Assessment of asphalt durability tests: Part 2, Comparison of wheel tracking tests using European standards

Transcription

1 PPR536 Assessment of asphalt durability tests: Part 2, Comparison of wheel tracking tests using European standards J C Nicholls, J P Harper, K L Green, J M Prime, R C Elliott and J Grenfell

2

3 Transport Research Laboratory Scott Wilson PUBLISHED PROJECT REPORT PPR536 Assessment of asphalt durability tests; Part 2, Comparison of wheel tracking tests using European standards by J C Nicholls, J P Harper, K L Green, J M Prime (TRL), R C Elliott (Scott Wilson) and J Grenfell (University of Nottingham) Prepared for: Project Record: 2/462_080 Client: Implementing asphalt durability tests (SATS and heavy duty wheel tracker) Highways Agency & Ministry of Defence, Network Services Directorate, Technical Services Division, Pavements Team & Defence Estates (Donna James & John Cook) Copyright Transport Research Laboratory November 2010 This Published Report has been prepared for Highways Agency & Ministry of Defence. The views expressed are those of the authors and not necessarily those of Highways Agency & Ministry of Defence. Name Date Approved Project Manager J C Nicholls 08/05/2009 Technical Referee A Hannah 20/05/2009

4 When purchased in hard copy, this publication is printed on paper that is FSC (Forest Stewardship Council) registered and TCF (Totally Chlorine Free) registered. TRL PPR536

5 Contents Executive summary Overall Project Task 2, Wheel tracking comparison iii iii iii Abstract 1 1 Introduction 1 2 Test Programme Mixtures Test methods Planned programme Sample preparation Testing Initial tests on DBM Repeat tests on DBM Tests on EME Comparison of Test Methods Overview Large device to BS EN and BS 598: Small device to BS EN and BS 598: Large and small devices to BS EN Influencing Factors Air voids content Compaction method Laboratory 20 5 Conclusions 22 Acknowledgements 24 References 24 Appendix A Results 25 TRL i PPR536

6 TRL ii PPR536

7 Executive summary Overall Project A project carried out for the Highways Agency (HA) between 2000 and 2005 involved the determination of the critical factors affecting the durability of long life asphalt pavements. This project included developing means of assessing the key properties and evaluating the specific risks associated with High Modulus Base (HMB) 15/25. The work led to the development of a robust procedure for the assessment of high modulus asphalt mixtures, known as the Saturation Ageing Tensile Stiffness (SATS) test, which is the first procedure of its kind that combines the ageing/water damage mechanisms to which an asphalt pavement is subjected in service within a single laboratory test. SATS makes use of tried and tested methods of assessment (based on mixture stiffness) that have been developed to meet the specific objectives set. Since the earlier work was commissioned, HMB 15/25 has been withdrawn from use and a new high modulus asphalt with a successful track record of use in France, Enrobé à Module Elevé 2 (EME2) has been the subject of a collaborative research programme overseen by the HA together with the Quarry Products Association and the Refined Bitumen Association. EME2 is currently being applied on the Trunk Road network in England with sites being monitored. Against this background, HA wants to widen the applicability of the SATS test so that it can encompass a broader range of materials, particularly the principal base materials, EME2 and Dense Bitumen Macadam (DBM) 50. At the same time, HA intends to move towards the large size Wheel Tracker (WTT), as specified in EN , for testing base materials (as a check on laboratory prepared material) and there is, therefore, a need to generate data for typical materials, particularly EME2 and DBM50, to provide benchmark values. Defence Estates (DE) has a large stock of airfields with old pavements, some of which have experienced very rapid deterioration over a short period of time, leading to operational difficulties. Therefore, DE has an interest in an assessment procedure that could be used to identify selected asphalt base layers at risk of deterioration, potentially involving the use of the SATS and the large size WTT, with potential added value if the SATS test could also accommodate the testing of Marshall asphalt (MA) with low air voids contents. The objectives of this research project are threefold: To widen the applicability of the SATS test so that it can encompass a broader range of materials, including but not limited to EME2, DBM50 and MA (Task 1). To generate data on mixtures such as those in Task 1 using the large size WTT (Task 2). To develop an assessment procedure for the prediction of durability of in situ material (Task 3). This report is the final output on Task 2. Task 2, Wheel tracking comparison The results obtained in this project found poor reproducibility and showed no clear relationships between the different methods. However, the poor reproducibility is also found by others for these wheel-tracking tests. Nevertheless, some interesting observations about factors that affect the results were found. The European standard test method for wheel tracking introduces three procedures to replace the single procedure in the British standard method, two of which can be applied to any particular asphalt type. The results from each of these procedures are likely to be different for any particular mixture, so comparisons are needed to revise the TRL iii PPR536

