High Volume Transport

Vital transport research to ensure accessible, affordable and climate friendly transport for all.

Bituminous surfacings for heavily trafficked roads in tropical climates. The World Bank Regional Seminar on Innovative Road Rehabilitation and Recycling Technologies, Amman, Jordan, 24 – 26 October 2000


Publications with the same themes

View all

PDF content (text-only)

Department For International 1)FID ~~~Development TITLE: by: Bituminous surfacings for heavily trafficked roads in tropical climates H RSmith and C RJones Transport Research Laboratory Crowthorne Berkshire RG45 6AU United Kingdom PA 3334198 PA3334/98 SMITH, H R and C R JONES. Bituminous surfacings for heavily trafficked roads in tropical climates. Proceedings of Institution of Civil Engineers. Transport, 1998, 129, Feb., 28 -33. Bituminous surfacings for heavily trafficked roads in tropical climates H. R. Smith, MPhil(Eng), F1HT and C. R. Jones, MPhil(Eng), CEng, MICE, M1HT Proc. Instn Civ. Engrs Transp., 1998, 129, Feb., 28 -33 Paper 11513 Written discussion closes 19 July 1998 a Increasing numbers of new roads in developing countries are being built with thick asphaltic concrete surf acings to accommodate iacreasing traffic volumes. However, in areas of high traffic stresses, such as climbing lanes and junctions, the use of asphaltic concrete designed by the Marshall method is often not appropriate, particularly athigh pavement tempera- tures. In these situations the commonly- used 75-blow Marshall compaction method underestimates the effect of secondary compaction under traffic and many of these surfacings suffer structural instability leading to severe plastic deformation. This paper illustrates the importance of retain- ing sufficient voids in the mix after traf - ficking to prevent plastic deformation and discusses limitations associated with the Marshall design procedure. Many developing countries have limited facilities for bituminous mix design and an improved method of design is proposed which requires only commonly available or inexpensive equipment. The procedure uses a combination of the Marshall test method and determination of a 'reference refusal density' using a vibrating hammer in accordance with the percentage refusal density test. Keywords: bitumen & tar; research & * ~~development; roads & highways Introduction * ~~~The standard of road networ ks in developing countries has continued to increase; gravel roads and earth tracks have been upgraded to all-weather standard by applying a sealed road surface, existing sealed roads have been strengthened, and new roads have been built with thick asphaltic concrete (AC) surfacings to acconmmodate increasing traffic volumes. lHowever, in areas of high tiafll stiesses, such as climbing lanes and junctions, the use of AC designed by the Marshall miethod is often not appropriate, particularly at high pavement tem- peratures. A fundamental assumption in the Mat-shall procedure is that the density obtained during the test represents the ultimate density of the AC in the road pavement after years of secondary compaction under traffic. This is 28 very difficult to predict and, where loading con- ditions are severe, it is most likely that the commonly-used 75-blow Marshall compaction will underestimate the effect of secondary comn- paction. 2. When secondary compaction of the AC surfacing is underestimated there is a high risk that structural instability may develop and result in plastic deformation. Typical sur- facing failures both on level roads and climb- ing lanes are illustrated in Figs 1 and 2. Such failures are an expensive waste of non-renew- able resources because, in the majority of cases, the deformed asphalt must be removed and replaced with new material. 3.' Plastic deformation of thick bituminous surfacings is by no means a new phenomenon. An OECD review in 19751 referred to wide- spread problems in Europe, North America and japan and noted that, although the problems were more severe in countries with hot cli- mates, hilly terrain and heavy traffic, there were also problems in temperate climates in more sensitive areas such as bends, junctions, etc. The report proposed bituminous mixes that were less susceptible to secondary comn- paction by traffic and suggested methods of designing these mixes that included: lowering the binder content, using a stiffer or a polymer modified binder, and improving aggregate angularity and grading. 