CORRELATING LABORATORY TEST METHODOLOGIES TO MEASURE SKID RESISTANCE OF PAVEMENT SURFACES

Transcription

1 CORRELATING LABORATORY TEST METHODOLOGIES TO MEASURE SKID RESISTANCE OF PAVEMENT SURFACES Adelia D. Nataadmadja (Corresponding Author) PhD Candidate Department of Civil and Environmental Engineering The University of Auckland Private Bag 92019, Auckland 1142, New Zealand Tel: anat022@aucklanduni.ac.nz Douglas J. Wilson, PhD. Senior Lecturer Department of Civil and Environmental Engineering The University of Auckland Private Bag 92019, Auckland 1142, New Zealand Tel: +64 (9) Fax: +64 (9) dj.wilson@auckland.ac.nz Seosamh B. Costello, PhD. Senior Lecturer Department of Civil and Environmental Engineering The University of Auckland Private Bag 92019, Auckland 1142, New Zealand Tel: +64 (9) ext Fax: +64 (9) s.costello@auckland.ac.nz Minh Tan Do IFSTTAR Department of Planning, Mobility and Environment Route de Bouaye, CS4, Bouguenais, FRANCE minh-tan.do@ifsttar.fr Submission Date: August 1, 2014 Word Count: Manuscript 5,737 (abstract = 222) each = 7,487 Prepared for presentation at the Transportation Research Board 94 th Annual Meeting, Washington D.C. and possible publication in the Journal of the Transportation Research Board

2 ABSTRACT The skid resistance performance of surface aggregates can be assessed in the laboratory by using a polishing device that is designed to simulate traffic polishing actions coupled with a skid resistance measuring device. The standard laboratory test has historically been the Polished Stone Value (PSV) test as specified in BS EN :2009. However, as the technology has advanced and traffic volume and composition on our roads have changed, other devices and methodologies have been developed in an attempt to better assess the skid resistance performance of surface aggregates. This paper discusses the usage of two alternative laboratory tests, the Auckland Pavement Polishing Device (APPD) and the Wehner/Schulze (WS) devices, to assess the skid resistance performance of three different New Zealand Greywacke aggregates. The paper discusses the variations in APPD and its corresponding skid resistance measuring device, the Dynamic Friction Tester (DFT) and the effect of macrotexture in the APPD-DFT test. The results show that there is a good relationship between the APPD-DFT and the PSV tests, while there is no relationship between the WS and the PSV tests. However, the APPD and the WS tests correlate very well. There are only three Greywacke aggregates used in this research, and hence, no conclusive results can be drawn on the relationship between these laboratory tests as yet. Keywords: skid resistance, aggregate, polishing, PSV test. INTRODUCTION Skid resistance is described as the friction generated between the tyre and pavement surface. Two pavement characteristics primarily influence skid resistance properties, namely microtexture and macrotexture (1). However, skid resistance values are not constant and vary in time depending upon a number of factors, which can be classified into four main groups: surface aggregate, load, environmental and vehicle factors (2). Out of all the factors listed above, only the surface aggregate factors, and partially the load factors, are controllable (3). Aggregates deteriorate over time although at that time they did not have a good understanding how the combined action of traffic, weather, and other factors polished the aggregates (4). A number of recent studies have been dedicated to improving the methods for predicting the field performance of aggregates under the influence of these in-service factors (2, 5). Since aggregates influence the skid resistance of the road, deterioration of the aggregate will reduce the available skid resistance. Therefore, it would be beneficial for highway engineers to be able to forecast such deterioration before the pavements are constructed, in order to achieve appropriate design life and to enable maintenance interventions to be planned in advance. Laboratory tests are generally used to assess and predict the skid resistance performance of pavement surfaces. The Polished Stone Value (PSV) test is the most popular laboratory test used worldwide to assess the skid resistance performance of surface aggregates as specified in BS EN :2009 (6). However, the ability of the PSV test to predict both the initial and long term aggregate surface performance has been challenged by a number of researchers (7-9). Limitations that affect the ability of the PSV test to predict the long term performance of road surfaces include: The PSV test primarily measures the influence of the microtexture of aggregates (10). The traffic volume and composition have changed markedly since the PSV test was first developed (2). The standard PSV test uses a set polishing time of six hours, and hence, there is a possibility that the aggregates have not reached the equilibrium skid resistance by then (7).

