: ASPHALT! MIX DESIGN — REFUSAL DENSITY. APPROACH FOR ‘ | HEAVILY TRAFFICKED ROADS * — Dr. S.K. Rao®, JK. Dast* & Prost Rovenowornmy¥** ABSTRACT Asphalt paving mixes designed by the Marshall method have been failing prematurely on our roads, One of the reasons for such foitures is inadequate initial compaction. Densities achieved under 75-blow Marshall compaction in the laboratory do not sirtulate the field densities of the mix after it has undergoue secondary compaction due 10 traffic. When the air voids in the mix decrease to below 3 per cent during such densification and as the viscosity of asphalt in the mix decreases sharply in summer, the mix permanently deforms as a rut under the wheel loads Three factors contribute to good performance of an asphali mix carrying heavy axle loads in hot climates. They are adequete initial compaction so that secondary compaction uader traffic is minimized, suificient asphalt content for durability of the mis and enough air voids in the mix for its stability. AU the three factors axe influenced by the VMA of the mix. A high VMA would permit the incorporation of higher asphalt content while ensuring enough air voids under increased compaction. Aggregate shape and surface texture influence the VMA to some extent but it is largely influenced by the aggregate grading. Dense gradings give rise to low VMAS and open gradings to high VMAs. The Paper suggests a modification to the Marshall design procedure we follow for the design of asphalt mixes. The modification involves adjusting the oggregate grading 10 achieve higher VMA values for creating more space to incorporate higher asphats coments and checking the mixes for ‘refusal density” for ensuring its stability under secondary compaction due 10 traffic 1.INTRODUCTION (a) select an aggregate grading fitting the grading band of the MoSRT&H specifications (b) cast a set of Marshall test specimens of asphalt mixes with aggregates raded to the selected grading and with varying asphalt contents (c) test the specimens for the Marshall Most of the important highways of the country are surfaced with asphalt, Bituminous concrete (BC) and dense bituminous macadam (DBM) are the commonly used asphalt courses. Mix designs for BC and DBM are based on the guidelines given in the Asphalt Institute (USA) Manual MS-2 ‘Mix Design Methods for Asphalt Conerete and Other Hot-Mix Types’. Several of the mixes designed by the Marshall test procedure in accordance with MS-2 have failed prematurely on our roads, sometimes within the first few months of construction. The object of this paper is to scrutinize the asphalt mix design procedure as practiced in our projects, point out why it falls short of expected performance and suggest a few modifications that could avert such failures and enhance the performance of the mixes. 2. ASPHALT MIX DESIGN PROCEDURE ‘The steps involved in the mix design procedure are properties and (d) choose the asphalt content offering the optimum Marshall properties in accordance with the criteria suggested in MS-2, as the design asphalt content. A design example is given below to show the limitations and deficiencies of this procedure as applied to our roads. 2.1, Asphalt Mix Design Example A gist of mix design results for bituminous concrete intended for a pavernent in a high-density corridor of a National Highway is given below. The MoSRT&H grading band and the adopted grading that lies within the band shows in Fig. 1. The Marshall Test properties of the mixes made with varying asphalt contents. The Marshall properties are per cent * Senior Consultane '* General Manager 66 Managing Dinctor [email protected] 4+ Written comments on this Paper are invited and wall be received upto 31% Augast LEA Associates South Asia Pvt. Ltd, Mathura Road, New Delhi 00734 Dr. Rao, Das & ROYCHOWDHURY ON air voids (Va), per cent voids in mineral aggregate (VMA), per cent voids filled with bitumen (VFB), unit weight, stability and flow. Shown in Fig. 2 to7.Marshall ‘n criteria as adopted by the MoSRT&H Specifications are given in Table 1. 2 E Damen Content) Fig. 4 & zg Fig. 1 Anerezate Grading fn the Design Example i a0 445 ses os 0 Bitumen Content) Fig. 5 E Z sw00 £ é : zim 1600 g ae & soo 40 45 so 58 60 3 1400 Bitumen Content (2) 1300 2 % 7 3 120 Fig. 2 i Te a ee 180 5.00 - - 450 F400 aot 7 140 E 350 Ht e zB 300 Tae a enne + Z 250 | 2.00 Y L B09 ta Coo +++. 