4. The importance of maintaining sufficient voids in the mix (VIM) after trafficking was emphasized in the Desert Roads Manu~a1 2which required that heavy duty dense bitumen maca- damis for use in hot arid areas should i etain a VIM of niot less than 30/ at refusal density. Further ~voi- k3showed that the resisi.-.ince to deformtationi of continuously graded mixes increased as the voids in the mineral aggregate (VMA) decreased until refusal density was approached at very low values of VMVA, when resistance to deformation began to decrease rapidly. The initial increase in r esisrance to deformttotion was thought to be caused by the increase in aggaregate contact as the coarser pat ticles~ wexre forced together. llowever, as compaction was increased further, the fines/ filler mnortar gradually filled more of the void space between the coarse particles, eventually reducing aggregate contact and making the mix less resistant to deformation. These results again infer that a minimum level of Harry R. Sm7ithi TransportResearch Laborator r.. Crowthor itc, Berkshire Colin R. Jones. TransportResearchLaboratory,Crow! horne, Berkshire BITUMINOUS SURFACINGS IN TROPICAL CLI[MATES VIM at refusal density would prevent the majority of failures through plastic deforma- tion. 5. D~espite these and other published design modifications for continuously graded mixes in hot climates, there are still far too many occa- sions in the developing world when limited maintenance resources are being used to iesur- face roads that have failed through premature plastic deformation. Full-scale trials and failure investigations 6. The Transport Research Laboratory (TRL), inl cooperation with 1)oth the Public Works Institute in Malaysia (INRAM) and the Institute of Road Engineering (IRE) in Indone- sihas undertaken research to investigate the nature of premature plastic deformation and also developed low-cost mix design procedures that w~ill prevent its occurrence. 7. The research in Malaysia centred around two sets of experimental surfacing trials on a particularly severe climbing lane on0 the Kuala Lurnpur Karak highway, which carried more than 1000 commercial vehicles per dav.i 5~ The climbing lane had a grade of 6%,~ aod the average speed of the commercial vehicles was 1 5 kmi,hll. The studies have shown that the AC wearing course materials designed by the Mar- shall procedure were compacted by the slow- moving and heavy vehicles. F~ig. 3 illusti ates ti is effeci and shows that a rapid decriease in void c' onten t in the surfatcings., of i p1icaillv v~; olccurredI over the first few moot Is. Xltlthis. lhe rai e oif secondary compact in decrecased as the mat erials approached their re fusalI densi- ties. In thits case the refusal denii vt is the iilaxi mom11 density to which each particuliar Inal erial will be compacted by thle traficallesing the cli bihing laiie. S. Further monitoring of these takhas shown hat.i if the level of air voidsil-'ii the s.urfa- cings ca ii be maintained abov'' 3",a t refuisal te-i,then there is a highl protutolity thatl plastic deformation (of thle oltliwill not ociv.C 9.)I lit aiit oi to the research it n~ the results, of further investigations ool heaivily tratlicked orads aiid climbing la11(' ns I the N'liddile Vast aiid East Africa haive c~otfIrlmed these liodiitgs. Ilo these studies tite ntiteil l plpiisof (leforirmed AC surfacIngs w(el c Co~lill):ll-Cd(I toille Piopei-ties (If Iai I,.whicI hll pt d ini satliisfactotrilv 'Itle Alm o ill Fti ., coltnolr tholse fo11llid 1 ,IV sin naiiicl.that if mlixes canl be deindin suct a wa v ithat there is a resicl] von l~d cltntcit[ (If 3%. after compaetioo by trtatfic, tilL surfaci og is uitlikely to fail t hrodgh prinea tire plastic defoi-mntioin. Recent design developments 10. The occurrence of plastic deformation as a result of increased heavy traffic has led to corresponding changes in existing design pro- cedures and also to the development of other design techniques. Most notable of these is the recent change in the Marshall design proce- dure 1' and the development of the snmio~uvi. mnethod (of design as part of the Strategic Highway and Research Programme.' Both these methods emphasize the importance of maintaining a mninimnum level of voids after secondary compaction by traffic. Iii -The Mairshall method of determining the desired bitumen content is, by far the molst commi-on procedure used to design e:ont in 0(i.