考虑零膨胀特征和异质交互作用的高速公路隧道事故致因量化分析OA

Conventional accident modeling approaches struggled to effectively capture the zero-inflation characteris-tics of accidents and the heterogeneous interactions of risk factors,which undermined the credibility of conclusions regarding the influence of these factors on accident frequency.To address this issue,a multidimensional accident dataset was conducted for tunnel design features,traffic conditions,and pavement performance of mountainous expressways.The Lindley distribution and correlated random parameter modeling techniques were introduced to develop the random parameters negative binomial(RPNB)model,the random parameters negative binomial Lindley(RPNBL)model,and the correlated random parameters negative binomial Lindley(CRPNBL)model.The effects of risk factors on tunnel accident frequency were quantitatively evaluated by parameter estimation and marginal effect/elasticity analysis.The results show that the CRPNBL model,by incorporating the Lindley distribution,improves compatibility with zero-inflated data,the interactive effects of heterogeneous risk influences are captured by intro-ducing the random parameter matrix,thereby achieving the superior model fit;There are significant heterogeneous interactions among tunnel entrance presence,number of lanes,and proportion of articulated trucks,which produce notable synergistic effects on accident frequency.A 1％increase in traffic volume and the proportion of articulated trucks causes a 2.44％and 1.09％increase in accident frequency,respectively;The accident frequency at tunnel entrances is 1.21 times per kilometer higher than the average for other tunnel sections;Conversely,each 1％in-crease in tunnel spacing and skid resistance index(SRI)reduces accident frequency by 0.82％and 1.06％,respec-tively;Furthermore,the implementation of all-weather lighting reduces accident frequency by an average of 0.73％.That three-lane tunnel sections exhibit a 0.89％lower average accident frequency than two-lane sections.This re-search achievement helps to deeply understand the mechanism by which various risk factors affect tunnel safety,and formulate effective remedial measures.