Cost-benefit analysis of the construction of different flexible pavement structures considering the axle load and type of binder

The status of Brazilian highways reflects a deficient pavement performance when they are subjected to loadings imposed by heavy traffic. Current legislation, as enacted by Contran (National Traffic Council), has increased the axle weight limit for cargo vehicles by up to 10%. Therefore, the aim of this study was to determine a cost-benefit ratio by using different types of structures, asphalt binders and load intensities. Typical pavements were determined and then analyzed by the software AEMC (SisPav) to obtain the horizontal tensile strain (εt) values at the bottom of the asphalt concrete layer and, later, the NFATIGUE value. It was found that the increase in weight, within values covered by legislation, might result in a reduction of approximately 50% in the NFATIGUE value for the pavement structures analyzed. As for economic impact, the same weight increase caused a mean increase of 120% in the cost of repeated loading on pavement structures (R$ NFATIGUE ). It was also observed that structures with more robust asphalt concrete layers can provide the best R$ NFATIGUE -1 ratios. The best results for granular materials were found with thinner layers, associated with a thicker coating. The benefits of modified binders were shown by the analyses of the best structural options: both the polymer-modified binder and the rubber asphalt binder offer significant structural and economic improvements to the structure.


Introduction
High investments in regular maintenance are required for a road network to offer adequate quality for users to drive along with comfort, safety and economy.Roads are exposed to aggressive environmental factors; furthermore, the high volume of traffic and excess loadings by different types of commercial vehicle axles may affect pavement performance.
The current legislation enacted by Conselho Nacional de Trânsito (Contran, 2014) increased the axle weight limit for vehicles by 10%.However, this rule is only valid for vehicles that do not exceed the 5% limit of total gross weight or of gross combined      In general 110 kN) caus the N FATIGUE v the R$ N FATI binders.

Effect of type of
At this poi the effect caus structural and Table 6 pres calculated wit the effect cau assessed N FAT applied by the    f asphalt co AC with mod rubber) cause the R$ N FATIG gs applied by th al binder was r was noted tha the cost per r e different appl inder was repla r, there was esser extent as t-benefit ratio e layer thickn he extreme se nces for the R ases evaluated onal binder, S bber asphalt the sets with considering r-modified as binder, were n evels of axle l t of five cases und in the res e loads evalua repeated in It can also be seen that the asphalt concrete module affects the classification of the data set; for the conventional binder, over 85% of cases have the most resilient moduli (5278 MPa) used when the applied load was 80 kN; the situation was repeated when axle loads were 100 and 110 kN.The first hundred sets obtained by using the polymermodified binder followed the same pattern: more than 70% of the cases showed the greatest resilient modulus adopted in three different evaluated loadings.When the asphalt rubber binder was used, the behavior of the sets changed; the greatest resilient modulus (5278 MPa) appear in less than 40% of the occurrences for the three different loading levels.There is also an intermediate modulus (4.280 MPa) in the AC layer for more than 30% of the cases, for the three loads applied on the pavement.
For the base of simple graded gravel (SGG), when the three different types of asphalt binders were used, smaller thicknesses (associated with higher AC thicknesses) was produced the best costbenefit ratios.A condition of very thick granular layers and low stiffness does not contribute to pavement performance against fatigue cracking.In the case of the asphalt layer with conventional binder, approximately 43% of the occurrences of the first one hundred positions had the thickest granular base used in the data sets (12 cm) for the 80, 100 and 110 kN loads.In addition, it is noteworthy that in the three different loadings, over 30% of the cases had intermediate thickness (15 cm).For the other evaluated binders, the occurrences showed similar behavior to that of the conventional binder, with only minor percentage differences.
Regarding the resilient modulus of SGG, as expected, for the structures of conventional asphalt binder, 53% of events had the highest determined resilient modulus (208 MPa) for graded gravel at three different loading levels applied on the surface.For the sets of polymer-modified asphalt binder and those with rubber asphalt, about 60 and 70%, respectively, have the highest RM adopted for the granular base evaluated for the three different axle loads.
The dry macadam sub-base (DM) showed that good results in the cost-benefit ratio might be obtained with small thicknesses.Approximately 40% of sets with conventional asphalt binder have a thinner determined granular sub-base (16 cm), for loadings of 80, 100 and 110 kN.In the first hundred positions of the sets comprised of rubber asphalt binder, the behavior was analogous to that of the conventional binder, while in sets with polymermodified asphalt binder, it should be noted that approximately 35% of cases have the lower thickness deployed for the sub-base and 35% have intermediate thickness (21 cm).
Regarding the RM of the dry macadam, more than 50% of cases have the greatest resilient moduli determined for the material (300 MPa) in the three loadings for the sets with conventional and polymermodified asphalt binder.For groups with asphalt rubber binder, over 60% of cases have the dry macadam with resilient modulus of 300 MPa.
The subgrade resilient modulus was an important factor in the cost-benefit analyses, and for sets with conventional binder as well as for sets with polymer-modified binder, over 70% of cases have the highest resilient modulus determined (150 MPa) for the three axis loads.For groups with rubber asphalt, 84% of the occurrences presented resilient modulus of 150 MPa for all three loads, and showed a significant increase of over 10% compared to the previously discussed binders.

