CALIBRATING MECHANISTIC-EMPIRICAL PAVEMENT PERFORMANCE MODELS WITH AN EXPERT MATRIX

Pavement performance modelling is an important element in the proper management of pavement infrastructure. Various factors such as material properties, traffic loads and climate plus construction and maintenance schedules must be utilized to develop performance prediction and life-cycle costs. This paper describes a methodology for the calibration and validation of a mechanistic-empirical flexible pavement performance model. Mechanistic-empirical design methods combine theory based design such as calculated stresses, strains or deflections with empirical methods in which a measured response is related to structural thickness and pavement performance. The design system presented incorporates elastic layer analysis to determine pavement response. It uses cumulative ESALs, subgrade type and layer thickness to determine the most effective design. The input and output variables are probabilistic and they have been re-calibrated to extend beyond the range of applicability to the more extreme conditions (i.e. extremely low traffic volumes or extensive heavy truck traffic loading). The new mechanistic-empirical model also separates the environment and traffic effects of performance. In effect, the total pavement performance (P) is the cumulative effect of the damage due to the environment and the damage due to traffic. Hence, the regional differences (eg. between Southern and Northern Ontario in the example) can be quantified. The system calculates either roughness in terms of the International Roughness Index (IRI) or Riding Comfort Index (RCI) or in terms of performance as a Pavement Condition Index (PCI). Although this system was developed for Ontario conditions, the mechanistic-empirical performance model can be re-calibrated to apply to other conditions. Examples are provided to show the relative deterioration/performance curves for various design situations. For the covering abstract of this conference see ITRD number E201066. (A)