-1051v graded mnixes in developing countries. It has proved to be t (bust and inexpensive and is enshrined in inost of the general specittcatllllt found in the developing world. The method requires test speciniens, having increasingl bitumen contents, to he moulded at a plite Fig. 1. Deformlotionl oii level r-oad Fig. 2.Iefrai 29 SMITH AND JONES compactive effort under standard conditions. The compactive effort varies according to the traffic level, with a maximum of 75 blows of the Marshall hammer on both sides of the spe- cimen usually being assumed as appropriate for traffic levels of more than one million equivalent standard axles. In the past 8the optimum bitumen content of the mix was then determined by taking an average of the bitumen contents which gave (a) maximum unit weight (b) maximum stability (c) air voids of 4%. To be acceptable the mix also had to comply with a number of empirically-derived criteria, namely, minimum values of stability and flow and VIM and VMA within defined ranges. 12. One disadvantage of this method was that the optimum bitumen content was often higher than that bitumen content which corre- spo nded to 4% air voids and therefore the design mix could possess air voids somewhat less than 4% and often close to 3%. When this occurs the mix becomes more susceptible to plastic deformation, particularly if the compac- tion under traffic is greater than the equivalent of 75 blows of the Marshall hammer. This is often the case on most climbing lanes. 13. In the latest versions of the design manual1 69there are two major changes. First, the introduction of a criterion to limit the voids filled with bitumen, particularly at high traffic levels; and second, a recommendation to select the design bitumen content as that which will result in 4% void content. Both these changes tend to increase the air voids in the design mix and therefore reduce the possi- bility of the surfacing reaching an unaccepta- bly low level of air voids after secondary comnpaction by traffic. 1.4. The 1994 version of the manual also states 'mixtures that ultimately consolidate to less than 3% air voids can be expected to rut and shove' and emphasizes that the design range of air voids (3-5%) is the level desired after several years of compaction by traffic. 1-owever, it is not explicit on how this level of compaction can be simulated in tile test proce- dure, stating that 'the design range (3-5% air voids) will normally be achieved if the mix is dresigned at the correct compactive effort and the air voids after construction is about 8%'. Tlie results from the study in M~alaysia tends to cottflrn this fact with the voids in the wl~lahdecreasing by -3-4,i1 o'-er a period of two years (see Fig. 3). However, the surfa- cing mnaterial outside the wheelpaths is unli- kely to receive appreciable secondary compaction and the binder may be prone to ageing and subsequent cracking. t10.11 10 9 8 7 o06 n0:?4 31 020 400 Dasafter construction Laboratory trials The effect of compaction level chosen for mix design 15. It is clear that the level of compaction used for designing a mix, which is to be sub- jected to severe loading conditions, is fundaL- mental to the long-term performance of the material. Laboratory tests have shown that the selection of a fixed number of blows in the Marshall test is an arbitrary one if there is no prior knowledge of the degree of secondary compaction that will occur under traffic. 16. In the laboratory, samples of an AC wearing course were subjected to 75-, 120- 300- and 600-blow Marshall compaction and also to refusal compaction using an electric vibrating hammer. The vibrating hammer test is based on an extended form of the compac- tion procedure used in the percentage refusal density (PRD) test 12and is discussed more fully later in this paper. 