A methodology for operations-based safety appraisal of two-lane rural highways : Application in Uganda

Sammanfattning: The majority of the road infrastructure in developing countries consists oftwo-lane highways with one lane in each travel direction. Operational efficiency of these highways is derived fromintermittent passing zones where fast vehicles are permitted by design to pass slow vehicles using the opposite traffic lane. Passing zonescontributeto reduction oftravel delay and queuing of fast vehiclesbehind slow vehicles. Thishoweverincreases crash risks between passing and opposite vehiclesespecially at high traffic volumesdue to reduction of passing opportunities.Reduction of passing-related crash risks is therefore a primary concern ofpolicy makers, planners, and highway design engineers. Despite the wide application of passing zones on two-lane highways, there is limited knowledge on the underlying causal mechanisms that exacerbate crash risks, and the essential tools to assess safety of the passing zones.This thesis presentsa methodology to appraisesafety of two-lane rural highways based on observed operation of passing zones.Theproposed methodology takes into accountthe impact of traffic and geometric factors onthe rate passing maneuvers end insidepassing zonesand in the no-passing zones, adequacy ofthe designpassing sight distance,and time-to-collision at the end of passing maneuvers.Thethesis is comprised offive papers addressing capacity and safety aspectsof passing zoneson two-lane rural highways. Paper Ipresents a review of the literature on capacity and safety of passing zones. Paper IIdiscusses adequacy of the design passing sight distance based on the sight distance required to complete a passing maneuverusing observed data. Paper IIIdiscusses formulation, estimation, and application of a model to predict the passing rate using geometric and traffic factors, and applications. Paper IVdiscusses risk appraisal of the passing process based on the probability to complete passing maneuvers with time-to-collision less than 3.0 seconds taking into account the accepted gap in the opposite direction and the passing duration. Paper Vdiscusses formulation and estimation of models to predict the probability and the rate at which passing maneuvers end in a no-passing zone, and applications.Resultsshow that passing zones of lengths between 1.30and 2.50km aregood for both operational efficiency and safety.Passing zones of lengths between 0.50and 1.30km exhibitincreasing crash risks resulting from delayed passing maneuvers thatend in the no-passing zone where the sight distance is limited to evadepotential collisions. Safety of these passing zones could be enhanced with additional signage to indicate the farthestpoint along a passing zone that maneuvers can be initiated so as not to end in a no-passing zone. Passing zones less than 0.50km compel drivers to commence passing maneuvers close to the beginning of the passing zone,and should be avoided during design for safety reasons.The results further show that the passing rate depends on the length of the passing zone, absolute vertical grade, traffic volume in two travel directions, directional split, 85thpercentile speed of free flow vehicles,and percent of heavy vehicles in the subject direction. The peak-passing rate also known as the passing capacity occurs at 200, 220, and 240vph in the subject direction for 50/50, 55/45, and 60/40directional splits, respectively. The rate at which passing maneuvers end in a no-passing zone increaseswith traffic volume and unequal distribution of traffic in two directions, absolute vertical grade, and percent of heavy vehiclesin the subject direction. The thesis furtherdiscusses practical applications of the study findings in highway planning and design to enhancesafety and improve operational efficiency of two-lane rural highways.