Joint faulting of undoweled transverse joints contributes to longitudinal roughness, and subsequent user discomfort (poor ride quality) and needs for rehabilitation and repair. In this paper the authors propose two (based on support conditions) joint faulting prediction models based on the Strategic Highway Research Program (SHRP) - Long Term Pavement Performance (LTPP) database. The models consist of explanatory variables like shear capacity, pavement support, temperature, unconfined compressive strength, size ratios, and pavement response. Variables (from the database) like shear capacity, load transfer efficiency, pavement support, and crack widths were back calculated using mechanistic-empirical algorithms for more realistic estimates. Efforts were made to determine the form of the model and explanatory variables that made engineering sense, rather than blindly relying on statistical tools. The paper discusses the methodology used in selecting explanatory variables (independent variables) in the model development and results of the sensitivity analysis. The analysis of the data has shown that undoweledjoint faulting is a function of traffic, joint shear capacity, joint spacing, base type and pavement support. The paper provides a comprehensive reliability analysis (with illustrative example) of the models developed.
Read full abstract