8 specification limits with the change in method so that they set approximately the same level of deformation resistance after the change in test method. Therefore, a series of comparative tests were untaken on a limited set of asphalt concrete mixtures using different test procedures and methods of compaction by two laboratories. The initial finding from the research was that the large wheel-tracker from France, as used at TRL, cannot discriminate between different poorly-performing mixtures because the upward pressure on the two slabs becomes unbalanced when the deformation exceeds about 10 %. This finding limits the types of asphalt that can be assessed with this procedure. The results compiled from the test programme showed limited relationships between the different test methods assuming the same test temperature. Furthermore, the best relationship, with a correlation coefficient R² of only 0.57, between the rut depth using BS and the large device to BS EN is only applicable with standard depth samples because of the different units employed. Nevertheless, using the relationship found: The French requirement of a maximum rut depth of 7 % with the large device would approximate to a maximum rut depth of 1.5 mm to BS The UK thin surfacing requirements in clause NG942 of 4.0 mm, 6.0 mm, 7.0 mm and 10.5 mm to BS would approximate to 19 %, 28 %, 33 % and 50 %, respectively, for BS EN with the large device. The large size device has not been considered for use on surfacing materials in the UK, but that option is available in the current European standards. However, the equivalent to the UK thin surfacing requirements using the large device would be unrealistically high and would never be reached in a test. The correlation coefficients between the other test procedures are not considered to be adequate for setting specification limits. This poor correlation may be explained by influences observed, including: The deformation occurring increased in mixtures with low air voids contents such that the bottom end of the permitted range of air voids content should be raised to at least 3 %, preferably without raising the upper end to risk the durability. Vibratory compaction produces less deformation than roller-compaction, particularly by pneumatic-tyred rollers, for materials which may be sensitive to deformation. TRL iv PPR536

9 Abstract The tests that are used in the UK to assess deformation resistance of asphalt mixtures by simulation from wheel tracking have changed in recent years with the introduction of European standards. The new standard allows either a small or a large device to be used depending on the maximum axle load being designed for. Each test method measures the deformation resistance using slightly different parameters. Comparative studies have been undertaken using each of the new methods plus the old UK method with two laboratories and two methods of compaction: roller and vibratory. The results have been compared to assess the relationships between the parameters, the influence of the method of compaction and an indication of the precision of the tests. Unexpected outcomes of the results are the inability of the traditional design of the large size device to deal with mixtures having limited deformation resistance and the affect of air voids content on the deformation resistance of category 2 Enrobé à Module Élevé mixtures. 1 Introduction The limit on axle loads on UK highways is currently less than 13 t, but the HA wishes to have the ability to allow in their designs for overloading that is known to occur. Under the European asphalt package of standards, there will be three performance-related methods for type testing in order to CE mark mixtures for deformation resistance when the mixtures are to be used in highways. These methods are: BS EN (CEN, 2003a) with the small wheel-tracking device to Procedure A for hot rolled asphalt used in pavements designed for axle loads of less than 13 t. BS EN with the small wheel-tracking device to Procedure B for other mixtures used in pavements designed for axle loads of less than 13 t. BS EN with the large wheel-tracking device for pavements designed for axle loads of 13 t or more. There is limited familiarity with these test methods in the UK, although the small device is similar to the equipment used for the superseded UK method BS (BSI, 1998). Some pre-normative research (Bonnot, 1997) was carried out prior to the finalisation of EN (CEN,???) in which 13 different asphalt were tested using four different test devices (the large with conditioning in air and pneumatic tyres; the small with conditioning in air and a rubber tyre; the small with conditioning in water and a rubber tyre; and the small with conditioning in water and a steel wheel) by six European laboratories. By allocating the results to one of four classes, it was found that the difference between the first three devices was never more than one class. However, the steel wheel was discounted because the difference between results from the latter two tests using the steel wheel and rubber tyre was greater than the differences between the large and small size devices. The classifications, which were not necessarily carried through into the standard, were based on proportional rut depth as follows: Large size in air after cycles: <5 % 5 to 10 % 10 to 15 % >15 % Small size in air after 945 cycles: <5 % 5 to 10 % 10 to 15 % >15 % Small size in water after cycles: <6.5 % 6.5 to 10 % 10 to 13 % >13 % Unfortunately, Procedure B of the standard for the small size device in air is a compromise between the latter two with condition being in air but the test result coming after cycles, so it is difficult to use the results for a comparison between the current methods. A later study (Nicholls et al., 2006) found a comparison between three wheel-tracking methods using the small device (BS , BS EN Procedure A and BS EN Procedure B) to propose numerical values for use in UK specifications. TRL 1 PPR536

10 Rut depth (mm) to BS EN Published Project Report However, the variation in the ratios of the results from different tests on the same mixture used for the analysis for different mixtures tested was sufficiently great to give rise to some uncertainty about the precision of these proposed values. Furthermore, any relationship used for the change from rut depth to proportional rut depth must be theoretically unsound unless used for a limited range of slab thicknesses. Details have also been provided by Jacobs Laboratory (Walsh, 2009) of tests comparing the small devices that, when the BS EN results are plotted against the BS results, do show relationships that have reasonable correlations, at least with Procedure A, when constrained to pass through the origin (Figure 1.1). It is interesting that the slope of the rut depths for Procedure B is only approximately twice that of Procedure A despite the test lasting for ten times longer Proc. A for 45 C Proc. A for 60 C Proc. A for 45 C Proc. B for 60 C Rut depth (mm) to BS Figure 1.1: Comparison of results provided by Jacobs Laboratory More recently, an inter-laboratory precision exercise has been carried out in Spain for the small device using Procedure B with 14 laboratories taking part. The results, which indicate fairly wide reproducibility, were presented by Baltasar Rubio of CEDEX at a meeting of CEN TC227/WG1 on 24 March 2009 and are reproduced as Table 1.1. These results include the density and percentage rut depth (PRD) as well as the rut depth and wheel-track slope (WTS). TRL 2 PPR536

