).

Figure 1(left). Static free body diagram of Figure 2 (right). Deterioration phases of granular pavements with
exaggerated camber effects thin surfaces, Adopted from [12]

P" cos 𝑎 − 𝜃 2𝐻
= , 𝑤ℎ𝑒𝑟𝑒, 𝑎 = 𝑡𝑎𝑛4" ( ) (Eq 1)
P# cos 𝑎 + 𝜃 𝐶

To convert this ratio of loads into a ratio of loads expressed in ESALs an exponent law is used as suggested by
Austroads. It states that the damage from each axle is calculated by dividing the respective axle load by the
reference load which is given for each axle configuration, and lifting the quotient to the EXP power [9, 10]. For
more than 40 years the power of 4 has commonly been used. It has been shown that the ratio of ESALs
experienced by the outside wheel path compared to the inside wheel path could be as high as 1.9 using the
example of a typical milk or log truck and the standard damage exponent of 4 given by Austroads [11].
1.2 Rutting deterioration Phases
It has been widely accepted that pavement deterioration can be broken up into three main stages as shown in
Figure 2. This also holds true for rut progression.
While many methods exist that allow for predictive capability (HDM-4 models and work done by Martin[12]).
Henning et al. [13] conducted research that showed a new approach for modelling rutting for New Zealand
conditions. Their research follows the three stage modelling approach. The first is a simplified model to predict
Page 2
the initial densification/ consolidation. The second is a linear model to predict the progression of rutting during
the relatively stable rut stage. Finally, a third model is used to predict the probability of a pavement undergoing
accelerated rut progression associated with failure of the pavement [13]. This is arguably the most important
element of their research. This knowledge allows the use of probabilities in calculating when intervention should
take place. Henning et al.’s probabilistic model for accelerated rutting is shown in Equation 2. Accelerated
rutting is defined to have occurred when the rut rate reaches twice the expected stable progression rate. In New
Zealand the stable rut rate is considered between 0.4 to 0.6 mm per year [14].
"
P Rut =>>?@ = (Eq 2)
" A?BC 4D.FGH∗"JKL ∗MN=O A#.PQP∗NRS – S.U.V.

where:

ESAL = Number Equivalent Standard Axels

SNP = Ajusted Structural Pavement Number

PTF = Pavement thickness facor. For base layer thickness < 150 mm, PTF =
4.426 & For base layer thicknes > 150 mm, PTF = 0.4744

Both Phase 2 and Phase 3 models were calibrated against test data from the Canterbury Accelerated Pavement
Testing Indoor Facility [13], which allows deterioration testing under controlled conditions [15, 16].

1.3 Scope and objectives


This research will show that there is a cost of using camber in terms of unequal pavement deterioration between
the outside and inside wheel paths.

• Using the New Zealand LTPP database, the differences in outside and inside wheel path rutting will be
investigated.
• Specific investigation into Phase 2 and Phase 3 rutting be conducted.
• Using the above points, it will be shown that due to camber there is a higher likelihood that the outside
wheel path will fail before the inside wheel path over time.
2. METHOD

All available LTPP sites that had had no maintenance except for emergency repairs (‘sterile’ sites)) in
Canterbury New Zealand have been used for this research. The New Zealand LTPP programme was established
to record accurate pavement data starting early 2000s. The programme monitors a number of 300 m long
sections of roadway across New Zealand. These sections (called calibration sections) together form a
representative sample of the New Zealand road network. Two main types of LTPP sections exist, sterile and
non-sterile. Since the sterile sections became part of the LTPP network, no maintenance has been performed on
them other than safety related maintenance, whereas non-sterile section experience normal maintenance. This
research will focus on sterile sections to limit variability introduced by maintenance. The LTPP condition data
contains the rutting data; this data is measured by a purpose built tool that is designed to measures the rut on
both wheel paths to an accuracy of ± 0.2mm resolution [17]. More details on the used set up can be found in [17,
18]. The LTPP contains the outside and inside wheel path rutting data from test sites collected over a number of
years. This data has been used to construct Figure 3.

Henning et al.[13] model for accelerated rutting gives the probability of accelerated rutting. The model was
calibrated against the CAPTIF data where load was not controlled by camber but precisely controlled by
researchers. Therefore, it is assumed that Henning et al. model [13] for accelerated rutting has no effect of
camber in it. Using the increased ESAL calculations based on loads from Equation 1 as an input into the
accelerated rutting model from equation 2 will result in the data, shown in Figure 4.
Page 3
3. RESULTS/ ANALYSIS

Figure 3. Comparison of left (outer) and right (inner) wheel path for multiple LTPP sections. Note that the y
axes are different.

Figure 3 shows a direct comparison of max, mean and min of the inside (LWP) and outside (RWP) wheel paths
for the sterile New Zealand LTPP sections. It is clear from these figures that there is significantly more rutting
occurring in the outside wheel path (LWP) compared to in inside wheel path (RWP). This effect is more
pronounced when looking at the max rut depths than the mean and min comparisons.