3 As a result, other devices and laboratory test procedures have been developed in an attempt to overcome these limitations. In addition, the standard issued by the New Zealand (NZ) Transport Agency for the laboratory measurement of skid resistance now allows the use of other laboratroy tests instead of the PSV test. This paper discusses the skid resistance performance of three different NZ Greywacke aggregates assessed with the PSV test and two alternative laboratory tests, namely the Auckland Pavement Polishing Device (APPD) and the Wehner/Schulze (WS) device. BACKGROUND The majority of the NZ road network (approximately 65%) is constructed as a flexible pavement with a chip seal or surface dressing (3). In contrast, in other countries most of their strategic road network consists of predominantly asphalt, and in some limited cases concrete surfaces. Most laboratory test methodologies developed to predict skid resistance performance of road surfaces therefore focus on assessing asphalt surfaces. Mineralogy of Aggregates Rocks from different geological sources can be used as aggregates in surface dressing, for example igneous or volcanic rocks (including Basalt and Andesite), metamorphic rocks, and sedimentary rocks (including Greywacke). Each of them are further characterised for use as road aggregates by their strength, durability, resistance to polishing action of traffic, and rates of wear. Sedimentary rocks in NZ (typically Greywacke) usually have the highest resistance to polishing, whilst some igneous rocks also have reasonably high resistance to polishing. Some metamorphic rocks are not suitable to be used as surface aggregates because the natural process of how they are formed creates foliated structures which reduces the resistance to abrasion. However, some of these may provide a reasonably high skid resistance, such as quartzite which has a hard mineral content and dense interlocking texture (11). Approximately 75% of the aggregates (both for underlying pavement layers and surfacings) produced in NZ are sourced from Greywacke (12). However, not all parent geological rock sources perform the same and it has been found that there are large variations across NZ in geological properties, degrees of metamorphism, and therefore engineering performance. More recently, roads and road sections with high demand for skid resistance and/or locations with recognised wet and loss of control crash related problems have used artificial aggregates that are more resistant to polishing. Artificial aggregates available in NZ for surfacings include Melter Slag, Electric Arc Furnace, and Calcined Bauxite (12). Three different Greywacke aggregates are used in this research, denoted as G1, G2, and G3. They are all quarried from different areas around the North Island of NZ. G1, G2 and G3 are coarse grained, medium grained and fine grained Greywackes, respectively. Aggregate Polishing Using the APPD An alternative laboratory test to the PSV was developed by Wilson (2) using an accelerated polishing, called the Auckland Pavement Polishing Device (APPD), to polish the specimens and the Dynamic Friction Tester (DFT) to measure the skid resistance. This test is called the APPD-DFT test. Although the APPD was developed at the University of Auckland (2), the design was adapted from the Three Wheel Polishing Device (TWPD) device developed by the National Centre of Asphalt Technology (NCAT). The TWPD was developed to polish Hot Mixed Asphalt surfaces, whilst the APPD was specifically developed to polish chip seal surfaces. Modifications were required in terms of its mechanical operation so that the device could accommodate the more irregular and higher macrotexture that is common of NZ aggregates.