17. Figure 5 shows the relationship between VIM and bitumen content for the EastAfrica *1 Malaysia MiddleEast 0 800 Fig. 3. Reduction in VIM in the wheelpath; wearing course mix designed by Marshall Procedure (75 blows) Fig. 4. Occurrence of plastic deformation in the wheelpath Increased Decreased probability of probabilt;y of plastic deformatio plastic defoirniatton 0 0 0 0 0 0. *** * * 4* 00 0 0 *4*mw0*44 2 * Plastic deformation 0 No plastic deformation 0 0 4 6 8 1 0 Voids in mix: 0 30 BITUMIOUS SURFACINGS IN TROPICAL CLIMATES selected levels of compaction. The results demonstrate that the use of 75-blow compac- tion for design, when secondary compaction in the road is better represented by 300-blow com- paction, will probably result in plastic defor- mation. This is because at 75-blow compaction, 3% air voids is obtained at a bitumen content of approximately 4-7%. If secondary compac- tion is indeed equivalent to 300 blows in the Marshall test then the resulting air voids at a bitumen content of 4-7% will be reduced to approximately 1-8%, that is, considerably less than the 3% criterion. 18. The refusal density obtained in the vibrating hammer test could be adopted as a reference density'. This is because a mix which is capable of retaining 3% air voids at this reference density can not undergo a further reduction of air voids to below 3% under secondary compaction, since it is unli- kely that the vibrating hammer produces a density which is significantly less than the absolute maximum that can be achieved. Bituminous surfacings for severely loaded sites 19. The capacity of a given graded aggre- gate to carry bitumen is controlled by the VMA after compaction. The'effect of VMA on the relationship between VIM and bitumen content when subjected to compaction to refusal is shown in Fig. 6 for six mixes with- different aggregate gradings. 20. The mix which had VMA of only 9-9% can only carry 3% bitumen if it is to retain 3% air voids at refusal density. This mix was typical of a wearing course material and would be too stiff to be workable during construction. As the VMA at refusal density increases, so the bitumen content, which can be carried without reducing VIM to less than 3%, also increases. With information such as that shown in Fig. 6 it is possible to choose an aggregate grading which simultaneously meets dhe requirements of sufficient bitumen for good workability during construction and sufficient voids to maintain a value of VIM1 of 3% at refusal density. 21. For severe sites, the basecourse specifi- ciosgiven in TRL's Overseas Road Note 311:3 are lilkely to produce the most appropriate mix. The aggregate grading of these materials is summrarized in Table 1. 22. E.nsuring that the composition of a mix is correct and that the VIM value will not fall below 3%) is a vital part of the tlcsign process. H-owever, the degree of aggregate interlock and friction betw~een particles also has an impor- twni barn on the resistance to sheica failure of a bitumninous mix. Uncrushed rounded gravel could meet the minimum VIM require- ment when compacted to refusal in a mould, but: aggregate interlock is unlikely to be suffi- cient to prevent shear failure under heavy *Marshatt 75 btows 7 - Marshatt 120 btows O Marshall 300 blows 6 *Marshatt 600 btows J15 - Vibratory hammer W 2- 2-5 3 3.5 4 Bitumen content: 0/ traffic. Additional tests are therefore required to determine the likely performance of asphal- tic surfacings under heavy traffic. Such tests can include 4-5 5 5.5 Fig. 5. Effect of compactive effort on weari~ng course mi~x (a) determinations of mix stiffness moduli (b) wheel tracking rates at temperatures of 450C and 600C. 23. Many developing countries do not have the facilities for such performance testing but do have considerable experience of the perfor- mance of materials designed by the Marshall method. It is therefore recommended that in these cases the normal Marshall design proce- dure, 6using 75 blows on each face, should be completed first to ensure the normal Marshal] design parameters can be met. This, together with past experience, is likely to ensure that Table I. Basecourse grading BS test sieve: mm %~ by mcss passing sieve Asphattic concrete Dense bitumen macadsmi 28 100 100 20 80-100 95-100 14 60 -.0 65 85 10 52-72 6.3 39-55 5 36-56 3 35 32-46 2 36 28-44 18i 20 3.1 (ii6 15 27, 023 10-201 7-21 (-1 5 5 13 0:07- 2-6 2-8 Bitumen gride: pen 80/100 or 60/70* 80/100 or 60/70` Bitumen contein: % 4-8-6 If 5-0 ± 0-5k *60/70 is preferred. tDetermined by Marshall design method. 