11 Table 1.1: Spanish inter-laboratory precision exercise results Mixture Property No. of results Mean value Repeatability Reproducibility σ r R σ R R BBTM11A Density (kg/m³) Rut depth (mm) WTS (µm/cycle) PRD (%) AC22S Density (kg/m³) Rut depth (mm) WTS (µm/cycle) PRD (%) The Highways Agency (HA) required the comparisons between the BS EN methods and BS to be extended to include the large device method. HA also wanted to establish whether the method of compaction, which can be either by vibratory compactor to BS EN (CEN, 2003b) or by roller compactor to BS EN (CEN, 2003c), affected the result. It was also intended for the comparison to provide estimates of the precision of the tests, for which no values are currently given in BS EN for the small device using Procedure B. Defence Estates (DE) also sponsored the research because of the need to control deformation on military airfields: some of aircraft apply very high point loads through their narrow tyres. The interest is despite the test for deformation resistance of airfield mixtures is currently Marshall stability. The work was undertaken by TRL and Scott Wilson (SW). The University of Nottingham (UoN) carried out the testing for SW whilst Surrey County Council (Surrey) undertook some of the testing for TRL. TRL 3 PPR536

12 TRL 4 PPR536

13 2 Test Programme 2.1 Mixtures The large size wheel-tracking device is a heavier piece of equipment than the small size device and is widely regarded as particularly applicable to base and binder course mixtures whereas the small scale device has traditionally been used to assess the deformation resistance of the surface course mixtures. Because the primary intention is to generate data for the large size device on the structural layers, the programme is designed around base and binder course mixtures. Therefore, the testing has been undertaken on one dense bitumen macadam (DBM) mixture and two category 2 Enrobé à Module Élevé (EME2) mixtures. The selected DBM was the 0/20 mm dense binder course mixture with the heavy duty macadam option at the mm sieve and 40/60 pen bitumen, as set out in Tables 15 and 16 of BS (BSI, 2005). The target grading was taken as the middle of the grading curve and crushed rock was used, so that the composition is as given in Table 2.1. Table 2.1: Target composition of mixtures Mixture DBM EME2 A EME2 B Sieve size (mm) Proportion passing (%) Bitumen Grade (pen) 40/60 15/25 15/25 Proportion (%) The two EME2 mixtures were kindly provided by the industry, with their target compositions also being given in Table Test methods The three tests that were compared using the three mixtures are: BS (BSI, 1998). BS EN (CEN, 2003a) with the small wheel-tracking device to Procedure B. BS EN with the large wheel-tracking device. Testing to BS EN with the small wheel-tracking device to Procedure A is not included because hot rolled asphalt mixtures were not included in the programme. The TRL small size device, which was used with different configurations for both BS and Procedure B of BS EN is shown in Figure 2.1 and the HA large size device, operated by TRL, is shown in Figure 2.2. TRL 5 PPR536

14 Figure 2.1: TRL small size wheel tracker Figure 2.2: HA large size wheel tracker Separate samples of the three mixtures were compacted for each of the three test methods using one of two compaction methods: Roller compactor to BS EN (CEN, 2003c). Vibratory compactor to BS EN (CEN, 2003b). Roller compaction is the more commonly used method, but vibratory could become popular if more testing is required because it does not require specialised equipment. With the roller compactor, samples for the large size device are generally compacted by a pneumatic-tyre roller (the HA equipment, operated by TRL, is shown in Figure 2.3) whilst those for the small size device are generally compacted by steel-wheeled rollers (as shown in Figure 2.4). As currently written, however, either type, or the sliding steel plate option, could be used for either type of specimen. In this test programme, only TRL 6 PPR536

15 the usual equipment (pneumatic-tyre roller for the large device and steel wheel for the small device) was used for the comparison. Figure 2.3: HA pneumatic compactor Figure 2.4: TRL steel-wheeled compactor The vibratory compaction is made with a conventional pneumatic hammer (Figure 2.5). The size of the foot can vary, with the size used by TRL being 115 mm by 105 mm. TRL 7 PPR536

16 Figure 2.5: Vibratory compaction at TRL 2.3 Planned programme The concept was to manufacture and test samples using different laboratories and different compaction methods in order to obtain an understanding whether either of these aspects was significant to the final result, as well as to use different wheel-tracking methods to try to calibrate the methods against each other. The original test programme is set out in Table 2.2 and is the same for each of the 3 materials tested. The same test temperature of 60 C was used for each of the test methods. The test programme required duplicate test results to both BS EN methods (which require two samples per test) and a single test result to BS (which requires six samples per test) with both compaction methods at both laboratories. The compaction of all the large slabs by roller compactor was undertaken at TRL because SW did not have the equipment. 2.4 Sample preparation When compacting a trial DBM slab on the pneumatic-tyred compaction equipment, the mixture appeared far too binder rich and also compacted too easily. The air voids content was measured at 1.2 % voids, which would be expected to lead to poor deformation resistance. Therefore, the mixture was adjusted by???? to produce a target air voids content of 4.1 %. TRL 8 PPR536

