A probabilistic model for joint-movements in jointed concrete pavement

Joint movement can be considered to be the result of horizontal movements of adjacent slabs to a joint. Slab movements are caused by the change of slab temperature and moisture level, and are resisted by subbase friction. The AASHTO model assumes that slab length equal to joint spacing. However this assumption is valid only if there are no penetrated transverse cracks between joints, and mechanistic discontinuity at every joint is satisfied. The occurrence of joint freezing and transverse cracking causes the variation of integrated slab length (ISL) in a given pavement section. And the variation of ISL may cause the variation of the magnitude of joint movement. But these have not yet been considered in the estimation of joint opening. In this study, a model (Lee-Stoffels model) is proposed to predict the probability of the magnitude of joint opening. In the Lee Stoffels model, joint freezing and transverse cracking is considered as factors to induce the variation of joint movement in a given pavement. The probability of a pair of ISLs adjacent to a joint is estimated based on the probabilities and distribution characteristics of joint freezing and transverse cracks for a given pavement section. The magnitude of joint opening is calculated by using a mechanistic model (SSM Program) for the given pair of ISLs (distances from a joint to adjacent mechanistic discontinuity). The Lee-Stoffels model is validated based on the comparison between predicted and measured joint openings in sixteen the Long Term Performance Pavement Seasonal Monitoring Program (LTPP SMP) pavement sections. The Lee-Stoffels model has advantages over the AASHTO model, since it is capable of estimating the magnitude of joint opening along with its probability. The reasonably largest joint opening (joint opening at the 90th to 95th percentile) estimated from Lee-Stoffels model may be used as a better criterion for joint seal design maintenance than the estimation of joint opening based on the AASHTO model.