Conclusion
Based on the numerical simulations carried out in accordance with the methodological procedures of the present study, it was concluded that: a) The increased load applied by the studied axle (TWSA) from 80 to 110 kN (within the load range allowed by the new law enacted by Contran) implies in a reduction of approximately 50% (mean decrease of assessed sets) in the N FATIGUE value.This reduction occurred for the two assessed sets with three different types of asphalt binders used in the composition of the coating layer.The economic impact of this increase also has a significant magnitude, and the same load increase by the TWSA caused an average increase of 120% in the cost per repeated load applied to the pavement.For structures with polymer-modified binder, this percentage was more than 140%.Nevertheless, the increased load from heavy vehicles is a point of no return in the face of technological advances in the automotive industry.The increase in the maximum legal axle load, regulated by the recent legislation, just follows this trend.Road engineers must be prepared for that, and use more robust structures and better asphalt mixtures to adequately support the new road loads.
b) In terms of performance of the modified binders, both the rubber asphalt and the polymermodified binder showed better economic and structural performance over conventional asphalt binder, especially the asphalt rubber binder, which has increased by over 80% the N FATIGUE value in three different loads, and reduced by more than 40% the cost per repeated load as compared with structures with conventional binder.The percentages generated by the structures with polymer-modified binder were also higher, resulting in an increase of over 60% in N FATIGUE and reduction of more than 30% in the cost per repeated load when the applied axle load was 80 kN.c) Among the best combinations analyzed for the cost of repeated load, it was determined that the largest measured asphalt concrete thickness (12.5 cm) is the best option for structures with any one of the three evaluated binders.Regarding the granular layers (base and sub-base), it was observed that lower thicknesses, associated with thicker AC layers, result in better R$ N FATIGUE -1 ratios with the three different asphalt binders.As for stiffness of the materials, it was observed that all layers showed the best cost-benefit results (for three different asphalt binders) when the pavement components were evaluated with their maximum resilient moduli.
Figure 3 modified Cle the low values objectiv most ec ratio w applied of the a Tab the eff N FATIGU most ec types o signific

Table 2 .
1. Section type for Determination o Base c = 22.715 + 0.247 x D 96.82 anical pavement izontal tensile s e layer (εt), the sed (application ctures); it is p y Franco (2007) for the calcu ents, with rout uses the multila mister in 1945 the program J ar. t, i.e., the pa e resilient m ls, the value (N FATIGUE ) we a class I highway of transportation co cost (R$ t -1 ) D Cold: Cb = 22.2 for cost analysis a ost for bituminou Based on IS/ DG 244 + 0.223 x D Ja 14 Cor G number 02 of Janu Paving index FGV anuary, 2009 Marc 224.886 26 rrection factor: 1. v. 38, n. 4, p. 44

Table 3 .
Final un

Table 6 .
Effect N FATIGUE value of M

Table 7 .
Avera the R$ N FATIGU

Table 8 .
Sets of best and worst performances for the R$ N FATIGUE -1 ratio.