17 Table 2.2: Proposed testing regime Set Compaction Mould (mm) No. Compacted Method Tested 1a BS EN x 180 x TRL BS EN Large TRL Roller compactor 1b 305 x 305 x 50 4 TRL BS EN Small TRL 1c 305 x 305 x 50 6 TRL BS TRL 2a BS EN x 180 x TRL BS EN Large SW/UoN * Roller compactor 2b 305 x 305 x 50 4 SW/UoN * BS EN Small SW/UoN * 2c 305 x 305 x 50 6 SW/UoN * BS SW/UoN * 3a BS EN x 180 x TRL BS EN Large TRL Vibratory 3b compactor 305 x 305 x 50 4 TRL BS EN Small TRL 3c 305 x 305 x 50 6 TRL BS TRL 4a BS EN x 180 x SW/UoN * BS EN Large SW/UoN * Vibratory 4b compactor 305 x 305 x 50 4 SW/UoN * BS EN Small SW/UoN * 4c 305 x 305 x 50 6 SW/UoN * BS SW/UoN * * The University of Nottingham undertook the manufacture and testing on behalf of Scott Wilson. The DBM mixture was still fluid and not easy to compact with the vibratory compactor. Fatting up (the binder and fines rising to the surface) occurred, presumably because of the vibration action, and there were more voids concentrated around the outside of the slab than in the centre. The slab compactor appeared to push the material into the edges more effectively to give more uniform compaction. The compaction of the EME2 mixtures was found to present no problems despite using stiff bitumen, probably due to the high binder content. The laboratory equipment was found to compact the material to give air voids contents that were generally below the target. For the large samples, it was found to be easier to get the correct value with vibratory compaction rather than with the pneumatic-tyred roller. The manual control involved in vibratory compaction meant that all the compactive effort could be kept where it was required. 2.5 Testing Initial tests on DBM The compaction and wheel tracking was undertaking on all the combinations by both laboratories with the exception of the large device on roller compacted samples. The compaction of all the large slabs by roller compactor was undertaken at TRL because SW did not have the equipment. The results are given in Table A.1 of Appendix A, where the results are shown in both the relevant units for the test itself and in the same manner as in one of the other tests to ease comparisons, the conversions being as follows: The BS wheel-tracking rate in millimetres per hour is converted into microns per cycle. The BS rut depth in millimetres is converted into proportional rut depth in percent. TRL 9 PPR536

18 Deformation (%) Deformation (%) Published Project Report The large-size BS EN proportional rut depth in percent is converted into rut depth in millimetres. The results of the large size wheel tracking with roller compaction appeared to be inconsistent between the two laboratories, as plotted out in Figure 2.6. The results from SW were, on average, 33.5 % greater than those from TRL, as shown when factoring the TRL results by 1.33 in Figure 2.7. Both laboratories checked the calibration of their equipment, but found no errors that would explain the difference. After re-calibration, TRL repeated the test with two further samples, which are shown on Figure 2.6 as TRL5 and TRL TRL1 TRL2 TRL3 5 TRL4 TRL5 TRL6 SW1 SW2 SW3 SW Number of Cycles Figure 2.6: Large wheel-tracking results for roller compacted samples TRL1* TRL2* TRL3* TRL4* SW1 SW2 SW3 SW Cycles Figure 2.7: Large wheel-tracking results for roller compacted samples after factoring TRL results by 1.33 All these results are well above the limit used in France where the rut should not be more than 7 % of the slab depth after 30,000 cycles. TRL 10 PPR536

19 Rut Depth (mm) Published Project Report Repeat tests on DBM BP Bitumen agreed to assist in establishing the reason for the difference by carrying out some comparative testing with TRL. BP Bitumen has had the large size equipment for many years and, as such, has more experience than any other UK laboratory. TRL manufactured four additional samples, tested two of them and sent the other two to BP Bitumen. Unfortunately, the samples were warped when they arrived at the BP Bitumen laboratory making it difficult, but not impossible, to get them into the moulds for testing. BP Bitumen stopped the test at 7,375 cycles when the rut depth was around 15 % of the slab depth, double the French limit, whereas TRL carried on their test to 10,000 cycles. The results are given in Table A.1 of Appendix A and shown in Figure TRL7 TRL8 BP1 BP Cycles Figure 2.8: Large wheel-tracking comparative results from BP and TRL Similar differences occurred between the results from the two samples in each pair tested together, with the differences emerging after 1,500 cycles when the deformation was already past the 7 % limit. The material in the mould was still quite fluid at the end of compaction, making it difficult to achieve a perfectly flat surface, and the resulting surface irregularity could have contributed to the difference. Nevertheless, it would appear that the large size device from France, as used by TRL and BP Bitumen but not the University of Nottingham, is not designed to cope with such poorly performing mixtures. It is believed that the upward pressure on the two slabs to present the surface at the appropriate level becomes unbalanced when the amount of deformation exceeds about 10 % Tests on EME2 The wheel tracking on EME2 mixture A was to have been undertaken using all the combinations by both laboratories. However, UoN had a problem with their large wheeltracker, breaking down during the first test after between 1000 and 3000 cycles and destroying the samples. The results that were obtained are given in Table A.2 of Appendix A. The wheel tracking on EME2 mixture B was undertaking using all the combinations by TRL only due to the delay caused by repairing the UoN equipment. The results are given in Table A.3 of Appendix A. TRL 11 PPR536