4 The three pneumatic castor wheels rotate around a central pivot at 50 rotations per minute. The total load imposed on the specimen is 57 kilograms. The TWPD was reported to only polish the microtexture of the prepared specimen and, given the similarity in the test devices, this raises a similar question regarding the APPD. This research investigates whether the APPD polishes both macrotexture and microtexture by monitoring the changes in macrotexture of the polished specimens by using the Circular Texture Meter (CTM). In addition, the APPD has not been fully standardised as yet, but this will form the basis of research going forward. Preliminary studies have shown that the APPD-DFT test result does not correlate well with the PSV test result, but suggest that it correlates well with historical infield skid resistance data for one type of NZ greywacke aggregates. Detailed explanation of the findings from these studies can be found in (13). The WS Test The WS device used for this research was available from IFSTTAR Nantes, France. It was developed in Germany and was initially designed for use with asphalt mixes. It has three rubber heads that are used to polish the surface sample and a separate measuring head with three rubber sliders to measure the skid resistance, and thus, it can record how the skid resistance of specimens changes during the polishing process. An advantage of this device is its ability to test both laboratory-prepared specimens and core samples. The diameter of cored samples and prepared specimens is 225 mm. The rubber cones (Figure 1) rotate at 17 km/h and have a contact pressure of 0.4N/mm 2. The slip between the cone and the specimen surface is between 0.5% and 1%, which is similar to the slip between vehicular tyres and the road surface. A mix of water and quartz is also sprayed onto the specimen while it is being polished FIGURE 1 Wehner/Schulze Machine (14) A number of research projects have been undertaken to assess the ability of the WS to predict skid resistance. Some researchers have claimed that the WS device has the potential to provide a better prediction of road surface performance in comparison to the PSV test (15). Moreover, several research projects have attempted to correlate the WS and the PSV test results. Some found that they correlate well (16) and others found that there is no correlation between them (17, 18). Therefore, further investigation is required. EXPERIMENTAL PROCEDURES The PSV Test

5 The PSV test for this research project was undertaken according to the procedure stated in BS EN :2009 (6). According to the standard, four aggregate sample specimens and four control sample specimens are to be made for each aggregate, but in this research project, six specimens were made for both aggregate and control samples aggregate. The original equation used to calculate the PSV (6) is shown below: PSV = S + (52.5) C Equation 1 where S = the average value for the four aggregate specimens C = the average value for the four control sample specimens Due to the extra specimens used in this research, the S and C in Equation 1 are modified to be the average values for all aggregate and control specimens, respectively. Laboratory Test Using the APPD-DFT Test Specimen Construction The procedure for constructing the sample specimens for the APPD-DFT Test has been adopted from (2, 3). Figure 2 shows a plan of the specimen constructed for the APPD-DFT test. Blue and red lines are drawn to mark the test device location of the Dynamic Friction Tester (DFT) and the CTM equipment, respectively. For each type of aggregate, three specimens were constructed. One specimen is left unpolished as a control specimen and two specimens are polished, denoted as Polished 1 and Polished FIGURE 2 The diagram of the sample specimen and its location on the APPD Platform

