Influence of Rain on Highway Breakdown Probability

Abstract

Capacity has been used to describe a deterministic value that represents the maximum volume of traffic supported by a road. Studies have pointed out the importance of not using a single value for capacity, but rather the concept of probability of occurrence of a traffic-flow breakdown. In this paper the probabilities of breakdown for a Brazilian highway under different weather conditions are compared. Data collected from inductive loop detectors and pluviometric data from automatic rain gauges are combined. Two methodologies of breakdown identification are then compared. The most consistent methodology for identifying breakdowns is used to generate breakdown probability distributions using the product limit and maximum-likelihood methods with the Weibull distribution. The results indicate significant differences in probability of breakdown for each studied climatic condition, including a maximum difference greater than 50% between dry and heavy rain conditions under the same traffic flow.