20 Large Rut Depth (%) Published Project Report 3 Comparison of Test Methods 3.1 Overview In order to identify any relationship between the results using different procedures, the results have been plotted as follows: Figure 3.1 The rut depth to EN BS with the large device in percent against the rut depth to BS in millimetres. Figure 3.2 The wheel-tracking slope to EN BS with the small device in µm/cycles against the wheel-tracking rate to BS in millimetres per hour. Figure 3.3 The rut depth to EN BS with the small device in millimetres against the rut depth to BS in millimetres. Figure 3.4 The rut depth to EN BS with the large device in percent against the rut depth to EN BS with the small device in percent. The linear trend lines for all the data are plotted with a solid line together with one constrained to pass through the origin as a dashed line on each graph. The slope and constant for each line, together with the correlation coefficients, are given in Table DBM, rolled EME2 A, rolled EME2 B, rolled DBM, vibrated EME2 A, vibrated EME2 B, vibrated BS Rut Depth (mm) Figure 3.1: Rut depth by BS against large-size equipment to BS EN TRL 12 PPR536

21 Small Rut Depth (mm) Small Wheel-Tracking Slope (µm/cycle) Published Project Report DBM, rolled EME2 A, rolled EME2 B, rolled DBM, vibrated EME2 A, vibrated EME2 B, vibrated BS Wheel-Tracking Rate (mm/h) Figure 3.2: Wheel-tracking rate by BS against small-size equipment to BS EN DBM, rolled EME2 A, rolled EME2 B, rolled DBM, vibrated EME2 A, vibrated EME2 B, vibrated BS Rut Depth (mm) Figure 3.3: Rut depth by BS against small-size equipment to BS EN TRL 13 PPR536

22 Large Rut Depth (%) Published Project Report DBM, rolled EME2 A, rolled EME2 B, rolled DBM, vibrated EME2 A, vibrated EME2 B, vibrated Small Rut Depth (%) Figure 3.4: Rut depth by BS EN for small- against large-size equipment Table 3.1: Statistics of trend lines Dependant variable Independent variable Constant Slope R² Rut depth to BS EN with large equipment (%) Wheel-tracking slope to BS EN with small equipment (µm/cycle) Rut depth to BS EN with small equipment (mm) Rut depth to BS EN with large equipment (%) Rut depth to BS EN with small equipment (mm) Rut depth to BS (mm) Wheel-tracking rate to BS (mm/h) Rut depth to BS (mm) Rut depth to BS EN with small equipment (%) Rut depth to BS (mm) Constrained * Constrained * Constrained * Constrained * Constrained * * For regression through the origin (the intercept model), R² measures the proportion of the variability in the dependant variable about the origin explained by regression. This value CANNOT be compared to R² for models which include an intercept. Including data from Jacobs and TRL656. The correlation coefficients between most of the test procedures are not considered to be adequate for setting specification limits, but this poor correlation may be explained by influences discussed in Chapter 4. TRL 14 PPR536

23 3.2 Large device to BS EN and BS 598:110 The correlation coefficient, R², between the rut depths to BS EN with the large device and to BS 598:110 of 0.57 is not particularly high but, given the poor precision that can be achieved with wheel-tracking tests (Section 1) and the limited number of data points, it is of the order to be accepted. However, any relationship found will only be applicable when the specimens for the large device are c.100 mm thick and those to BS are c.50 mm thick because of the different units employed. If the rut depth is proportional to the slab depth, then the BS results will change with sample thickness whereas, if the rut is independent of slab depth, then the large size rut depth will be the one that changes. In practice, it is expected that the rut depth will change with slab depth, but not proportionally. Constraining the trend line through the origin is appropriate because a very deformationresistant sample should produce zero deformation irrespective of the method used. For Figure 3.1, the constraint does not appear to be too severe, so that the appropriate relationship is given in Equation 3.1. d (3.1) L d BS where d L is the rut depth of 100 mm thick slabs to BS EN with the large device (%); and is the rut depth of 50 mm thick slabs to BS (mm). d BS Based on Equation 3.1, the French requirement of a maximum rut depth of 7 % under BS EN with the large device would approximate to a maximum rut depth of 1.5 mm to BS assuming the same test temperature. Similarly, the UK thin surfacing requirements in clause NG942 of the Notes for Guidance on the Specification for Highway Works (reproduced as Table 3.2) of 4.0 mm, 6.0 mm, 7.0 mm and 10.5 mm to BS would approximate to 19 %, 28 %, 33 % and 50 %, respectively, for BS EN with the large device. These high limits are unrealistic and would not be reached in the test. Table 3.2: Wheel tracking levels in Notes for Guidance on the Specification for Highway Works Level Test Specimen Maximum wheeltracking Maximum rut temperature thickness Criteria rate depth ( C) (mm) (mm/h) (mm) Mean * Maximum <30 Mean * 1/6 x thickness # 7/30 x thickness # Maximum 1/4 x thickness # 7/20 x thickness # Mean * Maximum <30 Mean * 1/15 x thickness # 2/15 x thickness # Maximum 1/10 x thickness # 1/5 x thickness # Mean * Maximum <30 Mean * 1/6 x thickness # 7/30 x thickness # Maximum 1/4 x thickness # 7/20 x thickness # 0 No requirement No requirement * mean = mean result of 6 consecutive determinations on individual specimens maximum = maximum result from 6 consecutive determinations on individual specimens # thickness = thickness of specimens tested = nominal depth + thickness of regulating ability TRL 15 PPR536

