Metodología para el análisis de la distancia de visibilidad disponible en carreteras, la coordinación del trazado y su relación con la seguridad vial

Resumen

Geometric design is a factor of the utmost importance in highway safety. In this regard, sight distance is a fundamental principle to establish criteria both for the geometric design of new highways and to evaluate the alignment of in-service highways. The outcome of most current sight distance estimation methods may underestimate or overestimate the available sight distance. Only in a few cases, high-performance methods of proven accuracy have been applied to in-service highways, and none of them has been applied to establish linkage between sight distance and highway safety. Moreover, the relationship between sight distance and highway safety is not always clear and the literature shows diversity of results. The present work develops a methodology for the study of sight distance in three dimensions in order to assess the alignment of in-service highways through sight distance, from the point of view of highway safety. For that purpose, the high performance of a geographic information system is exploited, where more comprehensive analyses can be performed (including sight distance, design consistency, traffic volume, collision frequency, etc.). In total, 11 two-lane rural highways and a motorway section have been examined. First, the difference between the available sight distance and the stopping sight distance has been studied considering different elevation models as inputs. It was observed that the use of digital elevation models that comprise other elements besides the bare ground, such as vegetation, road furniture or buildings, produce more lifelike results. The differences between the available sight distance values on the models used were found to be statistically significant. These results were then used in conjunction with reliability theory to research the implications of available sight distance on highway safety. Two methods are presented contemplating the difference between the available sight distance and the stopping sight distance as the LSF to compute the probability of non-compliance. The variables involved are incorporated by means of a probabilistic approach instead of the classical deterministic models still in force in highway geometric design standards. The first method conducted a continuous study of the stopping sight distance along a highway section. The second one analyzes the stopping sight distance in horizontal curves from the available sight distance outputs in two and three dimensions. It was observed that the differences in the probability of non-compliance results, derived from each of the modelling methods contemplated, are statistically even more significant than in the values of the available sight distance. The results of the latter method are also used to fit a generalized linear model in horizontal curves that enables the evaluation of the accident risk caused by insufficient available sight distance. The probability and exposure variables used were statistically significant. Finally, the three-dimensional alignment coordination of highways and its possible implications on traffic safety are broached. For two-lane highways, the effect of the use of different elevation models on both the detection of these design shortcomings and on their characteristic parameter values was studied. In addition, the alignment coordination variables which most often favor the existence of sight-hidden dips have been identified as well as their relation to the characteristic parameters. On the other hand, the three-dimensional alignment coordination of motorways, where the horizontal and vertical alignments and the cross section are larger, is studied to develop design criteria that prevent the appearance of the main spatial alignment deficits related to sight distance and the perspective of the motorway.