31 SMITH AND JONES the aggregate being used will be satisfactory in terms of having good particle interlock. Refusal density design for severely loaded sites 24. Some authorities have adopted a proce- dure using increased numbers of blows of the Marshall hammer to design surfacings which will retain a required minimum level of VIM after compaction by traffic. An alternative is to use the vibrating hammer as described earlier. Neither of these methods exactly reproduces the mode of compaction which occurs under heavy traffic; however, the latter procedure is preferred because it allows a degree of knead- ing of the mix which is more representative of field compaction and it is a much quicker method. 25. After the Marshall design binder content, for 75-blow compaction, has been determined, samples are made with bitumen contents decreasing in O.5% increments from the design value. These samples are then sub- jected to vibratory compaction to establish the bitumen content at which 3% VIM is retained at refusal density. Vibrating hammer compaction 26. The equipment and the method of com- paction used in the vibrating hammer test pro- cedure are in accordance with the PRD test.' 2 The sample is compacted in a 152-153 mm dia- meter mould to approximately the same thick- ness as will be laid on the road. Two tamping feet are used, having diameters of 102 mm and 14 6mm. 27. The smaller tamping foot is used for most of the compaction sequence, which involves moving the foot from position to posi- tion to cover the whole of the surface of the saniple. At each position comipaction should continue for between 2 and l0s, the limiting factor being that the material should not be allowed to 'push up' around the compaction foot. The compaction sequence is continued for a total of 2 mi + 5 s. The larger tamping foot is then used to smooth the surface of the 28. Irrespective of layer thickness, a spare base-plate should be used so that the mould can be inverted. The sample is forced to the niewd base-plate with the larger tamping foot and the compaction sequence repeated to ensure that refusal density is, achieved. 29. To ensure that the reference refusal density is obtained in thick layers, it may be necessary to repeat this procedure a second time. It is suggested that trial mixes with a. bitumen content which corresponds to approxi- matelv 6% VIM in the Marshall test are used to (a) determine the mass of material required to o(1 0.9) ' 3. - -- -- - - (9.9) 2 02.5 3 3.5 44! Bitumen content: % give a compacted thickness of approxi- mately the same thickness as for the layer on the road (b) determine the number of compaction cycles which will ensure that the reference refusal density is achieved. Transfer of laboratory design to compaction trials 30. After the standard PRD cycle, samples of basecourse which have been compacted from the loose state can be expected to have densi- ties between 1.5% and 3% lower than for the same material compacted in the road and then cored out and subjected to the PRD test. This is an indication of the effect of the different compaction regime and is caused by a different resultant orientation of the particles. The dif- ferences between the densities for laboratory compaction and field samples after refusal compaction should be measured to confirm whether this difference occurs. 31. A minimum of three trial lengths should be constructed with bitumen contents at the laboratory optimum for refusal density (3% VIM) and at 0-5% above and below the optimum. The trials should be used to (a) determine the rolling pattern required to obtain a satisfactory density (b) establish that the mix has satisfactoi y workability to allow a minimum of 93% of PRD) (standard compaction' 2) to he achieved after rolling (c) obtain cores so that the maximum binder content which allows 3% VILVI to be iretaiined at refusal density can be con- tinnmed. 32. For a given level of compact ion in the Marshall test, VMA reduces to a minimnum and then increases as bitumen content is increased. However, samples compacted to refusal density will have sensibly constant values of VMA over a range of bitumen contents before the 5 5 5.5 Fig. 6. Relationships between mix properties at refusal density 32 BITUMINOUS SURFACINGS IN TROPICAL CLIMATES aggregate structure begins to become 'over- filled' and VMA increases. This means that during the trials it will be a relatively simple matter to determine the sensitivity of the mix to variations in bitumen content and to confirm the bitumen content required to give a minimum of 3% VIM at refusal density. If necessary the aggregate grading can be adjusted to increase VMA which will reduce the sensitivity of the mix to changes in bitumen content. 