24 Small Rut Depth (mm) Published Project Report 3.3 Small device to BS EN and BS 598:110 For the comparison of the rut depth to EN BS with the small device, against those to BS , the same units of millimetres were used for the results in order to avoid any relationship being limited to specific slab depths: EN BS allows results to Procedure B with the small device to be given in either millimetres or percent. The correlation coefficient, R², for the wheel-tracking slopes of 0.20 is disappointing and that for the rut depths is only marginally better at Looking at just the rut depth and constraining the trend line through the origin, the appropriate relationship is given in Equation 3.2. d (3.2) S d BS where d S is the rut depth to BS EN with the small device (mm); and d BS is the rut depth to BS (mm). Based on Equation 3.1, the UK thin surfacing requirements in clause NG942 of the Notes for Guidance on the Specification for Highway Works (reproduced as Table 3.2) of 4.0 mm, 6.0 mm, 7.0 mm and 10.5 mm to BS would approximate to 7.5 mm, 15 mm, 13 mm and 20 mm, respectively, for BS EN with the small device assuming the same test temperature. These limits again appear high, but they are not as unrealistic as those derived for BS EN with the large device. For more accurate estimates, more data is needed, which should become available from industry with comparative testing on projects with the small device. Some additional data is already available from that provided by Jacobs (Section 1) and from TRL report TRL656 (Nicholls et al., 2006). These data have been added in Figure 3.3 with the statistics from the trend lines included in Table 3.1, which disappointedly shows a reduced value of R² to DBM, rolled EME2 A, rolled EME2 B, rolled Jacobs DBM, vibrated EME2 A, vibrated EME2 B, vibrated TRL BS Rut Depth (mm) Figure 3.3: Rut depth by BS against small-size equipment to BS EN with additional data TRL 16 PPR536

25 3.4 Large and small devices to BS EN The correlation coefficient, R², for the rut depths to EN BS using the large against small devices is again low at 0.07, at which level there appears to be no linear relationship. TRL 17 PPR536

26 Proportional rut depth (%) Wheel-tracking slope (µm/cycle) Published Project Report 4 Influencing Factors 4.1 Air voids content Compaction is an important parameter for durability of most mixtures, with high air voids content being taken as an indicator of a potential to deteriorate relatively quickly. However, excessively low air voids contents can make a mixture, particularly an asphalt concrete mixture, susceptible to deformation. Therefore, the wheel-tracking slopes and proportionate rut depths of each determination with all the mixtures are shown in Figures 4.1 and 4.2, respectively. Figure 4.2 shows the proportionate rut depths for all three methods whilst Figure 4.1 only shows the BS and BS EN smallsize results because the BS EN method for the large-size equipment does not include a measurement of the slope DBM / BS EME2 A / BS EME2 B / BS DBM / Small EME2 A / Small EME2 B / Small Air Voids Content (%) Figure 4.1: Influence of air voids content on wheel-tracking slope Air Voids Content (%) DBM / BS EME2 A / BS EME2 B / BS DBM / Small EME2 A / Small EME2 B / Small DBM / Large EME2 A / Large EME2 B / Large Figure 4.2: Influence of air voids content on rut depth TRL 18 PPR536

27 Proportional rut depth (%) Published Project Report The figures show that the highest deformation occurs at relatively low air voids contents, more so for proportional rut depth than wheel-tracking slope. However, there are still so many variables shown that it is difficult to see clearly any trend. Therefore, the results for the large device to BS EN are shown separately in Figure 4.3. These results include two trial determinations of the EME2 B mixture deliberately targeted at an air voids content of just over 2 % together with two at higher values DBM / Large EME2 A / Large EME2 B / Large Air Voids Content (%) Figure 4.3: Influence of air voids content on rut depth using the large-size equipment to BS EN In Figure 4.3, the DBM can be seen to be the poorest performing of the materials, as would be expected, and that the proportional rut increases when the air voids content is at lower end of the permitted range, below about 3 %. This latter observation indicates that the bottom end of the permitted range of air voids content for these materials should be raised to at least 3 %, preferably without raising the upper end to risk the durability. 4.2 Compaction method The results obtained were paired off for the two compaction methods, roller-compactor (whether steel-wheeled or pneumatic tyred) and vibratory. The inputs for each pair were the test results (mean of two or six determinations, depending on method) on the same material tested at the same laboratory. Where the combinations were replicated, the first roller-compacted result was paired with the first vibratory-compacted result and the second roller-compacted result was paired with the second vibratory-compacted result. The pairs are plotted in Figure 4.4 for wheel-tracking slope (which, therefore, excludes the large-size device to BS EN ) and Figure 4.5 for proportional rut depth. For both parameters, the majority of the results lie reasonably close to the line of equality (shown as a chain-linked line on the figures) with some outliers. Whilst the outliers could be explained by the poor precision found elsewhere for wheel-tracking (at least for the small device to BS EN , see Section 1), the two worst outliers are for the proportional rut depth with the large device to BS EN on DBM, where the roller-compacted samples had significantly higher rut depths than the vibratorycompacted samples. The remaining pairs tended towards the same bias with more points below the line of equality than above. This observation implies that vibratory compaction produces lower results than compaction by pneumatic-tyred rollers for materials which are sensitive to deformation. TRL 19 PPR536

