Analysis of crash frequency in motorway tunnels based on a correlated random-parameters approach

The paper provides an analysis of the frequency of total accidents (accidents involving material damage, physical injuries and fatalities), which occurred in 226 unidirectional motorway tunnels over a four-year monitoring period, based on unrelated and correlated random-parameter models. The so-called random-intercept model, in which only the regression intercept is assumed to be random, was also developed a priori for recording the random-effects (temporal correlations among accidents occurring in the same tunnel in different years). The independent variables were: tunnel length (L), annual average daily traffic (AADT) per lane, percentage of trucks (%Tr), presence of a sidewalk (SW), longitudinal slope (LS), and mechanical ventilation (MV). The comparison among the aforementioned three models showed that the correlated random-parameters model, which takes into account the cross correlation among the random-parameters, provided a better goodness-of-fit than the corresponding uncorrelated random-parameter and intercept-random models. This means that more precise estimations of accidents can be obtained when the random-parameters are assumed to be correlated in statistical analysis. The developed model also offers additional insights into showing how different combinations of parameters affect tunnel safety. In particular, through the correlation coefficient matrix of random-parameters, we found that the non-constant longitudinal slope (LS) alleviates the effect of the annual average daily traffic (AADT) per lane on increasing crash frequency. In addition, the presence of the mechanical ventilation (MV) in tunnels makes less significant the influence of AADT per lane on increasing crash frequency, too. The knowledge of these correlations may be useful for future applications, for example, for road engineers in designing tunnel. The model proposed can also be used by Tunnel Management Agencies (TMAs) for estimating possible variations in accident frequency in a specific tunnel due to modifications of traffic control systems.