Work Plan for the CSM working group by neolledivine

VIEWS: 0 PAGES: 9

									                                           Safety Unit

                               CST Working Group

 Introduction to approaching CSTs for parts of
               the railway system




introduction to approaching CSTs for parts of the railway system_10   Page: 1/9
Document elaborated by:                                  European Railway Agency
                                                         160 Boulevard Harpignies
                                                         F-59300 Valenciennes
                                                         France

Released by:                         Anders Lundström

Reviewed by:                         Thierry Breyne, Roberto Piazza, Anders Lundström

Author:                              Marcus Andersson

Version:                               1.0

Date:                                 April 2007

Type of document:                      Working document

Status of document:                    Final


Amendment records
Version      Issuer Modified sections                                                 Distribution
0.1          MA     All new sections – first version for internal review.             Internal draft
0.2          MA     All sections reviewed according to TB and RP                      Internal draft
                    comments
0.3          MA     Minor changes                                                     Internal draft
1.0          MA     Minor changes after revision by AL                                DG TREN

References
N° Description                                                           Reference               Version
/1/ Directive 2004/49/EC on safety on the Community’s                    OJ L220, 21.6.2004,
    railways (Railway Safety Directive)                                  p.16 (Corrigendum)
/2/ Mandate to the ERA for the development of the 1st set                04/49-MA02              EN04
    of Common Safety Targets
/3/ Mandate to the ERA for the development of the 1st set                04/49-MA02              EN04
    of Common Safety Methods




introduction to approaching CSTs for parts of the railway system_10                               Page: 2/9
Table of Content
1. Introduction ............................................................................................................................ 4
2. Considerations for the development and application of CSTs to macro parts ....................... 5
3. Some first ideas on approaches to safety targeting for macro parts of the railway system ... 6
4. Some first conclusions ........................................................................................................... 8




introduction to approaching CSTs for parts of the railway system_10                                                                Page: 3/9
1. Introduction
The work in developing Common Safety Targets for the railways in the EU Member States is
progressing. The main focus and direction of the work up to today is to create Common
Safety Targets for the railway system as a whole. This focus is logical since the analysis so far
has given that it is only at this very high level that an approach to safety targeting is possible
that satisfies the prerequisites stated in the Safety Directive /1/ that the CSTs should:

-   Be based on current safety performances.
-   Ensure that the level of safety is not reduced in any Member State.

Already when analysing the safety performances at this very high level representing the
overall outcomes at Member State level, several problems arise in regards to the availability
of useful data representing the current safety performance. These problems arise from a lack
of official harmonised data representing a long enough time series and a large enough
statistical population to make robust estimates of the safety performances. Knowing this it is
easy to imagine that these problems would be even larger when analysing the safety level at
more detailed levels and for parts of the railway system.

Effectively, this situation makes it very complicated, not to say virtually impossible, to today
create CSTs for parts of the railway in a similar way as for the railway system as a whole in
the Member States. Disaggregate data for parts of the railway system are not available and
will not be available to a sufficient extent in the short term future since this disaggregation is
as of today not included in the CSIs. This also creates problems in statistically monitoring the
achievement of any CSTs directed to parts. An amendment of the CSIs in an absolute
necessity for the methodology to be feasible, and even in this case it will be a complicated
exercise to try and derive CSTs from very limited statistical resources.

It is however indicated in the CST Mandate (section 2.2) /2/ that the CSTs should cover both
the system as a whole as well as parts of the railway system. Also, the Safety Directive
nevertheless clearly states that:

“common safety targets (CSTs) means the safety levels that must at least be reached by
different parts of the rail system (such as the conventional rail system, the high speed rail
system, long railway tunnels or lines solely used for freight transport) and by the system as a
whole, expressed in risk acceptance criteria”

In order to satisfy the demand for “CSTs for parts”, it is obvious that it is necessary to first
define a feasible approach to develop these CSTs. It is also important that the approach
chosen will be of added value to the stakeholders. Regardless of what approach is preferred it
is important as a first step to define for what parts of the railway system CSTs should be
developed. The list of parts mentioned in the safety Directive is not necessarily exhaustive,
and what parts that are suitable for CSTs should be analysed.