28 Proportional rut depth of vibratory-compacted specimens (%) Wheel-tracking slope of vibratory-compacted specimens (µm/cycle) Published Project Report DBM / BS EME2 A / BS EME2 B / BS DBM / Small EME2 A / Small EME2 B / Small Wheel-tracking slope of roller-compacted specimens (µm/cycle) Figure 4.4: Influence of compaction method on wheel-tracking slope DBM / BS DBM / Small DBM / Large EME2 A / BS EME2 A / Small EME2 A / Large EME2 B / BS EME2 B / Small EME2 B / Large Proportional rut depth of roller-compacted specimens (%) Figure 4.5: Influence of compaction method on proportional rut depth The next worse outliers are the slopes to BS for EME2 A and DBM, but because they on opposite sides of the line of equality, it is assumed that they result from the inherent variability within the test. 4.3 Laboratory The results were also paired off for the two laboratories where the testing had been undertaken, TRL and SW/UoN. In this case, each pair would be a test result on the same material compacted by the same method. The pairs are plotted in Figure 4.6 for TRL 20 PPR536

29 Proportional rut depth at SW/UoN (%) Wheel-tracking slope at SW/UoN (µm/cycle) Published Project Report wheel-tracking slope (which, therefore, excludes the large-size device to BS EN ) and Figure 4.7 for proportional rut depth DBM / BS EME2 A / BS EME2 B / BS DBM / Small EME2 A / Small EME2 B / Small Wheel-tracking slope at TRL (µm/cycle) Figure 4.6: Influence of laboratory on wheel-tracking slope DBM / BS DBM / Small DBM / Large 2 EME2 A / BS EME2 A / Small EME2 A / Large EME2 B / BS EME2 B / Small EME2 B / Large Proportional rut depth at TRL (%) Figure 4.7: Influence of laboratory on proportional rut depth The rut depth results from SW/UoN are shown to be generally greater than those measured by TRL, although all data pairs are not far from the line of equality, shown chained dashed on the figures. The slope results were similar, although the bias was generally less pronounced with slopes less than 0.4 µm/cycle whilst the divergence from the line of equality was greater above that level. Therefore, these results support the poor reproducibility found elsewhere for wheel-tracking (at least for the small device to BS EN , see Section 1). TRL 21 PPR536

30 5 Conclusions The results obtained in this project found poor reproducibility and showed no clear relationships between the different methods. However, the poor reproducibility is also found by others for these wheel-tracking tests. Nevertheless, some interesting observations about factors that affect the results were found. The initial finding from the research was that the large wheel-tracker from France, as used at TRL, cannot discriminate between different poorly-performing mixtures because the upward pressure on the two slabs becomes unbalanced when the deformation exceeds about 10 %. This finding limits the types of asphalt that can be assessed with this procedure. The results compiled from the test programme showed limited relationships between the different test methods assuming the same test temperature. Furthermore, the best relationship, with a correlation coefficient R² of only 0.57, between the rut depth using BS and the large device to BS EN is only applicable with standard depth samples because of the different units employed. Nevertheless, using the relationship found: The French requirement of a maximum rut depth of 7 % with the large device would approximate to a maximum rut depth of 1.5 mm to BS The UK thin surfacing requirements in clause NG942 of 4.0 mm, 6.0 mm, 7.0 mm and 10.5 mm to BS would approximate to 19 %, 28 %, 33 % and 50 %, respectively, for BS EN with the large device. The large size device has not been considered for use on surfacing materials in the UK, but that option is available in the current European standards. However, the equivalent to the UK thin surfacing requirements using the large device would be unrealistically high and would never be reached in a test. The correlation coefficients between the other test procedures are not considered to be adequate for setting specification limits. This poor correlation may be explained by influences observed, including: The deformation occurring increased in mixtures with low air voids contents such that the bottom end of the permitted range of air voids content should be raised to at least 3 %, preferably without raising the upper end to risk the durability. Specimens compacted by vibratory compactor appeared to deform less than those compacted by roller-compactor, particularly by pneumatic-tyred rollers, for materials which may be sensitive to deformation. TRL 22 PPR536

31 TRL 23 PPR536

32 Acknowledgements The work described in this report was carried out in the Design & Materials Group, Infrastructure Division of the Transport Research Laboratory and the Nottingham Transportation Engineering Centre at the University of Nottingham under a contract let to TRL and Scott Wilson. The authors are grateful to Lloyd Cowley, who undertook the laboratory work for BP Bitumen, and to Arthur Hannah, who carried out the technical review and auditing of this report. The authors also wish to thank the industry for their help in supplying materials with which to carry out the testing. References Bonnot, J (1997). Prenormative research on wheel tracking test. N633, report to CEN TC227 WG1 (unpublished). British Standards Institution (1998). Sampling and examination of bituminous mixtures for roads and other paved areas; Part 110, Methods of test for the determination of the wheel-tracking rate of cores of bituminous wearing courses. BS 598: Part 110: London: BSI. Comité Européen de Normalisation (2003a). Bituminous mixtures Test methods Part 22: Wheel tracking. BS EN : London: British Standards Institution. Comité Européen de Normalisation (2003b). Bituminous mixtures Test methods Part 32: Laboratory compaction of bituminous mixtures by vibratory compactor. BS EN : London: British Standards Institution. Comité Européen de Normalisation (2003c). Bituminous mixtures Test methods Part 33: Specimen prepared by roller compactor. BS EN : London: British Standards Institution. Nicholls, J C, C Roberts and P Samuel (2006). Implications of implementing the European asphalt test methods. TRL Report TRL656. Crowthorne: TRL Limited. Walsh, I D (2009). Personal communication. Maidstone: Jacobs Laboratory. TRL 24 PPR536