40 45 50 35 60 Z ant ‘cai ne 0, miiuanes Content (95) Bitumen Content (%) Fig. = Fig.7ASPHALT Mix Deston - Reusat Densiry Apekoacn For HEavi Tearricken Roavs 55 se A. MasuaLt Mix DasiGn CRITERIA ACCORDING. ro MeSRT&H Sreciricai0ss Mix Criterion Heavy Trafic Min Max Compaction elfon (no, of blows on each face of specimen) w % Stability (KN at 60°C) 9 Flow (mm) 2 4 % air voids z 6 “% voids in mineral age (VMA) paraie table® % voids Filled with bitumen (VFB) 6 E © VMA for DBM & BC mixes varies benoeen 10 per cen and 16 per cent depending on the nominal mas size af aggregate and the design air void content 2.2. Choosing Design Asphalt Content Initially, the asphalt content corresponding (0 the median of percent air voids limits as shown inTable 1. is selected from Fig.2 to examine whether all other Marshall properties according to the design guidelines in Table. are fulfilled. If this is so, this asphalt content is taken as the initial design asphalt content. If the criteria are not met, the mix is re-examined for its suitability at other asphalt contents or it may have to be redesigned by selecting a different aggregate grading. In the given example, the initial design asphalt content is 4.8 per cent (corresponding to 4.5 per cent air voids). This is then compared with the minimum asphalt content stipulated in the MoSRT&H Specifications (for bituminous concrete). If it is more, it is taken as the design asphalt content. If itis less, it has to be examined whether all the Marshall test criteria are fulfilled at the MoSRT&H minimum asphalt content. If this is so, the MoSRT&H minimum is selected as the desiga asphalt, content. If not, the mix has to be redesigned, changing, the aggregate grading. It may be mentioned, however, that the Marshall method of MS-2 does not stipulate any minimum asphalt content as the design requirement. In the design example, since the initial design asphalt content (4.8 per cent) is less than the MoSRT&H minimum (5 per cent) and since all the Marshall test criteria are fulfilled at the latter the design asphalt content would be selected as 5 per cent 3. WHY DO OUR ASPHALT MIXE! UNDER-PERFORM? If the mix design formulated in the design example is submitted for approval it would be accepted for implementation since it satisfies the requirements of our specifications. But mixes having the same Marshall characteristics as in the example have been found to have undergone premature permanent deformation under our traffie loadings, particularly during the summer months for the following reason. In the Marshall procedure of MS-2, the laboratory compaction is intended to simulate the in-place density after the mix has endured several years of traffic! At the time of developing the method, heavy traffic was regarded something close to I MSA (one million standard axles) and a laboratory compuction of 75 blows on the two faces of a Marshall specimen was discerned to achieve the corresponding density level. In the high density traffic corridors of the country, loadings of that ‘magnitude occur within a few months of opening to traffic and the in-place densities increase beyond the laboratory densities with further movement of ti increases, air voids in the mix decrease. As the air voids decrease below a certain limit (around 3 per cent) the asphalt film enveloping the aggregate particles tends to push them apart, reducing the aggregate particle contact. This action becomes more and more predominant as the pavement temperatures rise in summer and decrease the viscosity of asphalt. Because of the falling viscosity of asphalt and the diminished aggregate particle contact, the mix loses its shear resistance and deforms as a rut under the vehicle wheel with an accompanying heave on the side. It may be realized that the shear resistance of an asphalt mix is contributed jointly by the viscosity of asphalt and aggregate internal friction. Photo. shows this kind of rutting observed on a newly constructed road within two months of its construction. Photo.2 shows the cores of bituminous conerete wearing course taken at three locations on a cross section of this deformed roud. The three locations are (a) in the rut, (b) on the Photo 1, Asphalt Mix Rutting on a newly laid Pavement56 Dr. Rao, Das & RoycHowonury on accompanying heave and (c) at an un-rutted location on the same cross section. The as-laid thickness of the BC layer is 51 mm (un-rutted location) and this thickness has reduced to 38 mm in the rut und increased to 86 mm. in the heave. The disturbed aggregate structure could be observed in the cores Pavement Rutting Due to Weak Asphalt Mix RUTTING IN THE BC LAYER Phove 2, Cores ofthe Deformed BC hat the probability of plastic deformation (rutting) of an asphalt mix is very high when the in-place air voids fall below 3 per