Three-Dimensional, Probabilistic Highway Design

Abstract

The potential usefulness of reliability analysis has recently been stressed in many engineering applications. Given the variability in the design parameters, a reliability-based probabilistic approach is well suited to replace the current deterministic highway design practice. However, progress in this regard is generally slow. In this study, the reliability analysis was used to estimate the probability of hazard (POH) that might result from insufficiency of sight distances. As an application, the available sight distance was checked against required stopping sight distance on an assumed road segment. Variation of the design parameters was addressed with Monte Carlo simulation using 100,000 sets of design parameters based on distributions available in the literature. A computer program was developed to use these sets of design parameters to calculate the profiles of available and required stopping sight distances in two- and three-dimensional projections as well as the profile of POH. The approach was applied to a horizontal curve overlapping with flat grade, crest curves, and sag curves in a cut section where the side slope would restrict the sightline. The analysis showed that the current deterministic approach yields very conservative estimates of available and required stopping sight distance, resulting in very low POH. The application example also showed the change of POH with the change of vertical alignment parameters.