Characterizing the roughness of Kansas PCC and Superpave pavements.

摘要

Accurate prediction of pavement performance over long time horizon represents a critical issue in the pavement surface type selection process that is performed by the Kansas Department of Transportation (KDOT) via life-cycle-cost analysis. Thus, reliable prediction of roughness progression on Portland Cement Concrete (PCC) pavements constitutes a very important issue since the current model used by KDOT is based on the pavement serviceability.; In phase I of this study, dynamic Artificial Neural Network (ANN) and statistical analysis approaches were used to develop roughness (International Roughness Index, IRI) prediction models for newly constructed Jointed Plain Concrete Pavements (JPCP) in Kansas. Database used in the model development process included construction and materials data as well as other inventory items such as traffic and climatic-related data. Utilizing a two-stage training approach, two time-dependent ANN-based roughness prediction models were developed. Both models were able to project the time-dependent roughness behavior with reasonably high coefficients of determination, R2 = 0.90 and R2 = 0.86, respectively. Similarly, using regression analysis, a SAS-based time-dependent roughness prediction model (R2 = 0.76) was also developed. To validate the developed models, IRI values were predicted for the time horizons that have not been encountered in the development stage. Using the developed ANN- and SAS-based models, a thorough sensitivity analysis was also performed. The analysis quantified, to some degree, the impact of various key input parameters on JPCP roughness.; In phase II, concrete material and mixture data along with the static ANN methodology were used to develop three initial roughness IRI prediction models for constructed rigid pavements (JPCP). Similarly, Superpave mixture data was used to develop a Superpave initial roughness IRI prediction model. The developed models projected the anticipated initial IRI roughness value for various PCCP and Superpave projects with a reasonable accuracy.