Response of Ultra-Thin Continuously Reinforced Concrete Pavement to traffic loading

Abstract

Ultra-Thin Continuously Reinforced Concrete Pavement (UTCRCP) is an innovative pavement type that has the potential to fulfil South Africa’s pavement repair strategy requirements. It consists of a 50 mm thick High Strength Steel Fibre Reinforced Concrete (HS-SFRC) layer that is reinforced with 50 x 50 mm aperture steel bar mesh, placed on a newly constructed or rehabilitated pavement. The current design procedure for UTCRCP was extrapolated from conventional concrete design procedures, incorporating the improved flexural properties of the HS-SFRC to design for fatigue cracking. However, the alternative nature of the thin HS-SFRC layer in comparison to a thick normal strength concrete layer has led to the proposal that the response of UTCRCP to traffic loading should be reconsidered to improve its design approach. A literature study revealed that the wheel load configuration, the relative stiffness of the concrete layer and its foundation and the complex response of foundation materials to stress influence the response of pavements to traffic loading. The effect of these aspects was investigated by making use of scaled physical modelling, as well as Finite Element (FE) modelling that incorporated Linear Elastic (LE) and advanced material models. The effect of load configuration on thin asphalt and thin concrete layers was investigated using LE FE modelling. A three-layer system of bound layer, base and subgrade was modelled. It was found that the response of a thin concrete layer is similar to that of a thin asphalt layer subjected to axle loading in that the maximum deflection is at the load location and that a hogging type deflection is induced in the axle centreline. The stress induced at the top of the concrete layer due to this hogging moment was high, indicating the necessity of including significant steel in both the transverse and longitudinal directions of UTCRCP. The difference in substructure response (in terms of horizontal and vertical displacement), modelled using single wheel or axle loading, showed that the compression of the substructure in the axle centreline can be critical, while it is ignored when load configurations are simplified to single wheel loading. A multivariable analysis of the concrete layer thickness and base material stiffness was conducted using LE FE modelling. A similar three-layer system of bound layer, base and subgrade was used. It was found that the location of the maximum deflection is in the axle centreline for pavements incorporating thick concrete layers, while the maximum deflection is in the wheel centreline for pavements incorporating thin concrete layers. The response of thin concrete pavements was more dependent on the substructure. The physical modelling consisted of 1:10 scale models tested in a geotechnical centrifuge. The models consisted of either thick concrete layer or thin concrete layer on compacted dry sand, as well as a thin concrete layer on a cement stabilized base supported by compacted dry sand. The most notable finding was that a rut forms in the wheel path of thin concrete layers on sand, although it cannot be observed from the surface. It was observed that the concrete layer rebounds when the pavement is unloaded, forming a gap between the concrete layer and substructure. An advanced material model for sand was used to explore the effect of incorporating the stress- dependent, elasto-plastic behaviour of granular materials in UTCRCP. A FE model, similar to the three-layer system of the LE FE modelling, was used and the base layer was modelled using a soil model called Nor-Sand, which is based on the critical state framework. The initial void ratio, lateral earth pressure at-rest and overconsolidation ratio were varied. It was found that the combination of the strain hardening and the stress dependence of the elastic material stiffness resulted in higher induced stresses close to the load location. The gap formation observed during the physical modelling was confirmed by the capability of Nor-Sand to model permanent deformation. Overall, the results of this investigation indicate that the response of UTCRCP to traffic loading differs significantly from that of rigid concrete pavements. The thin HS-SFRC layer is subjected to high tensile stresses and deflects significantly into the substructure at the load location. It is proposed that UTCRCP should be designed to limit rutting, as well as that stress dependent, elasto-plastic material models should be used to optimize its layer arrangement.