Comparing failure mechanisms of ultra-thin continuously reinforced concrete pavements (UTCRCP) under cyclic and monotonic loading

Abstract

Conventional concrete road pavements are widely used and their behaviour is reasonably well understood and readily predictable, generally by means of semi-empirical design methods. An alternative pavement type, referred to as Ultra-Thin Continuously Reinforced Concrete Pavement (UTCRCP), comprises a concrete pavement substructure with a 50 mm thick high-strength concrete overlay, heavily reinforced with steel mesh and fibres. In contrast to conventional road concrete road pavements, the thin and heavily reinforced overlay behaves in a flexible manner. Several uncertainties remain regarding this system’s behaviour and analysis in the context of pavement design and scope exists for the improvement of design methods of UTCRCP. These uncertainties include (but are not limited to) the bond and possible debonding between the concrete overlay and the material below, the build-up of pore water pressure within the soil during cyclic loading and the behaviour of the structure after a load higher than the assumed design load has been applied. Improvements can be affected when behavioural data from such pavements are available, but relevant data from the literature are limited. UTCRCP have recently been studied at the University of Pretoria using 1:10 scale physical models. Data from such studies have the potential to contribute to UTCRCP design methods. However, the amount of data currently available from these studies are limited and a need for further work exists. To add to the data base of UTCRCP behaviour in models, experiments were conducted in which 1:10 scale UTCRCP pavement models were tested under cyclic and monotonic loading. Two parallel strip loads were used to represent the wheel loads of a single axle, modelling pavement behaviour idealised to plane-strain conditions. Tests were carried out using UTCRCP model pavements with layerworks constructed from a fine silica sand and a graded gravel respectively. Experiments were conducted at optimum moisture content and under saturated conditions. The fine silica sand was used for the observation of typical failure mechanisms while the gravel was used as a more realistic road pavement material. The settlement under load application at the surface of the pavement was monitored with a Linear Variable Differential Transformer (LVDT) and Digital Image Correlation (DIC) was used to track deformation fields within the pavement surface during testing. Monotonic load tests were first carried out to determine upper bounds of the failure loads for each pavement model type. Thereafter, various percentages of the upper bound loads were applied to the road pavements under cyclic loading until excessive deformation had occurred or until deformation had stabilised. Deformation mechanisms from the monotonic and cyclic load tests were then compared at several settlement magnitudes. (“Settlement” refers to the settlement of the strip load bogey and can be also be viewed as the depth of imposed rutting). This was done to determine whether a load applied monotonically to cause a certain amount of settlement (rutting), would result in a deformation mechanism in the road layers that would be similar to that from cyclic loading, given the same amount of settlement (rutting). It was found that settlements remained small until a certain number of load cycles had been imposed. The settlement trend was found to be linear when plotted against the logarithm of the number of load cycles, matching the way in which creep movement proceeds. However, once a certain number of load cycles had been imposed, the settlement trend accelerated. The number of load cycles at which the settlement accelerated was strongly dependent on the applied load magnitude. This behaviour is typical of the deterioration of a road pavement which demonstrates that the models have value in simulating actual road behaviour and that they are therefore suitable to be used to study elements of pavement behaviour. When the deformation mechanisms were compared between cyclic and monotonic tests at the same settlement values, it was found that the mechanisms were different. Although slip mechanisms developed for both monotonic and cyclic loading at higher load magnitudes, the strain distribution resulting from cyclic loading was found to be more gradual and generally extended wider and deeper than the failure zone which resulted from a monotonically applied load. Under monotonically applied load, failure generally occurred along well-defined shear surfaces with little straining of the materials contained within the shear bands. This could be due to particle reorientation happening within each load cycle as load is relaxed, causing the slip plane to be disrupted during cyclic loading. Slip planes were constantly changing, resulting in slip bands over a wider and often deeper extent in the cyclic tests. An exception to this were the tests conducted