Predicting the number of accidents in rail-road level crossings using the Poisson zero-adjusted model and demographic factors approach

Rail transport holds a special place among transportation methods due to its cost-effectiveness and high safety. However, the intersection of road and rail networks at grade crossings compromises railway safety and results in significant financial and human losses. This study aims to identify the factors influencing the number of accidents at these crossings and provide a method for prioritizing grade crossings in Iran with a focus on improving safety. Based on data from the last 10 years leading up to 2022, the 1385 Persian calendar population blocks, and the characteristics of 130 grade crossings, this research utilizes the Zero-Inflated Poisson model, which offers higher accuracy compared to previous studies in Iran. The study also seeks to diversify the variables influencing accidents and explore their relationships with grade crossing incidents. The model consists of two parts: the Poisson section, which analyzes the number of accidents based on influencing factors such as road usage type, the male-to-female population ratio within a 5-mile radius, distance to the nearest intersection, total passing trains, total population within the 5-mile radius, and the number of tracks at the crossing. An increase in these variables correlates with more accidents, while a higher number of tracks at a crossing lead to fewer accidents. The second part of the model is the logistic section, which assesses the probability of zero accidents occurring. Factors such as the availability of an alternative route when the crossing is closed, the male-to-female population difference within a 5-mile radius, and the average daily road traffic all affect this probability. Increasing the first two factors raises the likelihood of zero accidents, while increased traffic lowers this probability