This study aims to advance the current IC technology for asphalt compaction and maximise its benefits for quality assurance and quality control (QA/QC) of asphalt compaction.
Although many field demonstration projects have been undertaken in the past 2 decades, an acceptable correlation between asphalt density and intelligent compaction measurement values (ICMVs) has not yet been established. This study aims to advance the current IC technology for asphalt compaction and maximise its benefits for quality assurance and quality control (QA/QC) of asphalt compaction. Experimental results indicated that ICMVs and asphalt modulus are insensitive to the change in asphalt density at temperatures above the softening point of the bitumen. Therefore it is not possible to achieve an acceptable correlation between ICMVs and asphalt density for use in asphalt compaction practice. It is proposed that the ICMVs recorded during the pre-mapping of existing layers before paving the new asphalt layer, along with the asphalt surface temperature map produced during the final roller pass of the newly laid asphalt layer, could be superimposed to identify the most critical regions where asphalt density should be tested. The effectiveness of this approach was successfully demonstrated during an asphalt testbed construction using the SPARC IC kit. The approach can be easily implemented in existing IC retrofit kits and original equipment manufacturer (OEM) IC systems.
Intelligent compaction (IC) technology involves attaching an accelerometer to the vibratory drum of a roller to map the equivalent modulus of pavement layers (down to the influence depth) in real-time (Sivagnanasuntharam, Sounthararajah, Ghobarni, et al. 2023). The accelerometer data is analysed to produce various Intelligent Compaction Measurement Value (ICMV) parameters, categorised into 5 levels by the Federal Highway Administration (2017). While Level 1 ICMVs, such as compaction meter value (CMV) (Thurner & Sandstrom 1980) and compaction control value (CCV) (Chang et al. 2014) can be calculated in real-time using only the accelerometer data, Level 3 ICMVs, such as roller-integrated stiffness (kb) (Anderegg & Kaufmann 2004) and vibratory modulus (EVIB) (Briaud & Saez 2015), require both accelerometer data and the position of the exciters. ICMVs are indices for equivalent stiffness of the layers in the roller influence depth, with the influence depth depending on the type and mass of the roller, amplitude and properties of the layer. IC roller and influence depth are illustrated in Figure 1.
Many studies have reported the use of IC technology for asphalt compaction (Maupin 2007, Savan et al. 2015, White et al. 2010, Hu et al. 2019, Chang et al. 2018, Foroutan et al. 2020). However, ICMV parameters recorded during asphalt compaction do not correlate well with spot density measurements. Road authorities therefore continue to use density measurements at discrete spots for QA/QC purposes.
IC technology is not widely adopted in Australia. According to a survey, the key reasons for this are a lack of knowledge about it, uncertainty regarding its accuracy and value, high initial costs, a shortage of trained operators and the absence of specifications for its use (Sivagnanasuntharam 2023). Even where it is used, it is primarily for tracking roller pass counts and asphalt surface temperature, rather than for density estimation. As a result, the full potential benefits of a real-time monitoring system such as IC are not being realised.
This study has 2 main objectives. First, it aims to gain a deeper understanding of the limitations of current IC technology in asphalt compaction, particularly its effectiveness as a reliable indicator of density. Second, it seeks to explore alternative methods that can be integrated with IC systems to enhance their overall efficacy.
Figure 2 shows the variations of CMV and CCV with asphalt air void content for 2 different temperatures of the asphalt mix. It can be seen that CMV and CCV do not vary with changes in asphalt density during the vibrating hammer compaction test on hot asphalt mix at 60 ̊C and 120 ̊C. The vibratory hammer was instrumented with an accelerometer to determine the CMV and CCV during the compaction of the asphalt specimens at different temperatures and densities; and the samples were prepared using a gyratory compactor to target densities and conditioned to a certain temperature in the oven before performing the vibratory hammer tests. Further investigation using various modulus tests, such as lightweight deflectometer (LWD), indirect tensile test, unconfined compressive test and flexural modulus test, revealed that the density of asphalt has negligible influence on its modulus when the temperature of the asphalt mix exceeds the softening point of the bitumen (an example graph for the unconfined compressive modulus of asphalt is presented in Figure 3). Therefore, it is not possible to use the asphalt modulus or its indices (including CMV and CCV) to estimate the density of asphalt during compaction in the field.
This study presents an innovative approach known as the SPARC-NCA (Smart Pavements Australia Research Collaboration-No Correlation for Asphalt) approach. It aims to pinpoint the most critical areas where spot tests for asphalt density (utilising NDG/Nuke or cores) can be performed.
