Research and Definition of Car Accident Scenarios for Roadway Safety Management of S:s: 372 "Telesina".

The present paper has been developed on traffic, car accident and geometric data surveyed on the 61 kms of the italian rural road SS (Strada Statale) 372 “Telesina”. The SS 372 starts from the motorway exit of Caianello, located on the Rome - Naples highway and ends in the town of Benevento. It has a variable cross section but in average it is about 9 m wide. The road alignment is typical of rural roads running in the hilly countryside of Campania region, southern Italy. The vertical alignment is made with several crests among upgrades but their values are lower than 3,5%; in the final part of the track heading Benevento there is just one 7% upgrade. All along the track there are many motorway exits made on different levels with small lenght acceleration and deceleration lanes. Car accident data related to a 11 year period ranging from 1993 to 2003 have been collected and analysed from Police reports. The resulting database is made of about 180 accidents. The ANAS statistics on 3 sections of the road track have been collected too, being ANAS the Italian society deputed to the management of most part of rural roads. So, it has been possible to share the whole track in 3 parts. The ADT (Average Daily Traffic) has been supposed to be constant within each part of the road. The preliminary general analyses done on the whole track showed not so high accident indexes. Nevertheless, the successive scenario research showed some specific matters. First, a database of the road has been built. The database has been shared in three parts: 1) geometric data; 2) traffic data; 3) accident data. The first part has been built with the help of the existing topographic data: the whole track has been summarized with specific regard to horizontal and vertical alignment and location of motorway exits. The second part has been built on ANAS data as previously stated. The third part has been built on the police data. Accident location, traffic and road pavement conditions have been filed. Finally, a matrix data in which the following quantities have been associated to every accident has been built: “heading (to Benevento or Caianello) of the car which caused the accident, extracted from part 3; light conditions (day or night) at the moment of the accident, extracted from part 3; road pavement conditions (wet or dry) at the moment of the accident, extracted from part 3; horizontal alignment (curve or tangent), vertical alignment (positive, negative or plan) in the area of the accident, extracted from part 1 and finally the sight distance of the point in which the car accident happened (yes or no) calculated with informations provided from parts 1 and 3. Part 2 data have been just used to share the road into three homogeneous traffic conditions. Not all components have been used to build scenarios. Finally, 96 different possible scenarios have been built. As soon as each scenario has been characterized on the temporal and spatial point of view their hazardousness has been evaluated with the accident index. Results showed that all 3 most dangerous scenarios included the component “lack of passing sight distance”. Details showed that the first and most dangerous scenario is made with the following components: “daytime, wet, curve, lack of passing sight distance”; the second one “daytime, dry, curve, lack of passing sight distance”. So, it seems clear that the component ”lack of passing sight distance” has a prevalent influence on the hazardousness of scenarios; in order to improve safety it should be necessary to work on this component for instance by enlarging the shoulders of the road or making adequate excavation in curve