In principle, two different possible ways of developing CSTs for parts can be identified:

1) Define the CSTs for parts in a similar way as for CSTs for the system as a whole, eg by
measuring the current safety performance for the parts in question and deriving the CSTs
from the measured values. This would first require as stated above an analysis of more


introduction to approaching CSTs for parts of the railway system_10                        Page: 4/9
precisely what CSTs should be developed in order to amend the list of CSIs accordingly.
After this has been achieved, it will be necessary to collect a number of years of data in order
to have a statistically significant population. The time frame needed for this would be
significantly larger than the time frame for the CSTs for the whole systems as the number of
events collected for each of the parts will naturally be smaller than for the aggregate of events
for all parts.

2) Use of a different approach based on assumptions, analysis of existing targets and other
kinds of information and reasoning. This approach would rather focus on defining targets for
the safety levels of parts of the systems on a predictive basis, using the target as risk
acceptance criteria for risk analysis procedures. This makes this approach fundamentally
different from the CSTs for the system as a whole as they are being developed, where the
CSTs are derived using accident statistics and the verification of achievement relies on
measurement of the achieved safety level rather than on risk analysis results. This approach
links naturally with the development of CSMs.

This document is meant as an introduction to possible ways of creating CSTs for parts of the
system as is requested in the safety directive and also to explain the present limitations in
possible approaches. Another frequently discussed issue in the field of CSTs have been how
to deal with safety targets and requirements for the construction of new and significantly
modified lines. It is therefore also discussed whether these safety requirements as design
criteria on the level of macro parts could relate to or be merged into the CSTs for parts.

2. Considerations for the development and application of CSTs to
macro parts
As stated above, the Safety Directive suggests a number of parts for which CSTs should be
developed:

         -    The conventional rail system
         -    The high speed rail system
         -    Long railway tunnels
         -    Lines solely used for freight transport

The purpose of defining CSTs specifically for these parts is not defined, and it is not
expressed if or why the safety target of these parts should be different from each other and
from the CST for the system as a whole. If it would be possible, through already available
statistics, to base CSTs for these parts on recorded safety performance it is rather likely
though that rather large variations between the categories would be discovered. For instance,
the high speed rail system is generally made up of modern technologies, infrastructure and
rolling stock whereas the other categories would generally represent a mixture of modern and
older technologies. Given the development of railway safety it is not unrealistic to assume
that lines using modern technologies would generally be safer than older lines.

Some issues that arise when studying these categories is that some of the categories for the
individual risk, as stated in the Safety Directive Art 7 (4), seem less relevant to apply for
certain parts. For example it does not seem valid to apply a safety target to the individual risk
to passengers on lines solely used for freight traffic. High speed lines rarely have level cross-
ings why this risk category would be of marginal importance to target specifically for these
lines.

introduction to approaching CSTs for parts of the railway system_10                       Page: 5/9
The above mentioned limitations as to disaggregated data for parts does create problems not
only in setting the CSTs themselves, but also in monitoring and verifying their achievement.
Today, disaggregated accident data is not collected and even if it was decided to collect this
data, the return of events would most likely be so marginal even if studying the safety
performance over a very large time period that it could be a pointless exercise trying to
specifically monitor the safety performance of parts. This is of course especially true in the
Member States with limited rail operations, and it is a condition that puts significant limits on
the feasibility of using the first approach as described in the previous section. The problems
encountered in the work regarding the CSTs for the system as a whole regarding scarcity of
data and statistical significance is the main issue to resolve and a large part of the resources
are used to find ways to deal with this. These problems would naturally be amplified for the
parts of the railway system, making the development and especially monitoring of CSTs for
parts difficult if not impossible if this approach is used. Also, it is hardly reasonable to
demand from Member States to deliver accident statistics disaggregated in to the different
parts especially considering the limited added value of this data.