6 Sample Polishing The prepared sample specimen and the APPD are placed on the designated area on the APPD platform. The final arrangement is shown in Figure 2. This is undertaken to ensure that the APPD always polishes the same area on each polished specimen. Water is sprayed onto the specimen during the polishing process; however no additional contaminants are added unlike the PSV test method. The APPD is stopped at certain time intervals (15, 30, 45, 60, 90, 120, 180, 240, 300, 360 and 479 minutes) in order to record the skid resistance and macrotexture of both polished and control specimens with the DFT and the CTM, respectively. The polishing is stopped at 479 minutes (about six hours) because experience has shown that the skid resistance will have already reached the equilibrium skid resistance (ESR) state. This is checked by ensuring that the skid resistance values measured at 300, 360 and 479 minutes are within 0.01 of each other. Skid Resistance and Macrotexture Measurements The DFT is used to measure the skid resistance of specimen surfaces before polishing and after the specimens are polished for certain time intervals, as mentioned above. The DFT is placed on top of the surface specimen and in the designated area. The initial speed is set at 60 kph and the skid resistance is measured three times. The skid resistance values are recorded at 20 kph (DFT 20 ) and 40 kph (DFT 40 ). For this research, the skid resistance values are reported as the average values of DFT 20 of the three measurements. After the skid resistance was measured, the specimen was left to dry before the macrotexture measurements were taken using the CTM. The macrotexture is measured as the Mean Profile Depth (MPD). The measurements were undertaken three times and the average value was reported. The WS Test Specimen Preparation and Construction Four sample specimens were constructed for each type of aggregate, denoted as Sample A, B, C and D. Three of the specimens were polished (Sample B, C, and D) and one was left unpolished as a control specimen (Sample A). Before the specimen was constructed, the aggregates are washed with water until they are reasonably clean and then oven dried. Before the aggregates are placed into the mould, liquid suspension of waxes is applied to the mould and the mould is left to dry out for approximately 30 minutes. The aggregates are hand-placed into the mould with the flat surface facing down. Elongated and/or flaky aggregates are removed by hand. After all aggregates are placed, fine sand is poured to fill the gaps between the aggregates. A mixture of 300 grams of liquid epoxy resin, 100 gram of hardener and 800 gram of fine sand is mixed well. The mixture is left for one hour before it is poured into the mould. The mould is gently knocked a few times so that the epoxy resin mixture is evenly distributed. The specimen is then left overnight to harden. After the specimen is hardened, it is pulled out of the mould and the edge is trimmed to remove any uneven surfaces. Skid Resistance Polishing and Measurements Both the polishing and measurement of skid resistance of test specimens are done using the WS device. The polishing was stopped at predetermined number of passes (2115, 4230, 6345, 8460, 12690, 16920, 25380, 33840, 42300, 50760, 67500, 90000, and passes) in order to monitor the evolution of skid resistance due to polishing. The number of polishing passes has been calculated so that they are

7 equivalent to the polishing minutes in the APPD by knowing that one rotation is equal to three polishing passes. Sample D was also stopped at Ni, which denotes the number of passes at the intersection point between the slopes of initial and final skid resistance from Sample C RESULTS Skid Resistance Performance of Greywacke Aggregates by the APPD-DFT Test Figure 3a shows the skid resistance deterioration of G1. The initial skid resistance values of the Unpolished, Polished 1 and Polished 2 specimens from the APPD-DFT test are 0.69, 0.73 and 0.76, respectively. During the first two hours of polishing, the skid resistance values of all specimens decrease and then the skid resistance values of the unpolished specimen fluctuate, while the skid resistance values of the Polished 1 and Polished 2 specimens keep decreasing. After six hours of polishing, the skid resistance values of the Polished 1 and Polished 2 specimens decrease by about 42% and 39%, resulting in the final skid resistance values of 0.42 and 0.46, respectively. The skid resistance of the unpolished specimen remains higher than the other two specimens. Both Polished 1 and Polished 2 specimens have similar skid resistance values throughout the polishing process FIGURE 3 The skid resistance of both polished and unpolished G1 (a), G2 (b) and G3 (c) aggregates The initial skid resistance of the unpolished G2 sample specimen is 0.76, which is higher than the other two specimens, which are 0.66 and 0.59 (Figure 3b). For a period, the skid resistance values of all specimens fluctuate. After six hours of polishing, the skid resistance values of the Polished 1 and Polished 2 specimens decreased by about 47% and 41%, respectively.

