Modeling Reliability in Pavements (TRB 931133)

The correct application of reliability to pavement design is essential to the objectives of pavement design which are to produce quality pavements to serve the traveling public in comfort and safety, being built to be durable in service at a minimum life cycle cost. Reliability is a technical term which is defined in mathematical terms and is therefore objective in its application. As applied to pavements, it makes use of either empirical or mechanistic pavement performance equations to predict an expected value and variance of either the traffic or the distress for which the pavement must be designed. Explicit expressions are found for the quantities in terms of the expected values and coefficients of variation (cv's) of the independent variables which appear in the performance equation(s). Although ample data are available to determine these for all pavement independent variables, experienced engineers can estimate realistic values of the cv's almost as well as they can be measured. Several examples of the application of reliability to pavement performance models are given in this paper including:

1. Proper methods of analyzing field data to develop an empirical equation for the number of load cycles to reach pavement failure due cracking.
2. Pavement performance equations for cracking which incorporate reliability in determining the number of load cycles for which the pavement should be designed.
3. Methods of incorporating reliability into the design of overlays while taking into account the variability of the cracking that occured in the old pavement.

In all cases, the mathematical nature of reliability is demonstrated. It is because of its ability to be defined in precise, unambiguous terms that makes it simple to apply and possible to set objective standards for the performance of pavements in order to achieve its objective of providing an acceptable level of risk that pavements will carry the traveling public in comfort and safety while being built durably at minimum life cycle costs.