cent has been observed in several countries experiencing high pavement temperatures as revealed in Fig. 8°. In several of our project roads, which have experienced such rutting, the in-place air voids (determined from the cores taken from the site) have been observed to be lower than 3 per cent. These observations emphasize the fact that the air voids level in the mix should remain above 3 per cent, if plastic deformation is to be avoided. In fact, MS-2° warns ‘it has been shown that miatures that ult to less than three per cent air voids can be expected to rut and shove if placed in heavy traffic locations” | | ee | ately consolidate setae £1244 ° rr ewdind 3 °° 4” | Fig. 8. Probability of Plastic deformation in Asphalt Wearing, Course in Relation to Air Voids Souree: A Guide to the design of hot mix asphalt in tropical and sub-tropical countries. Resistance to premature rutting can however, be achieved if we reduce the asphalt content and design the mix for an initial air void content closer to the upper limit. Butsuch mi rden quickly leading to top-down cracking, a common phenomenon observed on our rouds. Brittleness of asphalt, premature cracking, raveling and stripping are the types of disiresses that are associated with low asphalt content and high air voids content in the mix. This. the primary reason why a minimum asphalt content has been stipulated for the asphalt mixes in the MoSRT&H Specifications. In essence, the reasons for the poor performance of asphalt mixes on our project roads could be attributed to one or more of the following characteristics of the mixes. . * Inadequate initial compaction making the mix vulnerable to high secondary compaction under watfic. © Relatively high asphalt contents that permit the reduction of air voids to less than 3 per cent under secondary compaction, leading to rutting under heavy axle loads when pavement temperatures rise in summer, © Low asphalt contents and high air voids in the mix leading to top-down cracking, raveling and stripping making the mix less durable. 4, HOW CAN WE IMPROVE THE MIX DESIGN PROCEDURE TO ACHIEVE BETTER ‘ORMANCE? An asphalt mix designed to serve as a wearing course in a heavy-duty pavement should possess several desirable characteristics, chief among them, are high resistance to deformation, high resistance to fatigue and g00d workability to allow adequate compaction. For designing a mix having all these characteristics we need to address conflicting design indicators. For example, dense and durable mixes with high asphalt content are prone to plastic deformation, but more open graded mixes designed to resist plastic deformation are vulnerable to accelerated bitumen oxidation and hardening and are therefore Jess durable. Asphalt mix design is thus a delicate process requiring a good understanding of the mix behavior besides accurate laboratory test procedures and analysis. But the mix design procedure in MS-2 is not comprehensive in this regard, emphasizing, as it does, on the selection of asphalt content rather than on otherAsrnaur Mix Desicr - Rerusat. Dexsiry Aprsoacn For Heavity Trarricxep Roaps ST parameters like the material properties and aggregate grading characteristics. The limitations of this method have since been realized by the formulators of the method, the Asphalt Institute of USA. It has now come up with an advanced method of mix design, called the “SUPERPAVE” (Superior Performing Asphalt Pavements) Mix Design‘. The Superpave system incorporates material characterization, traffic related level of compaction and a mix design and analysis procedure whose aim is to achieve improved performance by controlling rutting, fatigue cracking and low temperature cracking. The Superpave system requires special testing equipment for testing the physical properties of asphalt and aggregate and a gyratory ‘compactor for compacting the mixes. It may be eventually necessary to switch over to this method of mix design for achieving better performance of our roads but such a change may take some time as the complex te: equipment used in the method, besides being costly, requires training the professionals for conducting the tests, to a certain degree of precision and reliability. Till such time it is expected that the Marshall method would continue and it is therefore, necessary to devise modifications to the method to achieve better performance of the mixes. Under similar requirements, pavement investigators working for the Transport Research Laboratory (TRL), UK have come up with a modification, which is based on their observations of asphalt mix