A risk-based approach to modeling life-cycle costs associated with warranty specifications for transportation infrastructure

To improve quality and reduce the overall costs, in recent years, many state highway agencies have started investigating innovative contracting methods, such as performance-based warranty contracting. This contracting method is structured to shift the performance-related risk from the agency to the contractor, by means of warranty provision. Even though the application of performance-based warranty contracting methods allows for a “win-win” situation, where agencies hedge the performance-related risk, and contractors have more flexibility in the design and construction processes, there are many concerns with its implementation. One of the most important concerns is how to quantify the risk cost. This dissertation is focused on the development of a robust and flexible methodological framework for quantifying the risk cost associated with warranty specifications for transportation infrastructure. The key components of this framework for studying performance warranties include: characterization of the warranty systems, development of probabilistic performance models based on the method of moments, formulation of the risk cost quantification models, and formulation of the models for determining the optimal design strategy and maintenance schedule. In this dissertation, three types of warranty systems are characterized and elaborated upon in detail: short-term, long-term, and maintenance performance warranties. To test the accuracy of the method of moments for developing reliability functions, the current AASHTO method for design of pavements is employed to provide a case study. The results from the comparison analysis of the methods of moments with Monte Carlo simulation indicate that the method of moments yields accurate predictions of the failure probabilities; in general, the quality of estimation improves as the order of the central moments in reliability indices increases. Finally, the methodology is illustrated with numerical examples to show that models for quantifying the risk cost associated with warranty specifications can be developed.