8 Similar to what was observed for the G1 aggregates, the skid resistance values of the unpolished specimen keep decreasing for up to two hours of polishing and then increased at 360 and 479 minutes. The values are constantly higher than the two polished specimens. The initial skid resistance values of all three specimens of the G3 aggregates are similar, that is 0.63, 0.62 and The skid resistance of G3 aggregates generally decreases over time (Figure 3c). However, unlike the other two aggregates, during the first two hours of polishing, the skid resistance of all specimens fluctuate. The unpolished specimen has a higher skid resistance than both polished specimens. After six hours of polishing, the skid resistance values of both Polished 1 and Polished 2 specimens decreased by approximately 38%. Changes in Macrotexture Due To Polishing By the APPD The macrotexture of the aggregates is represented as the mean profile depth (MPD). Figure 4a shows that the MPD of G1 aggregates generally increases with polishing duration. The MPD of the unpolished specimen slowly increases from 1.6 millimetre (mm) to 1.8 mm over time. The MPD of both polished specimens, however, started off with a low MPD, about 1.2 mm, increased for a period and then remained relatively stable throughout the remaining polishing process. The final MPD for the Polished 1 and Polished 2 specimens are around 1.4 mm and 1.6 mm respectively FIGURE 4 The macrotexture of both polished and unpolished G1 (a), G2 (b) and G3 (c) aggregates The initial MPD of the G2 aggregates samples was 1.4 mm for both polished specimens and 1.5 mm for the unpolished specimen (Figure 4b). The MPD of the unpolished specimen fluctuates, while there is a clear increasing trend of the MPD of both of the Polished 1 and Polished 2 specimens. The final MPDs of the Polished 1 and Polished 2 specimens are 1.8 mm and 1.9 mm.

9 Skid Resistance Nataadmadja, Wilson, Costello and Do Unlike the other aggregates, the macrotexture of all G3 aggregate specimens varies over time (Figure 4c). The initial MPD of the unpolished specimen and the Polished 1 specimen is 1.2 mm, while the Polished 2 specimen has an initial MPD value of 1.5 mm. Comparison of All Skid Resistance Test Results Figure 5 shows the average skid resistance values of both polished specimens for the G1, G2 and G3 aggregates as assessed by the WS device and the APPD-DFT. Despite the fact that the skid resistance measurements were taken by different devices, it seems that the ESR values are similar. As assessed by the WS device, the G1 and G2 aggregates have a very similar skid resistance performance throughout the polishing process, which is better than the G3 aggregates. The skid resistance of the G1, G2, and G3 aggregates decrease by about 24%, 21% and 27%, respectively. In comparison, the skid resistance test done by the APPD-DFT shows that all three aggregates have similar skid resistance performance. However, the skid resistance of the G1 aggregate is higher than the other two aggregates, whose skid resistance values are almost identical. The skid resistance reduction of all aggregates is approximately 40%. Generally, the initial skid resistance of the Greywacke aggregates are higher when the aggregates were polished using the APPD as opposed to the WS device. The skid resistance values of G1 aggregates, as measured by the DFT, are higher than the skid resistance values as measured by the WS device regardless of the number of polishing passes. On the other hand, after being polished with about 40,000 polishing passes, the skid resistance values of G2 and G3 as measured by two different tests are starting to converge. There is a sudden increase in skid resistance of the APPD G1 after being polished by 25,380 passes, with further polishing the skid resistance decreases again Polishing passes Sample 1 APPD Sample 1 WS Sample 2 APPD Sample 2 WS Sample 3 APPD Sample 3 WS FIGURE 5 The skid resistance evolution of Greywacke aggregates as tested by both the APPD-DFT and the WS tests There is a very strong correlation between the APPD and the WS test results for all samples (Figures 6a, 6b and 6c) although it is recognised that the variation in skid resistance values is limited in both tests. The coefficients of determination (R 2 ) values are 0.93, 0.93 and 0.83 for G1, G2 and G3 aggregates, respectively.

10 FIGURE 6 The correlation between the WS and the APPD-DFT test results of G1 (a), G2 (b) and G3 (c) aggregates Figures 7 show the relationship between the PSV test results and the final APPD-DFT and the final WS test results respectively. The PSVs of all aggregate range between 47 and 56, while the skid resistance as measured by the APPD-DFT range between 0.35 and The skid resistance as measured by the WS device, however, shows the skid resistance of all aggregates is similar. The PSV correlates very well with the final skid resistance values generated by the APPD-DFT test (R 2 = 0.98), while it does not have any correlation with the WS (R 2 = 0.06).