Page 4
Figure 4a (Left). Illustrates Phase 2, expected rut depth at 1, 2, 3 % camber for thin and thick pavements.
Figure 4b (Right). Illustrates the mean probability of accelerated rutting for standard, LWP and RWP at camber
of 1, 2, and 3 % for thin granular pavements

Figure 4a illustrates how camber can have an effect on stable rut progression, Phase 2. This figure shows that
there is a difference between inside and outside wheel path’s stable rut progression. It is clear that this effect is
more apparent on thin granular pavements with a relatively high stable rut progression rate compared to thicker
granular pavements with lower stable rut progression rate.
From figures 4b it can be noted that the probability of accelerated rutting in the outside wheel path is greater
than the probability of accelerated rutting of the inside wheel path. As camber increases from 0% to 3% this
difference in probability increases. As accelerated rutting is occurring more often in the outside wheel path,
maintenance would need to be done more regularly in the outside wheel path than if there was no camber. It is
also important to note that as camber increases, the likelihood of the inside wheel path experiencing accelerated
rutting decreases. This means in reality less maintenance would need to be conducted on the inside wheel path
over time. This finding matches trends seen in the LTPP data.
4. DISCUSSION

Dawson [21] highlighted limitations of the Austroads power law used to calculate equivalent damage [10]. The
presented methodology uses the power rule with the exponent 4, however numerous authors reasoned that for
thin lightly trafficked granular pavements the exponent value would probably need to be larger than 4, with
values or around 7 used by various studies [19-21]. Using a higher exponent would increase the unequal
damaging effect due to camber.
5. CONCLUSIONS

The New Zealand LTPP data shows that there is a significant difference between rutting of the inside and
outside wheel paths. Due to having a non-horizontal pavement surface (or camber) the difference in wheel load
has been calculated using a simple free body diagram. It shows that there is a significant difference in load
experienced in the outer and inner wheel paths. As rutting is related to loading, it stands to reason that camber is
significantly responsible for the found difference in rutting. By adopting this point, the individual wheel path
rutting has been predicted using models developed in New Zealand. Results from this model show that as
camber becomes steeper, it significantly increases the probability of accelerated rutting occurring for outside
wheel path and decreases the probability for the inside wheel path. This will result in the outside wheel path
failing sooner and governing maintenance procedures over time. This result forces a re-think of anecdotal
homogenous pavement assumptions and methods to extend pavement life.

Page 5
REFERENCES:
[1] Tarefder, R., M. Zaman, and K. Hobson. A laboratory and statistical evaluation of factors affecting
rutting. International Journal of Pavement Engineering, Vol. 4, No. 1, 2003, pp. 59-68.
[2] Gransberg, D. D., and D. M. James. Chip seal best practices NCHRP. Transportation Research Board,
Washington, D.C, 2005.
[3] Lay, M. Handbook of Road Technology. Volume 2: Traffic and Transport. Gordon and Breach Science
Publishers, London (England), 1998.
[4] Ong, G. P., H. R. Pasindu, and T. F. Fwa. Critical Rut Depth for Pavement Maintenance Based on
Vehicle Skidding and Hydroplaning Consideration. Journal of transportation engineering, Vol. 138, No.
4, 2012, pp. 423-429.
[5] Robinson, R., U. Danielson, and M. Snaith. Road Maintenance Management-Concepts and Systems.
1998.
[6] Cenek, P., R. Henderson, M. Forbes, R. Davies, and A. Tait. The relationship between crash rates and
rutting January 2014. 2014.
[7] Hicks, R., J. Moulthrop, and J. Daleiden. Selecting a preventive maintenance treatment for flexible
pavements. Transportation Research Record: Journal of the Transportation Research Board, No. 1680,
1999, pp. 1-12.
[8] Pidwerbesky, B. A holistic approach to flexible pavement design improves performance.In ARRB
Conference, 26th, 2014, Sydney, New South Wales, Australia, 2014.
[9] Austroads. A Guide to the Structural Design of Road Pavements. Austroads Publication No. APGP17/04.
Austroads, Sydney, 1992.
[10] Austroads. 2004a Pavement Design-A Guide to the Structural Design of Road Pavements. Revision 2.In,
AUSTROADS, Sydney, NSW , Australia, 2004.
[11] van der Walt, J., E. Scheepbouwer, and S. Tighe. Differential rutting in Canterbury New Zealand, and its
relation to road camber. International Journal of Pavement Engineering, 2016, pp. 1-7.
[12] Martin, T. Pavement performance prediction deterioration model development: data review and
calibration of HDM-4 road deterioration models. AUSTROADS Project No: BS. AN, Vol. 552, 2003.
[13] Henning, T., R. Dunn, S. Costello, and C. Parkman. A new approach for modelling rutting on the New
Zealand State Highways. Transport Research: A Journal of Australian and New Zealand Research and
Practice, 2009.
[14] Henning, T. F., S. Costello, T. Watson, and N. Land Transport. A review of the HDM/dTIMS pavement
models based on calibration site data. Land Transport NZ, 2006.
[15] Alabaster, D. J., A. W. Fussell, and N. Land Transport. Fatigue design criteria for low noise surfaces on
New Zealand roads. Land Transport New Zealand, 2006.
[16] Pidwerbesky, B. D. Accelerated dynamic loading of flexible pavements at the Canterbury accelerated
pavement testing indoor facility. Transportation research record, No. 1482, 1995, pp. 79-86.
[17] Brown, D. Measuring pavement condition data for the establishment of a long-term pavement
performance study on New Zealand roads.In International surface friction conference: roads and
runways: improving safety through assessment and design. Christchurch, New Zealand, 2005. pp. 1-4.
[18] Henning, T., and D. Roux. Pavement deterioration models for asphalt-surfaced pavements in New
Zealand. NZ Transport Agency research report 367, Wellington, New Zealand 2008.
[19] Jameson, G. Origins of AUSTROADS design procedures for granular pavements, Report ARR 292.
September, ARRB Transport Research Ltd., Vermont South, Australia, 1996.
[20] Dorman, G. Design curves for flexible pavements based on layered system theory. Highway Research
Record, Vol. 1, No. 71, 1965, pp. 69-84.
[21] Dawson, A. Rut accumulation and power law models for low-volume pavements under mixed traffic.
Transportation Research Record: Journal of the Transportation Research Board, 2008, pp. 78-86.

Page 6

You might also like