33 Appendix A Results Table A.1: Dense Bitumen Macadam (Cont.) Compaction Method Test Method Laboratory Air Voids content (%) Wheel Tracking Slope Rut Depth (mm/h) (µm/cycles) (mm) (%) Steel Roller BS TRL Steel Roller Pneumatic Roller BS EN , Small BS EN , Large SW/UoN TRL TRL SW/UoN SW/UoN TRL TRL SW/UoN SW/UoN TRL 25 PPR536

34 Table A.1: Dense Bitumen Macadam (Cont.) Compaction Method Test Method Laboratory Air Voids content (%) Wheel Tracking Slope Rut Depth (mm/h) (µm/cycles) (mm) (%) Pneumatic Roller BS EN , Large TRL TRL * 17.0* 14.2* 14.2* BP Vibratory BS TRL Vibratory Vibratory BS EN , Small BS EN , Large SW/UoN TRL TRL SW/UoN SM/UoN SW/UoN SW/UoN * After 10,000 cycles rather than 30,000 cycles After 7,375 cycles rather than 30,000 cycles TRL 26 PPR536

35 Table A.2: Category 2 Enrobé à Module Élevé mixture A (Cont.) Compaction Method Test Method Laboratory Air Voids content (%) Wheel Tracking Slope Rut Depth (mm/h) (µm/cycles) (mm) (%) Steel Roller BS TRL Steel Roller Pneumatic Roller BS EN , Small BS EN , Large SW/UoN TRL TRL SW/UoN SW/UoN TRL TRL * 18.2* * 14.1* SW/UoN * * SW/UoN TRL 27 PPR536

36 Table A.2: Category 2 Enrobé à Module Élevé mixture A (Cont.) Compaction Method Test Method Laboratory Air Voids content (%) Wheel Tracking Slope Rut Depth (mm/h) (µm/cycles) (mm) (%) Vibratory BS TRL Vibratory Vibratory BS EN , Small BS EN , Large SW/UoN TRL TRL SW/UoN SW/UoN TRL * 7.8* * 8.0* TRL SW/UoN SW/UoN * After 10,000 cycles rather than 30,000 cycles Samples destroyed due to fault with LVDT on wheel-tracker After 3,000 cycles rather than 30,000 cycles TRL 28 PPR536

37 Table A.3: Category 2 Enrobé à Module Élevé mixture B Compaction Method Test Method Laboratory Air Voids content (%) Wheel Tracking Slope Rut Depth (mm/h) (µm/cycles) (mm) (%) Steel Roller BS TRL/Surrey Steel Roller Pneumatic Roller BS EN , Small BS EN , Large TRL TRL trial TRL TRL Vibratory BS TRL Vibratory Vibratory BS EN , Small BS EN , Large TRL/Surrey TRL/Surrey TRL TRL TRL 29 PPR536

38

39 Assessment of asphalt durability tests: Part 2, Comparision of wheel tracking tests using European standards The tests that are used in the UK to assess deformation resistance of asphalt mixtures by simulation from wheel tracking have changed in recent years with the introduction of European standards. The new standard allows either a small or a large device to be used depending on the maximum axle load being designed for. Each test method measures the deformation resistance using slightly different parameters. Comparative studies have been undertaken using each of the new methods plus the old UK method with two laboratories and two methods of compaction: roller and vibratory. The results have been compared to assess the relationships between the parameters, the influence of the method of compaction and an indication of the precision of the tests. Unexpected outcomes of the results are the inability of the traditional design of the large size device to deal with mixtures having limited deformation resistance and the affect of air voids content on the deformation resistance of category 2 Enrobé à Module Élevé mixtures. Other titles from this subject area TRL674 Durability of thin surfacing systems. Part 4: Final report after nine years monitoring. J C Nicholls, I Carswell, C Thomas and B Sexton PPR497 GripTester trial October 2009 including SCRIM comparison. A Dunford PPR468 PPR458 PPR457 Enhanced levels of reclaimed asphalt in surfacing materials: a case study evaluating carbon dioxide emissions. M Wayman and I Carswell Review of UKPMS core functionality the minimum functionality all PMS should embody in the UK. BV Cleave, R A Cartwright, K A Gallagher and T Rasalingam SCANNER accredited surveys on local roads in England accreditation, QA and audit testing annual report P Werro, I Robinson, E Benbow and A Wright PPR437 Highways Agency 2009 National Falling Weight Deflectometer Correlation trials. S Brittain PPR393 Measuring skid resistance without contact progress report. A Dunford RN39 Design guide for road surface dressing. 6th edition. C Roberts and J C Nicholls PPR299 Automated detection of fretting on HRA surfaces. S McRobbie and G Furness Price code: 2X ISSN TRL Crowthorne House, Nine Mile Ride Wokingham, Berkshire RG40 3GA United Kingdom T: +44 (0) F: +44 (0) E: enquiries@trl.co.uk W: Published by IHS Willoughby Road, Bracknell Berkshire RG12 8FB United Kingdom T: +44 (0) F: +44 (0) E: trl@ihs.com W: PPR536