Human factor, road-rail safety policies,

available technologies at level crossing.

Towards a model to evaluate LCs risk.

Emilio CosciottiMassimo Costa

Salvatore De Marco Luciana Iorio

Roma, 7 giugno 2013

2

Index

UNECE

System definition

Identification of critical scenarios and associated incidents

Calculation of the frequency of accidents for “Investments of pedestrian at LC”

Calculation of the expected damage of accidents

Conclusions

3

4

UNECE – United Nations Economic Commission for Europe

WP1_ UNECE Road Safety Forum The WP1 is the ONLY Permanent intergovernmental body in the UN dealing with Road Safety ; Open not only to the UNECE members

CONVENTIONS • Conventions on Road Traffic 1968• Convention on Road Signs and Signals, of 1968• European Agreement supplementing the 1968 Convention on Road Traffic

Road Traffic Safety Frame provides a set of international agreed road

traffic recommendations aimed at the improvement of the efficiency

and safety of international road traffic

UNECE

5

SAFE SYSTEM APPROACH

Road Safety Policies have- for the next decades- ambitious targets based on the well known VISION ZERO.

The principles of

SAFE SYSTEM APPROACH (SSA)

already merged in the EU Commission strategy will also be reflected in the UNECE Road Safety Legal instruments in a comprehensive sound approach involving all the exogenous and endogenous factors of safe mobility .

UNECE

6

LEVEL CROSSING

A unique dangerous moment: Interaction of road / railways, two completely different modes.

Fatalities & Road Users

Attitude and risk demeanor at LXing are now under study at WP1 in collaboration with UIC and ILCAD

EXPERT GROUP

UNECE

7

EDUCATE IN EDUCATE OUT

Risk Social acceptance thwarts most of the road safety and mobility policies, and it is the most difficult to be dealt with

It is very common to be inattentive ( for many reasons, goals, mobile phones, music- short term goals i.e. going to school and pick up children at school, could be prioritize to wise long term goal, i.e. stay alive, that is why people take short cuts).

UNECE

8

THE LXING/ROAD MOMENTUM

The WP1 expects to have its Expert Group on road Safety at Level Crossing stepping in soon to start dealing with a crucial cumbersome challenge

ruling the interaction of two modes and two infrastructures to save lives

by upgrading mobility options

identify and evaluate key factors leading to unsafe conditions at level crossings, by bridging several factors such as the infrastructure, legislation, user behaviour, management, focusing on risk perception to mould the awareness, education and modelling enforcement.

UNECE

9

ESTABLISHMENT OF A GROUP OF EXPERTS ON SAFETY AT LEVEL CROSSINGS

UNDER THE MANDATE OF THE UNECE Inland Transport Committee,

a GROUP OF EXPERT will act and rules addressing key issues related to enhancing safety at level crossings

SOUGHT AFTER CROSS ACTION AMONG THE UNECE WP’s -Working Party on Road Traffic Safety (WP.1), the Working Party on Road Transport (SC.1) and the Working Party on Rail Transport (SC.2), bodies such as the European Railway Agency, in generalsafety specialists from the road and rail sectors so as to better understand the issues at this intermodal interface ( in accordance with ToR expertise of l UNECE member States, the European Union, Academia and the private sector

ECE/TRANS /WP.1 /2011/6

UNECE

10

The functions and elements of the system interact each other according to 3 types of installations considered: automatic LC with full barriers half-barriers automatic LC automatic LC with light signals and bells on the road side & St

Andrew’s cross

Road side protection and road side signals are operated by the following operation mode: passing train (automatic LCs) train control center (automatic LCs) users (private LCs) railway operator (manual LCs) other (without barriers LCs)

System definition

11

The system is also characterized by some boundary conditions: number of train per day number of tracks in the LC area maximum speed of the line average LC closure time per train road vehicles traffic pedestrian traffic influenced by the location of the LC in a urban

area or not road side visibility of the warning signs and of the signals railway side visibility of the LC area or road-side visibility of the

incoming train from the LC area

In some conditions aid equipment (e.g. CCTV) to the protection of the LC is required: crossing with barriers at a considerable distance intense heavy road traffic difficult and tortuous road layout obstructions on the normal road vehicle flow, due to crossings or

other things

System definition

12

The critical scenarios considered after some on site investigations are: investments of pedestrians at LC collision of a train with a vehicle trapped inside a LC with full-

barriers collision of a train with a vehicle dodging at a LC with half-

barriers

Identification of critical scenarios and associated incidents

Critical scenario

Frequency (accident/

period)

