Properties of coarse unbound granular material used in road construction

Abstract

Road pavements are designed to ensure safe, accessible roads according to the expected traffic volume and composition and climate conditions. These parameters are considered in overall evaluations balancing the effects on monetary costs, traffic safety, health and work environment, external environment, climate, traffic management preparedness and more. The introduction of a mechanisticempirical (ME) pavement design system with performance prediction tools facilitates improved project specific analysis of deterioration mechanisms based on material, traffic and climate data. Hence, the pavement design may be optimized based on a defined balance between the predicted pavement performance and costs. To utilize the potential of such tools, comprehensive characterization of materials and other input data is a prerequisite. Coarse crushed rock (CR) materials with maximum particle size ranging from 125 mm to 250 mm are commonly used in structural layers in Norwegian road construction. Such gradings are uncommon outside of Scandinavia, and the existing literature is limited. Hence, there is a need to characterize the properties of coarse unbound granular materials used in road construction. Stiffness of materials is one of the main parameters in pavement analysis. Laboratory investigation to characterize the stiffness of pavement materials is commonly conducted with the repeated load triaxial test. However, characterization of standardized Norwegian subbase materials would require sample sizes of up to 1250 mm in diameter and 2500 mm in height. Hence, this study investigates stiffness characterization with repeated load triaxial testing on downscaled materials combined with in-situ fullscale characterization using the plate loading test. The laboratory investigation with large-scale repeated load triaxial test was conducted with 300 x 600 mm samples of downscaled open-graded and dense-graded materials. The upper sieve size of the downscaled materials was 22–63 mm, and the results indicated that downscaling of open-graded materials could be applicable for characterization of stiffness. However, the results for dense-graded materials indicated reduced stiffness with decreased upper sieves size. The in-situ field test was conducted with three parallel sections with different subbase materials: Section 1 with open-graded CR 22/125 mm (layer thickness 600 mm); Section 2 with dense-graded CR 0/125 mm (layer thickness 600 mm); Section 3 comprising dense-graded CR 0/32 mm as interlocking layer (layer thickness 100 mm) on open-graded CR 22/125 mm (layer thickness 500 mm). Sections 1 and 2 were designed to compare open-graded and dense-graded subbase materials with the same upper sieve size. Two weeks after construction, the stiffness of the dense-graded subbase was 15% higher as compared to the open-graded subbase, measured with the plate loading test utilizing a 300- mm plate. The dense-graded subbase exhibited a significant increase in stiffness by 14% after one year with no further compaction or traffic loading. Consequently, the stiffness for the dense-graded CR 0/125 mm was 36% higher as compared to the open-graded CR 22/125 mm after one year. Sections 1 and 3 were designed to compare the influence of the interlocking layer. In Section 1, CR 22/125 mm was compacted at a 600 mm thickness and CR 0/32 mm was applied in a 0–50 mm depth to obtain an interlocked surface for the coarse open-graded subbase. Section 3 was designed with a 100 mm layer of CR 0/32 mm which could be considered a separate upper subbase layer. Based on the plate loading test using a 300-mm plate, the stiffness measured in Section 3 was 19% higher as compared to Section 1 after two weeks. After one year, with no further compaction or traffic loading, the difference increased to 36%. The plate loading tests with 300-mm plate were combined with earth pressure cell measurements directly below the test plate. Measurements were conducted at a 600 mm depth, which was the interface between the subbase and subgrade. The plate loading test is conducted in two cycles, and measurements during the six loading steps of the second cycle were compared for Section 1 and Section 2. The stress level measured in Section 1 was on average 55% higher for all loading steps as compared to Section 2. The ratio of (change in measured stress at 600 mm depth)/(change in applied stress at the surface) was computed for consecutive loading steps. The average ratio was 12.4% in Section 1 and 9.7% in Section 2, indicating a better load distribution for the dense-graded CR 0/125 mm as compared to the opengraded CR 22/125 mm. Sections 1 and 2 were analyzed by use of Boussinesq’s linear elastic theory and the KENPAVE ME design and analysis software. The applied loads were equal to the load and circular area of the plate loading test. The average vertical stress at 600 mm depth for all loading levels based on Boussinesq’s theory was 52% and 72% as compared to the stress measured with earth pressure cells in Section 1 and Section 2, respectively. Computations of vertical stress by KENPAVE were on average 36% and 51% for all loading levels as compared to the in-situ measurements in Section 1 and Section 2, respectively. Hence, both methods estimated a lower vertical stress below the 600 mm thick subbase, as compared to the in-situ measurements with earth pressure cells. The in-situ plate loading test results and earth pressure cell measurements indicate a higher stiffness and better load distribution properties for the dense-graded subbase as compared to the open-graded subbase. However, the results are limited to one rock type and one subgrade type measured with the plate loading test, and further research should be conducted with complementary rock types, gradings, and subgrades. To exemplify estimates of emissions related to unbound granular materials in pavements, a case study was conducted on frost protection of roads. Frost design according to the current design regulations and frost quantities based on the period 1981–2010 were compared to frost design based on projections for 2071–2100. The Intergovernmental Panel for Climate Change publishes Representative Concentration Pathways, which are future scenarios for climate change based on population size, economic activity, lifestyle, technology, energy usage, land use patterns, and climate policy. Two scenarios were employed in the study, with RCP8.5 reflecting “business as usual” and RCP4.5 representing an intermediate scenario between RCP8.5 and RCP2.6, which is the 2°C goal. In combination with the RCPs, three alternatives were computed: Alternative 1 with fossil fuel; Alternative 2 with electric machinery and biofueled trucks; Alternative 3 with transition from fossil fuel to all-electric machinery. The emission models were based on LCA tools VegLCA and Stein LCA. E6 Kolomoen–Moelv (E6 K–M) which is a 40 km four lane highway in inland Norway, and E18 Tvedestrand–Arendal (E18 T–A) which is a 22 km four lane highway on the Norwegian southern coastline were selected for the case study. The estimates with frost quantities for the period 1981–2010 and Alternative 1 with fossil fuel were 29,017 and 6,298 tonne CO2-equivalents for E6 K–M and E18 T–A, respectively. For Alternative 1, E6 K–M reductions were estimated at 31% and 46% under RCP4.5 and RCP8.5, respectively. E18 T–A reductions with Alternative 1 were estimated at 47% and 65% under RCP4.5 and RCP8.5, respectively. For Alternative 3, the estimated reductions under RCP4.5 and RCP8.5 were 71–78% and 77–85%, respectively. The 130 mm difference in pavement depth for E18 T–A under RCP 4.5 as compared to RCP8.5 was estimated at 50 kg CO2-equivalents per meter road for Alternative 1. Thus, if the implementation of ME design reduces material usage or increases the expected service life for most of the future public road projects, it could lead to considerable accumulated reductions in emissions related to pavement materials.