performance in hot climates under heavy axle loads*®"*, This modification consists in ensuring that a mix designed by the Marshall method with 75- blow compaction (say, for 4.5 per cent air voids) also does not compact to lower than 3 per cent air voids when compacted further to ‘refusal density’. Refusal density is a state of compaction where the density does ‘not change with further compactive effort. Without knowledge of the degree of secondary compaction under traffic loads, selection of a level of Marshall compaction to represent it becomes arbitrary. In comparison, compaction to refusal density, retaining a minimum air voids level of 3 per cent, ensures that the air voids would remain above that level under the severest compaction. Usually a compactive effort of around 300 blows on each face of the Marshall specimen is required to reproduce that state. ‘The effect of laboratory compaction on the air voids in the asphalt mix shows in Fig,9. A 200-blow compaction, which might fairly represent the level of secondary ‘compaction under the expected traffic, has brought down the air voids in the design example mix from 4 per cent (o 2.1 per cent at the selected design asphalt content of 5 per cent. That this mix in the field actually deformed within the first two months of its laying, achieving air s levels as low as 1.5 per cent establishes the inappropriateness of using the standard Marshall ‘compaction for designing the asphalt mixes for our roads. ‘Air Volas 09 uw 4s so ss “ Bitumen octet (4) [= ti apacion —O=10 fn ompncton Fig, 9. Effect of Laboratory Compaction on Air Voids 5. DESIGNING RUT RESISTANT ASPHALT MIXES How is it possible to achieve 3 per cent air voids level at refusal density while designing the mix with 75- blow Marshall Compaction? It is possible in two ways. (A) By reducing the asphalt content and designing the mix with a higher initial air voids content (say, close to the upper limit); (B) By choosing an aggregate grading with higher voids ratio, that is, one which results in a higher VMA (voids in the mineral aggregate) of the mix. But we have observed that mixes of Type A would be less durable and prone to premature cracking and raveling. For generating mixes of Type B we would have to study the aggregate grading characteristics and select a grading that yields higher VMA values, keeping in mind the physical properties of aggregate that also influence VMA. 5.1. Composing Aggregate Grading for the Asphalt Mix For adequate stability and sustained performance (durability) an asphalt mix must have ‘enough’ bitumen to bind the aggregate particles with adequate film thickness and also have ‘enough’ air voids to avoid failure by plastic deformation, This implies that the mix should have higher YMA (which is the space occupied by bitumen and air voids in the mix) so as to have room for both ‘enough? bitumen and ‘sufficient’ air voids. A high VMA value can be achieved by choosing aggregates with good physical properties, that is, of angular shape and rough surfacetexture and an appropriate blend of aggregate size fractions, that is, a suitable aggregate grading. Usually, the designer aims at the mid-point aggregate grading of the MoSRT&H specified aggregate grading, band. There is no basis for judging the suitability of an aggregate grading, except that a mix made with it should, satisfy the Marshall Design criteria, To examine the VMA generating capacity of an aggregate grading we need t0 compare it with the 0.45 power grading, also known as the PRA grading, representing the percentage passing a given sieve size d = 100*(d/D)*, where D is, the maximum size of aggregate of the selected grading. ‘This grading, shown in Fig. 10, is regarded as the densest grading and has been adopted as the reference grading for the Superpave mix design. Gradings that closely follow the 0.45 power grading give rise to low VMA values, as, does the grading in the design example, shown plotted in, Fig. 10. Gradings that plot above or below are capable of ‘generating greater VMA values. A grading which plots, above are a finer grading, that is, the percentage passing any sieve size is higher than for the 0.45 power grading. In such a grading the contribution of aggregate internal friction to the shear resistance of the mix is low. Gradings that plot below the 0.45 power grading line are coarser gradings. They also yield higher VMA values and should be preferred: but gradings plotting too much above or too much below the 0.45 power grading line give rise to tender or harsh mixes respectively, causing compaction or segregation problems