11 WS and APPD-DFT Nataadmadja, Wilson, Costello and Do y = x R² = y = x R² = PSV WS APPD-DFT Linear (WS) Linear (APPD-DFT) FIGURE 7 The correlation between the PSV, the WS and the APPD-DFT test results of G1, G2 and G3 aggregates DISCUSSION Understanding Variations in the APPD-DFT Test Results The APPD-DFT test is a newly developed test and the repeatability of the test results has not been rigorously tested previously in comparison to other devices. In this research project, two polished specimens were constructed. From Figures 3a, 3b, and 3c, it can be seen that for the first 30 minutes of polishing, some samples exhibited fluctuating skid resistance values. They did not necessarily follow a clear deterioration trend and this could be caused by the presence of residual clay in the surface macrotexture that was used to construct the specimen. Even though the clay was washed off after the epoxy was hardened, there was some clay still attached to the sample surface that can only be taken off by harshly scrubbing the surface. However, this was not undertaken because it could damage or modify the microtexture of the aggregate surface. After the sample surface was subjected to the APPD polishing, the clay was eventually worn away and the APPD tyres directly polished the aggregates. From this point the skid resistance values started to decrease. The specimen construction is a very important step in the APPD-DFT test, similar to all accelerated skid resistance test methods. The specimen making process was undertaken manually by hand, including selecting aggregate chips to be used to construct the specimen by placing the aggregate chips on to the prepared softened clay. To reduce reproducibility error due to operator variation, only one person constructed all of the specimens. It can be seen from Figures 4a, 4b, and 4c that the skid resistance values of the polished specimens from each aggregate are quite close to each other. This suggests good repeatability of data. Moreover, the two part epoxy used to harden the specimen was mixed by hand and sometimes the parts came from a different production batch, causing a slightly different consistency in the mix. This caused a few aggregates to be stripped from the specimen because the epoxy was not hard enough to hold the aggregate to the specimen. The specimen was repaired when necessary by putting another aggregate (with similar size) back into the matrix using more epoxy. However, there would be a slight change to the macrotexture of the sample. As mentioned before, there was a fluctuation in skid resistance values of the APPD polished G3 aggregates during the first two hours of polishing (Figure 3c). The sudden increase in

12 skid resistance occurred after the specimen was fixed because a few aggregates were dislodged. After that, the skid resistance deteriorated again. A limitation encountered in the research was the unavailability of a temperature-controlled room to conduct the APPD-DFT test. Unlike the WS test that was undertaken in a temperature-controlled room, the APPD-DFT test was conducted in an open space, directly exposed to air temperature. It was found that a significant drop in air temperature lowered the DFT rubber temperature and caused a higher skid resistance value as measured by the DFT (Figure 3b). However, the skid resistance continued to decrease after being polished longer despite the change in the temperature. Similar phenomenon occurred with the G1 aggregates (Figure 3c) where the skid resistance values of the unpolished specimen taken at 360 and 479 minutes increased. This is due to the fact that the measurements were taken in the morning when the temperature was low (approximately ten degrees Celsius). As observed in Figures 3a, 3b and 3c, the skid resistance of all specimens, both polished and unpolished, decrease to a certain point over time. At the earlier stage of polishing, the skid resistance decreases at a faster rate and as the aggregates get closer to the ESR point, the skid resistance decreases at a slower rate. Theoretically, the skid resistance of unpolished specimens should not change over time. However, during the skid resistance measuring process, the DFT rubbers rotate on the specimen and also polish the aggregates, although the polishing is not as harsh as the APPD. This is why the skid resistance of all unpolished specimens also decreases, but remains higher than the polished specimens. The Effect of Macrotexture on the APPD-DFT Test Results The macrotexture of all samples was measured using the CTM device. This is a laser based device, which is very sensitive to the presence of water. Testing the same surface under wet and dry conditions will give different results. All specimens were dried before any macrotexture laser based measurements were taken. Theoretically, the macrotexture of aggregates decrease as the aggregates are polished and/or abrade and are worn away. In the G1 and G2 aggregates, a different trend was observed (Figures 4a and 4b). The initial MPD of both polished specimens were lower. As the specimen surface became more polished, the MPD started to increase to a certain point and then continued with a relatively steady MPD. This phenomenon was caused by the presence of residual clay in the specimen surface as explained previously. Initially, there was some clay present on the specimen surface and covering the aggregate. The MPD is the average height of two surface peaks that are alongside each other. If there is clay filling the gap between the aggregates or even covering some parts of the aggregates, the MPD of the surface will be lower. The increasing MPD as the specimen was polished more and more shows that there is a process of clay removal by the APPD polishing either form the microtexture surface or removal of clay within the macrotexture of the surface. The relative steady values of MPD after the specimen was subjected to more polishing suggest that the clay had been removed to the point where it does not affect the macrotexture measurements anymore. In addition, the macrotexture of the unpolished specimens of G1 and G2 aggregates varies over time. This could be due to the uneven removal of clay by the DFT rubber polishing. Unlike the other two aggregates, all G3 aggregate samples have different initial MPD and the macrotexture of the G3 aggregate fluctuates more around the same MPD level. The polished samples were repaired at 30 minutes by adding a few aggregates to replace the aggregates that were stripped off. The replacement aggregates were carefully chosen so that the size is similar to the ones placed in the sample originally, to prevent any changes on the macrotexture of the sample. However, at first, there was a noticeable change in the macrotexture on the Polished 1 sample, but the repair did not seem to affect the macrotexture of the Polished 2 sample. By comparing the changes in macrotexture and the skid resistance of the G3 aggregate, it was observed that the skid resistance kept decreasing regardless of how the macrotexture changes.

