Development of Prediction Models for Joint Faulting using Long-Term Pavement Performance Database

10.6135/ijprt.org.tw/2013.6(5).658

Hsiang-Wei Ker；Ying-Haur Lee；Chia-Huei Lin

Concrete pavements ； Joint faulting ； Long-term pavement performance ； Modern regression ； Prediction models

International Journal of Pavement Research and Technology

6卷5期（2013 / 09 / 01）

658 - 666

英文

The main objective of this study is to develop improved faulting prediction models for jointed concrete pavements using the Long-Term Pavement Performance (LTPP) database. The retrieval, preparation, and cleaning of the database were carefully handled in a systematic and automatic approach. The prediction accuracy of the existing prediction models implemented in the recommended Mechanistic-Empirical Pavement Design Guide (NCHRP Project 1-37A) was found to be inadequate. Exploratory data analysis of the response variables indicated that the normality assumption with random errors and constant variance using conventional regression techniques might not be appropriate for prediction modeling. Therefore, without assuming the error distribution of the response variable, several modern regression techniques including generalized linear model (GLM) and generalized additive model (GAM) along with quasi-likelihood estimation method and Poisson distribution were adopted in the subsequent analysis. Box-Cox power transformation and visual graphical techniques were frequently adopted during the prediction modeling process. By keeping only those parameters with significant effects and reasonable physical interpretations in the model, various tentative performance prediction models were developed. The resulting mechanistic-empirical model included several variables such as pavement age, yearly ESALs, bearing stress, annual precipitation, base type, subgrade type, annual temperature range, joint spacing, modulus of subgrade reaction, and freeze-thaw cycle for the prediction of joint faulting. The goodness of fit was further examined through the significant testing and various sensitivity analyses of pertinent explanatory parameters. The tentatively proposed predictive models appeared to reasonably agree with the pavement performance data although their further enhancements are possible and recommended.