in the field. Therefore VMA increases should be achieved by an overall adjustment of the gradation. “By adjusting the proportional percentages of the aggregates that substantially contribute to the intermediate sizes, the gradation curve can be revised to plot further away from the maximum density Tine,” says the MS-2”, Our investigations and observations of several field mixes for bituminous conerete with 26.5 mm nominal ‘maximum size of aggregate revealed that gradings coarser inthe fine fractions up to 2.36 mm sieve size and finer in the coarse fractions from 9.5 mm sieve size upwards relative to the 0.45 power grading offer best WMA values and also better overall Marshall characteristics. One such grading is shown in Fig.11 in relation to the MoSRT&H grading band and in Fig.12 plotied on the 0.45 power ‘grading line, It may be seen that this grading is notaltogether within the MoSRT&H aggregate grading band, 58 Dr. Rao, Das & RovcuowDiury ON 10 | 5 0a8 Power Grading Design Example Gra 8 [ ge Be Percent Passing | eg Selve Size (mim) in the Design E: rading Chart Fig. 10, Aggregate Gradin ‘on the 0.48 Power Sse [ented ering Spel woper tnt wo e508 Percent passing Seve Sie pm) Fig. 12. Modified Aggregate Grading in the Design Example plotted on the 0.45 Power Grading Chart 5.2. Aggregate Properties Influencing VMA It is often difficult to achieve higher VMA values with any kind of adjustment of aggregate gradation alone. Aggregate physical properties play an important role. ‘Changing to angular shape, improving the texture of the coarse aggregate by crushing and switching from naturalAspuatt Mix Deston - Rerusat Densrry Aprroacu For Heavity Trarrickep Roans 59 sands t0 more angular manufactured sands can bring out a significant change in VMA. By this way the entire terlocking and packing mechanism of the aggregates modified giving rise to enhanced shear resistance of the mix. The Superpave mix design method recognizes this as an important requirement and stipulates that the entire course aggregate should have two or more fractured faces (for heavy traffic situations) and low- angularity natural sands should not be used as fine aggregate. It also stipulates a limit to the proportion of flat or elongated particles in coarse aggregate (sizes coarser than 4.75 mm). The type of crusher used for crushing the aggregates makes a substantial difference to the shape of the particles obtained. Vertical shaft impact (VSI crushers tend to produce more cubical particles than cone crushers. 5.3. Influence of Filler (< 0.075 mm) Content on YMA Reducing the filler content in the mix results in an increase of VMA and is often resorted to by the designer. This is not a wise step. Aggregate interparticle contact provides nearly all the internal shear resistance of an asphalt mix and the material passing 75 micron sieve together with the asphalt makes a major contribution 10 the mix cohesion, High cohesion provides the internal tensile strength and mix toughness to resist the shearing forces which cause rutting, It is therefore necessary 10 avoid low percentages of filler. However, excessive amounts of filler reduce the VMA, increase the aggregate surface area and reduce the asphalt film thickness thereby affecting the-durability of the mix. Typical specifications require the filler to bitumen ratio to be between 0.6 and 1.2 by weight. 6. MARSHALL PROPERTIES OF ASPHALT MIX WITH MODIFIED AGGREGATE GRADING Blending the aggregates used in the design example to obtain the grading as shown in Fig. 12 (with a view to increase the VMA value) the Marshall design has been carried out, making mixes with varying asphalt contents and testing them for their Marshall properties which are shown in Figs. 13 to 18, When we compare them with the Marshall properties of the mix in the design example are shown in Figs. 2 to 7 we observe that for any given asphalt content in the range of asphalt contents tested, the VMA is more with the modified grading than with the original grading. So are the air voids. The mix design with the modified grading translates as below. ‘The initial design asphalt content, that is, the asphalt content for achieving the median of air voids limits is 5.3, per cent, All other Marshall properties are also satisfied at this asphalt content and hence this can be taken as the design asphalt content. This is however: higher than the MoSRT&H suggested minimum (5 per cent) and the primary anxiety in this case is whether this asphalt content is not too high to make the mix prone to rutting under heavy traffic. To examine this we should resort