Weak underlying support may absorb a significant portion of the compaction effort during asphalt layer compaction, resulting in low densities in the asphalt layer. Moreover, areas with low surface temperatures demand particular attention, as asphalt workability decreases significantly with a reduced temperature (especially below the asphalt softening point). By superimposing the pre-mapping of ICMV (collecting ICMV data on the existing layer at lowered frequency and amplitude of vibration before paving the new asphalt layer) onto the asphalt surface temperature map produced in the final roller pass of the asphalt compaction, the pavement lot can be categorised into three regions. These are supercritical; critical; and normal (see Figure 4). Regions with low pre-map ICMV and low asphalt surface temperature are classified as 'supercritical', whereas those with either low pre-map ICMV or low asphalt surface temperature are 'critical'. The remainder of the pavement lot is considered in the 'normal' region. Spot tests for asphalt density should be performed in supercritical and critical regions.
The SPARC-NCA approach was evaluated during the construction of a 100 m long, 4 m wide and 50 mm thick asphalt overlay. The asphalt mix used was AC14 with a polymer-modified bitumen (A10E). A Dynapac double-drum vibratory roller, equipped with an accelerometer (B & K Triaxial type 4506 B 003) and an IR thermometer (DFRobot TS01), was used for asphalt compaction. Prior to placing the asphalt overlay, the underlying support was pre-mapped to segment the pavement lot into three regions based on the CMV data, as illustrated in Figure 5(a). The asphalt overlay was then divided into 2 regions based on the surface temperature measured during the last roller pass, as shown in Figure 5(b). Figure 5(c) was created by superimposing Figure 5(a) on Figure 5(b). The asphalt overlay was compacted, with uniform compaction effort, using a predetermined number of roller passes, and its density was measured at 14 locations using a nuclear density gauge (NDG) following AS/NZS 2891.14.2 (Standards Australia 2013), as marked in Figure 5(c).
Figure 5(c) presents a schematic representation of the asphalt pavement testbed, displaying the 6 categories of regions (namely, 1 supercritical region, 3 critical regions, and 2 normal regions). The locations where the spot density tests (NDG tests) were performed fall within 5 different regions, as shown in Figure 5(c). The air void contents at these spots, as determined from the NDG measurements, are listed in Table 1.
Table 1 indicates that the spot located in the supercritical region achieved the lowest level of compaction compared to the other spots. This could be attributed to weak underlying support and the low asphalt temperature during compaction. Despite spots L2S1 (rank 6) and L2S3 (rank 7) being assigned to critical region 2, and spots L1S5 (rank 8) and L2S4 (rank 9) being assigned to normal region 2, they exhibit a maximum difference of only 0.8% in their air void contents. As such, the approach demonstrates considerable promise.
The SPARC-NCA approach leverages pre-mapping ICMV data and asphalt surface temperature to segment the asphalt layer into supercritical, critical and normal regions to determine the locations for spot density tests. This method utilises data that all current IC systems for asphalt compaction rollers can record. These systems include a data analyser attached to the roller that processes accelerometer data to produce ICMVs in real-time. Implementing the SPARC-NCA approach requires no additional sensor and the data analyser can be customised to integrate this approach.
To effectively implement the SPARC-NCA approach once it is added to any IC system, the following steps can be followed:
Establishing an acceptable correlation between ICMVs and asphalt density is crucial for using ICMVs as control parameters in asphalt compaction. However, such a correlation has not been demonstrated yet. This study examines the reasons hindering such a correlation and proposes an innovative approach to maximise the benefit of the current IC technology.
The analysis of accelerometer data from vibrating hammer tests showed that ICMVs are unaffected by changes in asphalt density at compaction temperatures. Additionally, modulus assessments of an AC14 asphalt mix with C320 bitumen, conducted at different temperatures and densities using unconfined compressive loading, revealed that asphalt mix temperature is the key factor influencing modulus. Tests above the bitumen's softening point temperature were not effective for estimating air void content using a modulus measurement. Consequently, neither asphalt modulus nor ICMVs are reliable for field compaction control.
To realise current IC technology benefits, this study proposes an approach (SPARC-NCA) that leverages pre-mapping ICMV data and asphalt surface temperature to segment the asphalt layer into supercritical, critical and normal regions, thereby determining the locations for spot density tests. This approach was validated using data from an asphalt overlay testbed construction. No additional sensors are required; only a customised data analyser is needed to implement this method.
This research work was part of a research project (Project number: IH18.05.1) sponsored by the SPARC Hub (https://sparchub.org.au) in the Department of Civil Engineering, Monash University, funded by the Australian Research Council (ARC) Industrial Transformation Research Hub (ITRH) Scheme (Project ID: IH180100010). The financial and in-kind support of the Australian Flexible Pavements Association (AfPA), the Department of Transport (DoT), Victoria and Monash University is gratefully acknowledged. The financial support of ARC is also greatly appreciated. The authors also wish to thank NTRO for their support in writing this follow-up paper. 124 – 134