Damage (euro/

accident)

Risk of a LCDirect cause

Road vehicle in the railway area

Collision of a train with a

vehicle trapped inside a LC with full-

barriers

First event or incident

LC (unauthorised crossing/barriers striking by

road vehicles)

LC (unauthorised crossing by pedestrian or cyclists)

suicide

attempted suicide

person in the railway site/not at safety distance

Investments of pedestrian

at LC

Road user mistakes

Road user infringem

ents

Road user infringem

ents

Frequency (accident/

period)

Damage (euro/

accident)

unauthorised crossing of track by pedestrian

The incidents are identified by the Italian Railway Infrastructure Manager (RFI Spa) in its Safety Database (BDS)

Reference period of this study:July 2010 - August 2011

13

776655443322110 xxxxxxxy

railway side

visibility

railway side

visibility

Linear regression model for the estimation of the number of accidents y for a generic LC during a period equivalent to the one related to the study ( 1st attempt):

Calculation of the frequency of accidents for “Investments of pedestrian at LC”

urban/non urban areaurban/non

urban area

railway traffic

railway traffic

single/double track

single/double track

total faults/incidents

total faults/incidents

max speed of the railway line

max speed of the railway line

avg LC closure time

avg LC closure time

Coefficient Value t-Student

0

1 0.305 1.405

2 0.366 1.218

3 0.003 0.477

4 0.005 1.549

5 -0.334 -0.977

6 -0.051 -0.483

7 0.124 0.669

N. observations 23 2 0.878 F 16.465

Inv t 1.026

14

776655443322110 xxxxxxxy

railway side

visibility

railway side

visibility

Linear regression model for the estimation of the number of accidents y for a generic LC during a period equivalent to the one related to the study (1st attempt):

Calculation of the frequency of accidents for “Investments of pedestrian at LC”

urban/non urban areaurban/non

urban area

railway traffic

railway traffic

single/double track

single/double track

total faults/incidents

total faults/incidents

max speed of the railway line

max speed of the railway line

avg LC closure time

avg LC closure time

Coefficient Value t-Student

0

1 0.305 1.405

2 0.366 1.218

3 0.003 0.477

4 0.005 1.549

5 -0.334 -0.977

6 -0.051 -0.483

7 0.124 0.669

N. observations 23 2 0.878 F 16.465

Inv t 1.026

not useful

variables

not useful

variables

15

3322110 xxxy

railway side

visibility

railway side

visibility

Linear regression model for the estimation of the number of accidents y [n. accidents/period] (final attempt):

Calculation of the frequency of accidents for “Investments of pedestrian at LC”

urban/non urban areaurban/non

urban areamax speed of

the railway line [km/h]

max speed of the railway line

[km/h]

Coefficient Value t-Student

0

1 0.255 1.306

2 0.331 1.222

3 0.004 2.410

N. observations 23 2 0.868 F 43.710

Inv t 1.020

16

It is possible to calculate the damage per event [€/event], considering: number of fatalities multiplied for the Value of Preventing a

Casualty VPC (€ 1.500.000 in Italy) 10 serious injuries = 1 fatality 200 minor injuries = 1 fatality cost of environmental damage cost of material damage to rolling stock or infrastructure cost of the delays due to accidents (not available from the

databases)

It is possible to calculate the technological risk of the event:

R [€/period] = F [events/period] x D [€/event]

Calculation of the expected damage of accidents

17

The database was used to search a linear regression model for the prevision of number of accidents in a reference time period with regard to a collision of a train with a vehicle trapped inside a LC with full-barriers

The available database did not allow to obtain results, due to a low number of events in the period considered

The database will be extended in order to check the usability of such a type of model for this critical scenario and for the scenario of a road vehicle dodging at a LC with half barriers

Conclusions