13 Skid resistance is affected by both microtexture and macrotexture. As previously described, the APPD was developed according to the TWPD, which only polishes the microtexture of aggregates. Figure 3a, 3b and 3c show that, in general, the skid resistance decreased, while the macrotexture slightly increased or remained steady. The increase in macrotexture could be explained by the removal of clay, but the fact that the macrotexture was steady after the sample was subjected to longer polishing actions suggests that the APPD also only polishes the microtexture of aggregates, similar to the TWPD. It also suggests that the DFT only measures the changes in microtexture to calculate the skid resistance of the sample surface. Correlation between the APPD-DFT, the WS and the PSV Test Results Figures 6a, 6b and 6c demonstrate the strong correlation between the APPD-DFT and the WS test results for all three Greywacke aggregates. This is an interesting observation as the APPD-DFT test results show higher skid resistance values in the earlier stage of polishing (up to two hours) than the WS test results. The APPD-DFT testing ended after six hours of polishing. A critique posed to the PSV test is that the PSV test only polishes the sample specimen for six hours and it does not guarantee that the aggregates have already reached an equilibrium skid resistance (ESR). It is important that the skid resistance of aggregates is tested at an ESR level because prior to ESR, the skid resistance values can vary more than after ESR. The APPD-DFT test was also stopped after the specimen was polished for about six hours. Figure 7 indicates a strong relationship between the APPD-DFT and the PSV test result although the number of data points means statistical confidence is currently low. Furthermore, it was found that there is no correlation at all between the WS and the PSV, while the APPD-DFT and the WS test results correlate very well. There are only three greywacke aggregates used in this research and this is not enough to draw conclusive results between the various laboratory tests CONCLUSIONS This paper discusses the usage of two polishing laboratory tests, the APPD and the WS devices, to assess the skid resistance performance of three different NZ Greywacke aggregates. Results show that the APPD-DFT and the WS tests correlate very well despite differences between these two machines in terms of polishing actions and friction measurements. The research also shows that there is a good relationship between the final value of APPD-DFT test and the PSV of the aggregates, whereas there is a no relationship between the final value of WS test and the PSV. There are only three aggregates used in this research and this is not enough to draw a full picture of how these laboratory tests are related to each other. A greater number of aggregate specimen test results are required to confidently answer the question of which laboratory test can give a better prediction of the long term in-field skid resistance performance. There are lessons that can be learned from this research, including: Sample construction process is critical for the APPD-DFT device, especially due to the excess clay that may cover parts of the aggregate surface and/or fill the gap between the aggregates; There are other factors linked to the APPD-DFT test that need to be investigated further, such as temperature and human operator related variances; The data suggests that the APPD-DFT only polishes and measures the effect of a change in microtexture of the aggregates. Ongoing research at the University of Auckland is working towards correlating these laboratory test results to measured infield skid resistance and to study the contribution of both microtexture and macrotexture to the skid resistance performance of aggregates. The aggregate matrix is currently being expanded to include a greater variety of aggregates in order to better understand the effect of aggregate microtexture and mineralogy on the evolution of skid resistance.

