Research Projects

Rigid pavements, in particular Jointed Plain Concrete Pavements (JPCPs), are used on major roads that carry disproportionately high volumes of heavy trucks, such as expressways. Their performance depends on the extent of transverse cracking that they develop over their service life. A satisfactory JPCP design method should be able to predict the development of transverse cracking under traffic loads. The current state-of-the-art procedure is Mechanistic-Empirical (ME) design, in which the stresses developed in JPCP sections under traffic loads are first evaluated using a linear elastic structural model, and these stresses are then correlated to transverse cracking based on empirical data collected from in-service field sections. This second, empirical step tends to be highly variable and subjective, with dozens of models being reported in the literature that can give wildly varying designs. To overcome this, there is a need to develop Mechanistic design procedures, which use non-linear structural models to directly predict transverse cracking based on the fundamental quasi-brittle fracture behaviour of cement concrete, rather than through indirect empirical correlations. While a handful of studies have demonstrated the effectiveness of this approach using a Cohesive Zone Model (CZM), this method tends to be very computationally intensive (taking several hours for a single design). Thus, despite its benefits, no JPCP design procedure in the world has adopted a direct Mechanistic approach. This project proposes to develop an advanced numerical technique that reduces the time to directly evaluate transverse cracking in JPCP from hours to minutes, thus making it feasible to be used for JPCP design. This technique divides the cement concrete slab into uncracked and cracked sections. Uncracked sections will be modelled using a computationally efficient reduced-order Finite Element (FE) model based on Kirchhoff-Love plates, while only a small part of the slab around the transverse crack will be modelled using 3D FE continuum Mode I CZM elements. A novel interpolation scheme, taking advantage of the kinematic behaviour of Kirchhoff-Love plates, together with the method of static condensation, will be used to couple these two models into a single, computationally efficient model. This technique is expected to facilitate the adoption of direct Mechanistic approaches into JPCP design methods, reducing the need for repeated collection of empirical field data and the subjectivity and variability in current ME design methods. As no existing FEM solver is capable of implementing the proposed technique, it will be implemented in an open-source FORTRAN/C++-based program. This program will then be used to demonstrate the Mechanistic design of a JPCP section using the proposed advanced numerical technique.