Influencing-Factors Analysis of Freeway Incident Duration Considering Heterogeneity in Means

Abstract

To enhance the efficiency of freeway emergency management and mitigate the occurrence of secondary incidents, this study builds a random-parameter (RP) accelerated-failure-time (AFT) model with heterogeneity in means (RPHM) to investigate the influencing factors of freeway incident duration and their potential heterogeneity. Given the incident data collected from the G2504 Hangzhou freeway in China, the proposed RPHM AFT model is compared with the AFT model and the RP AFT model. Results show that the RPHM AFT model has the best fitting accuracy, and can better reveal the potential heterogeneous effects among influencing factors. Fourteen covariates, including collision with fixed objects, scratches, rollovers, and fire, significantly affect the incident duration. Among all factors, the parameter of lane-closure effect on incident duration exhibits spatial variation. Scratches, heavy vehicles, multi-vehicle incidents, rainy conditions, and Lane-1 closure display random characteristics. Collision with fixed objects has a significant heterogeneous effect on the mean value of the index RPs of heavy vehicles and positively affects the incident duration. Furthermore, the temporal factor of the peak-hour index does not have a significant impact, indicating that the operation and management of domestic freeways are significantly different from foreign freeways and previous studies may not be directly applied to China. The research results provide a theoretical foundation for relevant departments to formulate differentiated emergency-response plans for freeways. Moreover, these findings are of great significance in improving the safety-management level of China’s freeways.