to testing the mix volumetrics at refusal density and see whether the air voids level dips below 3 per cent. Before that, however, a mixture’s propensity to rutting can be judged with some confidence from its VMA curve shown in Fig.14, MS-2'° emphasizes ‘asphalt contents on the “left or right-hand increasing side of the VMA curve is avoided. Design asphalt contents in this range have tendency to bleed and/or exhibit plastic flow when placed in the field’. Since 5.3 per cent asphalt content is on the right-hand side of the VMA curve, indicating the mixture’s, propensity to ruiting, lower asphalt content should be selected as the design asphalt content. It is taken as 5 per cent, the MoSRTS&H minimum, as all the Marshall feria are fulfilled at this asphalt content also, although the VEB is just the minimum at 65 per cent. wo 0 60 so 0 AIR VOIDS (9 0 20 4s 50 ss 60 eum Content (6) Fig. 13 cme Consent (%) Fig.60 Dr. Rad, Das & ROYCHOWDHURY ON 6.1, Testing for Refusal Density Permanent deformation occurring in an asphalt mixture that densifies to a very low air void content is called tertiary flow in theological terms. The Super pave mix design system uses the ‘repeated shear test’ to analyses this property. Some highway agencies stipulate, as part of mix design procedure, camying out wheel- tack tests, like the Hamburg Wheel test, to measure ruts created by repeated passage of a wheel over asphalt concrete specimens of the designed mix. Rutting is then measured by the relative percentage reduction in the thickness of asphalt concrete in the wheel path. vera ea Ssanen Content) Fig. 15: On the other hand, refusal density testing is simply a procedure for ascertaining the air voids level when the mix has achieved its maximum density under Marshall compaction (or any other type of compaction). ‘This ait voids level is then compared with the threshold lower limit of air voids (3 per cent) that has been recognized as an indicator to the propensity of the mix for rutting, BULK DENSIEV Gene) It can be easily realized that when the compactive effort increases, the density of the mix increases and the air voids decrease. This state will continue till the density ceases to increase with further increase in compactive effort. When this happens the mix is said to have been compacted to ‘refusal density’. Although it is not clear how much compactive effort would represent the secondary compaction under a given intensity of traffic loading, the corresponding field density would however be not greater than the ‘refusal density’. So if the air voids level at refusal density of the mix is restricted to 3 per cent, the air voids in the field mix could be expected not to go down below 3 per cent, however high the secondary compaction is. It can be inferred that such a mix is not prone to rutting from the Fig. 17 consideration of secondary compaction under traffic. — g 5 5 itmen Contest (2%) For refusal density testing, Marshall specimens have = ; been made with the modified aggregate grading at the | | Vy | design asphalt content of 5 per cent, giving different compactive efforts ranging from 75 blows to 400 blows and the air voids of the mixes have been determined from the G,,,, (maximum specific gravity of the mix) and re TT | G,,, (bulk specific gravity of the compacted mix) values, 1k | I “The results are shown in Fig. 19. The mix did not density a further beyond 300 blow compaction and so the air voids did not change. Therefore a 300-blow compaction may be taken to impart refusal density to this mix, The air Fle. 18 voids have fallen to 2.6 per cent under this compactive FLOW cam) 4s 50 ss 60 Bidymen Content (%)‘Asrniatt Mix Desiow - Rerusat. Density Approack For Hravi ‘TrarrickeD Roaps OL effort but are above 3 per cent till a compaction level of 250 blows. We could achieve the 3 per cent air voids level under refusal density by adjusting the agaregate grading further but the current grading as such can be ecepted because it has offered 3 per cent air voids at 250-blow Marshall compaction which may be a good enough representation of the secondary compaction under the waffie situation. { ‘Compaction vs Air Vaids | so 1180 250 3003600 Marshall compaction in no. of blows NDED MODIFICATIONS TO ASPHALT MIX DESIGN Step 1: With the chosen aggregate grading within the MoSRT&H grading band, fix up the asphalt content, not lower than the MoSRT&H suggested minimum, so that the mixes made with 75-blow Marshall compaction satisfy the Marshall eriteria. Step 2 : Make Marshall Specimens of the same mix under increasing compactive efforts till “refusal