33. A minimum of 93% and a mean value of 95% of the standard PRD density is recom- mended as the specification for field compac- tion of the layer. From these trials and the results of the laboratory tests, it is then possi- ble to establish a job mix formula. After this initial work, subsequent compliance testing based on analysis of mix composition and refusal density should prove to be speedy, especially if field compaction is monitored with a nuclear density gauge. This initial procedure is time-consuming, but is justified by the long- term savings that can be made by extending pavement service life and minimizing eventual rehabilitation costs. 34. It is essential to seal mixes designed by this method with a surface dressing. This greatly reduces the risk of premature 'top down' cracking associated with bitumen age hardening. 10,11This is important not only during the period when secondary compaction occurs in the wheelpaths but also for long-term protection of those areas which will not be traf- ficked and which are likely to retain air voids above 5%. Summary 35. Research has shown that the risk of plastic deformation in asphaltic concrete surfa cings on severely loadcd sites can be minini mized if VIM of at least 3%/ can be retained after secondary compaction by traffic. 36. A methodology which combines two standard test procedures has been proposed for the design of bituminous sur facings for such sites in developing coulitries wvith tropical environments and where, a- is commonly the case, equipment for mix dosign is~ limited. While the procedurc dscsuibcd wvill bring immediate benefit to maw' road projects, improvements in the methodology can be expected with the further introduction of suita- blec laboratory mix perforiiancc tests. References1. ORGANIZATION FOR ECONOMIC COOPERATION AND DEVELOPMENT. Resistance of Flexible Pavements to Plastic Deformation. OECD report, 1975. 2. HALCROW. Desert Roads Manual. Sir William Halcrow & Partners, 1st edn, 1980. 3. COOPER K. E., BROWN S. F. and POOLEY. G. R. The design of aggregate gradings for asphalt basecourses. Proceedings of the Association of Asphalt Paving Technologists, 1985, 54, 324-346. 4. HIZAm H. and JONES C. R. The performance of polymer modified asphaltic concrete on climbing lanes in Malaysia. Proceedings of the 16th ARRB Conference, Melbourne, November 1992. 5. HiZAm H. and MOROSIUK G. A study of the perfor- mance of various bituminous surfacings for use on climbing lanes. Proceedings of the 8th REAAA Conference. Taipei, April 1995. 6. ASPHALT INSTITUTE. Mix Design Methods for Asphalt Con crete and Other Hot-mix Types. Asphalt Institute, Manual Series No. 2 (MS-2), 6th edn, Lexington, 1994. 7. COM.IN~SKY R. J. The SUPERPA VE Mix Design Manual for New Construction and Overlays. Strategic Highway Research Programme, National Research Council, SHRP-A-407, Washington, 1994. 8. ASPHALT INSTITUTE. Mix Design Alethods for Asphalt Concrete and Other Hot-mix Types. Asphalt Institute, Manual Series No. 2 (MS-2), Lexington, 1988. 9. ASPHALT INSTITUTE. Addendum: Mix Design Methods for Asphalt Concrete and Other Hot-mix Types. Asphalt Institute, Manual Series No. 2 (MS-2), Lexington, 1992. 10. RoL.T j., SMITH H. R. and JONES C. R. The design and performance of bituminous overtays in tropi- cat environments. Proceedings of the 2nd Interna- tionial Conference on the Bearin.g Capacity of Roads and Airfields, Plymouth, UK, 1986. 11. SMITH H. R., ROLT J. and WAMBURA j. H. G. The durabitity of bituminous overtays and wearing courses in tropical environments. Proceedings of the 3rd International Conference onl the Bearing Copacity of Roads and Airfields. Trondheim, Nor way, The Norwegian Institute of Technology, 1990. 12-ts BIT SH STAND)ARDS INSTITUTIiiON. lAlethods of Test for the Determination of DensitY and Compaction. BS 598 Part 104, BSI, 1989. 1 3. TREANSPORT RESEARCH LAOAI)YA Guide to the Structural Design of Birni ulcdRoads inTrpical and Sub-tropical Coi~ii,iies. Overseas R~oad Note 31, Transport Rese.ca; La boratory, Crowthorne, 1993. 33