14 ACKNOWLEDGEMENTS The authors would like to thank: Patrick Maisonneuve and Jean-François Le Fur from IFSTTAR for conducting the WS tests. Mohit Khale, an intern student at the University of Auckland, and Leorica Jambalaos from Downer for conducting the PSV tests REFERENCES 1. Oliver, J.W. Factors affecting the Correlation of Skid-testing Machines and a Proposed Correlation Framework. Road & Transport Research: A Journal of Australian and New Zealand Research and Practice, Vol. 18, 2009, pp Wilson, D.J. An Analysis of the Seasonal and Short-Term Variation of Road Pavement Skid Resistance, in Department of Civil and Environmental Engineering. The University of Auckland, Auckland, Kumar, B. and D.J. Wilson. Prediction of Pavement Surface Skid Resistance and the Effect of Smaller Chip Size. In IPENZ Transportation Group Conference. Christchurch, Henry, J. and W. Meyer. Prediction of aggregate and pavement polishing. Bulletin of Engineering Geology and the Environment, Vol. 29, 1984, pp Kane, M., et al. Exploring the Ageing Effect of Binder on Skid Resistance Evolution of Asphalt Pavement. Road Materials and Pavement Design, Vol. 11, 2010, pp British Standards. Tests for mechanical and physical properties of aggregates. Part 8: Determination of the polished stone value. British Standard Institution, London, Wilson, D.J. and P.M. Black. Comparison of Skid Resistance Performance between Greywacke and Melter Slag Aggregates in New Zealand. In International Safer Roads Conference. Cheltenham, Oliver, J.W.H., P.F. Tredrea, and D.N. Pratt. Seasonal Variation of Skid Resistance in Australia. Australian Road Research Board, South Vermont, Cenek, P., R. Henderson, and R. Davies. Selection of aggregates for skid resistance. NZ Transport Agency, Wellington, Won, M. and C. Fu. Evaluation of Laboratory Procedures for Aggregate Polish Test. Transportation Research Record: Journal of the Transportation Research Board, Vol. 1547, 1996, pp Smith, M.R. and L. Collis. Aggregates: Sand, gravel and crushed rock aggregates for construction purposes. The Geological Society, London, Black, P.M. Geologic Inventory of North Island Aggregate Resources: Influences on Engineering Materials Properties. The University of Auckland, Auckland, 2009, 13. Nataadmadja, A.D., D.J. Wilson, and S.B. Costello. Predicting Skid Resistance Performance of Chip Seal Surfaces in New Zealand. In MAIREPAV 7. Auckland, New Zealand, Do, M.T., et al. Pavement polishing Development of a dedicated laboratory test and its correlation with road results. Wear, Vol. 263, 2007, pp Dunford, A. The Wehner Schulze machine and its potential use to improve aggregate specification. In International Safer Roads Conference. Cheltenham Arampamoorthy, H. and J. Patrick. Potential of the Wehner-Schulze test to predict the on-road friction performance of aggregate. NZTA, Wellington, Allen, B., et al. Prediction of UK Surfacing Skid Resistance Using Wehner Schulze and PSV. In International Safer Roads Conference. Cheltenham Friel, S., et al. Use of Wehner Schulze to predict skid resistance of Irish surfacing materials. In Airfield and Highway Pavement 2013: Sustainable and Efficient Pavements. ASCE, Los Angeles, 2013.

15 15 478