density” and determine the air voids in the mixes from theirG., and G,,, values. Ifthe air voids at refusal density is not lower than 3 per cent approve the mix. Step 3: If the air voids at refusal density is less than 3 per cent, change the aggregate grading and repeat the design process until you achieve 3 per cent at asphalt content not lower than the MoSRT&H minimum. ‘The gracing may go out bounds of the MoSRT&H grading band but must comply the Superpave grading controls If refusal density is regarded as too severe a condition, we might choose a lower compaction level, something like 200 blows for 3 per cent air voids, but such a value should be established based on observations of respective field mixes. The modified mix design steps are shown in the Flow Chart given as Appendix. 8. CAN WE ELIMINATE MIX RUTTING BY USING MODIFIED BINDERS’ Mixture resistance to rutting is jointly contributed by the binder viscosity and the interparticle friction which is a function of the surface roughness of the aggregate and the inter-granular contact pressures. It has been an established observation that under slow moving heavy vehicles and at high pavement temperatures asphalt acts like a viscous liquid and contributes little to the stability of the mix which is then largely dependent on che aggregate internal friction", Asa binder alone, modified bitumen can be manufactured to possess greater rutting resistance at higher temperatures than straight-run bitumen’s but “the choice of asphalt binder grade alone \ill not eliminate permanent deformation which is strongly dependent upon the aggregate properties and volumetric properties of as-constructed paving mix”. The scope of the Paper does not include a discussion on modified binders. 9. CONCLUSIONS AND RECOMMENDATIONS, Asphalt mixes designed by the Marshall method in accordance with the guidelines of MS-2 and MoSRT&H specifications have been performing poorly under heavy traffic conditions, 75-blow Marshall compaction ‘s inadequate to represent the in-place densities attained by the field mixes under heavy axle loads. The mixes fail for one or more of the following reasons. * Inadequate initial compaction, making the mix vulnerable to high secondary compaction under traffic. * Relatively high asphalt content that allows the reduction of air voids to lower than 3 per cent under secondary compaction thereby leading to rutting when pavement temperatures rise in summer. ‘© Low asphalt content and high air voids, leading to top-down cracking, raveling and stripping, thereby making the mix less durable. ‘An asphalt mix must have ‘enough’ bitumen and ‘enough’ air voids to be both stable and durable. To fulfill this requirement the VMA of the mix should be high. To achieve high VMA we should use rough-textured (crushed) angular aggregate and an aggregate grading that is not too dense,62 De. Rao, Das & Raycrowpsuky on A grading which closely follows the PRA (0.45 power grading) grading is a dense grading. Its VMA ‘would be low. A mid-point grading of the MoSRT&H grading band plots close to the PRA grading shown in Fig. 10. A grading above or below the 0.45 power grading line offers higher VMA values. Our field obscrvations and laboratory investigations suggest that for bituminous concrete with 26.5 mm. nominal maximum size of aggregate (NMSA), a grading coarser in the fine fractions up to 2.36 mm sieve size and finer in the coarse fractions from 9.5 mm sieve size upwards relative 1 the 0.43 power grading offers the best VMA values and also better overall Marshall characteristics. These results are reported in this paper. ‘The propensity of an asphalt mix for rutting under secondary compaction due to traffic was judged by studying the air voids levels under increased Marshall compaction up to “refusal density’ It is observed that a grading which offers higher MA values not only permits the use of higher asphalt content but also records safer air void levels under refusal density compaction. Based on these observations a modified approach to the asphalt mix design has been suggested in the paper. REFERENCES: 1 Asphalt Institute (19961) “Mix Design Methods for Asphalt Concrete and Qther Hot-Mix Types”, Asphalt Institute Manual Series No.2 (MS-2), Siath Edition, p.1 2. TRL, UK (2002a) “A Guide to the Design of Hot-Mix Asphult in Tropical and Sub-tropical Countries", Overseas Road Note 19, p.5. Asphalt Institute (1996b) “Mix Design Methods for Asphalt Concrete and Other Hot-Mix Types”, Asphalt Institute Manual Series No.2 (MS-2), Sixth Edition, p.73. 