Back
Report numberRA-2005-70
TitleRisk analysis on motorways
SubtitlePart 3-Accidents models for Flemish motorways
AuthorsFrank Van Geirt
Erik Nuyts
Published byPolicy Research Centre for Traffic Safety 2002-2006
Number of pages56
Date01/10/2005
ISBN
Document languageDutch
Partner(s)PHL
Work packageOther: Infrastructure and space
Summary

We have built cross-sectional models to predict the expected number of accidents on Flemish freeways as a function of the infrastructural design of the road segment. Data on infrastructure are collected by ARAN of AWV, accident data are based on geo-coded injury accidents from 1996 to 2001. The best predictions were obtained when the road segments were split up into three groups: entrance ramps and their neighbourhood (up to one km from the entry ramp), exit ramps and their neighbourhood and the segments that are more than one km away from both entrance and exit ramps.

 

For every group four models are built: all injury accidents, deadly accidents, severe injury accidents and light injury accidents. The models for accidents with severe injuries or deadly injuries have few explaining variables, or did even not always converge, probably due to a limited number of this kind of accidents.

 

For all groups and for all models it is found that the number of accidents increases with the traffic intensity.

 

In the model including all injury accidents on segments away from entrance and exit ramps, it is found that traffic safety increases with larger lane width, larger shoulder width and larger redress lane width. Clearly, these results are due to more free space to manoeuvre in dangerous situations. Segments with four driving lanes are safer than segments with two or three driving lanes.

 

The entrance zone is defined from one kilometre before the beginning of the entrance ramp to one kilometre after the end of the entrance ramp. In general, this zone is more dangerous than the link segments. Within this zone, the segments beside the entrance lanes are the most dangerous. Weaving from passing and entering traffic is probably the explanation for this result. In the model including all injury accidents again it is found that traffic safety increases with decreasing traffic intensity, larger lane width, larger  shoulder width and larger redress lane width. We also find that road segments with a maximal speed limit of 100 km/u are significantly more dangerous than segments with the usual speed limit of 120 km/u.

 

The exit zone is defined from one kilometre before the start of the exit lanes to one kilometre after the end of the exit lanes. In general, exit zones are more dangerous than the link segments.

 

Also for the exit zones the model including all injury accidents predicts fewer accidents with decreasing traffic intensity, larger shoulder width and larger redress lane width. Segments beside exit lanes and especially segments after exit lanes are more dangerous than segments before exit lanes. We have no explanation for this result. The number of accidents is lower when the speed limit increases.

 

Cross-sectional models do not provide causal relationships, only correlations. Yet, we think that the increased safety on segments with larger redress lanes and segments with larger shoulder lanes (perhaps with a maximum width 3.7 to 4m) are important results for the government. The finding that segments with lower speed limits are still more dangerous than segments with the usual speed limit of 120 km/u deserves a special investigation.

DownloadPDF icon RA-2005-70.pdf
Lijn

Mission

The Policy Research Centre for Traffic Safety carries out policy relevant scientific research under the authority of the Flemish Government. The Centre is the result of a

cooperation between Hasselt University, KU Leuven and VITO, the Flemish Institute for Technological Research.

Partners

Leuven vito