4 Asphalt Institute (2001) “Superpave Mix Design’ Institute Superpave Series No.2 (SP-2), Third Ful Asphalt Sterling, A.B. and K.A. Zamhari (1997) “Designing Asphalt Mixes to Last” Road Research Development Project Published Paper PAS, TRL, UK in association with PT Yodya Kurya, Indonesia, pp.5-6, Smith, H.R. and C.R. Jones (1998) “Bituminous Surfacings for Heavily Trafficked Roads in Tropical Climates” Proceedings of Institution of Civil Enginoers, ‘Transport & Paper PA 333498, TRL, UK, pp. 32-33 Smith, H.R. (2001) “Bituminous Surfacings for Heavily Trafficked Roads in Tropical Climates” World Bank Regional Seminsr on Innovative Rond Rehabilitation and Recycling Technologies, Amman, Jordan & Paper PA3737/ OL, TRL, UK, pp. 10-14. TRL, UK (2002) “A Guide to the Design of Hot-Mix Asphalt in Tropical and Sub-tropical Countries”, Overseas Road Note 19, pp. 69- Asphalt Institue (1996c) “Mix Design Methods for Asphuilt Concrete and Other Hot-Mix Types”, Asphalt Institue Manual Series No.2 (MS-2), Sixth Ecition, p.36. Asphalt Institute (1996d) ~ Mix Design Methods for Asphalt Concrete and Other Hot-Mix Type |. Monismith, C.L (1961) “Asphalt Paving Mintures- Properties, Design and Performance” Short Course on Asphalt Paving, University of California, Berkeley, Los Angeles, p.93. Asphalt Institute (1996) “Superpave Performance Graded Asphalt Binder specification and Testing”, Asphalt Institute Superpave Series No.l (SP-1), p.6- National Research Council (1994) “The Superpave Mix Design Manual for New Construction and Overlays” Strategic Highway research Program Report SHRP-a-407, p36,Aspuatt Mix Desicn - RerusaL DeNstry Arproaci For HEAVILY TRAFFICKED ROADS 63 Further work for Rationalizing Asphalt Mix Design ‘The Marshall method of asphalt mix design given in MS-2! envisages that the adopted compaction simulates the in-place field density after the mix has endured several years of traffic, But a 75-blow compaction is inadequate to represent this state where the axle loadings are heavy and the pavement temperatures are high. There will be significant additional compaction due to traffic (secondary compaction) that usually leads to rutting of the mix. This could be mitigated if the mix is designed using greater compactive effort with accompanying adjustment of asphalt content and aggregate grading for balancing other mix characteristics. What should be this, additional compactive effort in terms of the number of blows of Marshall compaction should be established by studies of field performance of the laboratory designed mixes and their volumetrics under traffic loading. Field. investigations are under way in some of the on-going projects executed by usto study the densification of mixes under traffic. ‘The refusal density approach to asphalt mix design suggested in this paper is only a transitional step towards rationalizing mix design procedures. A rational mix design method should meet the performance requirements dictated by the traffic situation and environmental conditions (climate) at the particular pavement site and Asphalt Mix Design ASPHALT MIX (HMA) DESIGN Carry out mix design as per MS-2, determine OBC Appendix for Refusal Density LEGEND: HMA = Hot Mix Asphalt PDAC= Preliminary Design Asphalt Content DAC = Final Design Asphalt Content OBC = Optimum Bitumen Content AlC = Asphalt Content RD. = Refusal Density OBC is PAC Carry out Marshall tests, varying compaction till Refusal Density, draw ‘no. of blows Vs AIR VOIDS. y! Is OBC = o> . 1oSRTH minim MoSRTH minimum is pbac ———a PDAC is DAC }«—____. 2" Trial ‘The nearest A/C to MoSRTH minimum for 3% AIR VOIDS at Refusal Density is the DAC Is AIR VOIDS ~ or > 3% at RD. Change the aggregate grading and repeat the dosign steps Is There A/C = or > MoSRTH Minimum, for 3% AIR YOIDS at Refusal Density This A/C Is DAC64 Da. Rao, Das & Roycuowouuay on Aspuatr Mix Desicn - Rerusat Density Approack For Heavity Trarricken Roaps should be based on the performance-telated properties of the materials in the mixes as the criteria for mix design, In recognition of this fact, the Marshall method of mix design is gradually being replaced by the Super pave Mix design by several agencies in the world, In the Super pave system the mix is designed to resist three modes of failure. which are, permanent deformation, fatigue cracking and low temperature cracking, by taking into consideration performance-based characteristics of materials and by performance testing of mixes. ‘The authors are of the opinion that our mix design procedures would have to be altered on the lines of Super pave mix design and eventually adopt it as the standard mix design procedure.