It makes very little sense creating CSTs that there is no method or procedure to apply to
verify that the desired safety level has been met. For macro part CSTs to be useful and of
added value they need to be applicable to a framework where they can be used as verifiable
and useful criteria in a safety related framework.

3. Some first ideas on approaches to safety targeting for macro
parts of the railway system
One way of incorporating the CSTs for parts into a useful context could be to instead regard
these CSTs as requirements for the safety level to be reached in the design of new railway
lines and perhaps also when large changes are made to existing lines. This approach has a
number of potential benefits but is also a fundamentally different type of CST that will not
automatically fit in with the CST concept as it is currently described:

The issue that safety targets (or safety requirements) for the macro parts of the railway
systems needs further development in the CSM and CST concepts has been raised, and a task
force has been established already to deal with this in the CSM work. Developing “CSTs for
parts” in this way would help to breach this gap. These CSTs would also work in the direction
of harmonizing the overall safety levels of the Member State railways since as more lines are
built, changed or upgraded with the same safety target in the different member states, the
overall safety in the Member States will progressively stabilize at a harmonized level. This
process of course requires a long time horizon.

Another benefit is also that if the CSTs for parts also can be used as a safety requirement for
the design phase, this will allow them to be used within the CSM framework as harmonized
risk acceptance criteria in the risk assessment procedure. Ultimately, this will fulfil the
requirement that it should be possible through the use of CSMs to verify the achievement of
CSTs as stated in the CSM (2.3.1) and CST (2.3.2) mandates /2,3/.

In developing CSTs for parts of the railway system, the limitations earlier discussed will have
to be taken into account. Using statistics on accidents is not possible with the current CSIs.
Risk categories that are not relevant should be excluded. For new systems it should also be


introduction to approaching CSTs for parts of the railway system_10                       Page: 6/9
taken into account that the safety level is generally higher than in comparison with the
existing lines that constitute a mixture of old and new infrastructure and technology. Possible
sources for inspiration worth studying could be:

        -   Existing risk acceptance criteria for these parts in different Member States (if such
            exist).
        -   Design risk acceptance criterias for new railway lines or other transport
            infrastructure used in planning in different member states.
        -   Project based design risk acceptance criterias (eg Eurotunnel etc).
        -   Look at the NRVs for the “best in class”, eg to look at the best level that has been
            documented as possible to achieve in the operations phase and use this as a design
            criteria for new and significantly modified lines.

Regarding the list of macro parts and the risk categories it should be possible to exclude the
irrelevant risk categories. In some cases, these exclusions are obvious (eg passengers for
freight lines, level crossing users in tunnels). It is reasonable also to exclude level crossing
users from high speed lines, since there are generally no level crossings on high speed lines,
especially when looking at category 1 high speed lines. For tunnels; the category “others”
should probably also be excluded since this category refers mainly to persons outside of the
railway system that are affected by a railway accident. In the case of tunnels, any person
inside the tunnel not belonging to the category “staff” or “passengers” would have to be
considered an “unauthorised person on railway premises”. A first rough filtering excluding
the most obvious categories would still leave 20 CSTs to be developed.

Table 1. Relevant risk categories for macro part CSTs
Macro part                                      Relevant risk categories
The conventional rail system                    Passengers
                                                Staff
                                                Level crossing users
                                                Unauthorised persons on railway premises
                                                Others
                                                Societal/Collective risk
The high speed rail system (category 1)         Passengers
                                                Staff
                                                Unauthorised persons on railway premises
                                                Others
                                                Societal/Collective risk
Long railway tunnels                            Passengers
                                                Staff
                                                Unauthorised persons on railway premises
                                                Societal/Collective risk
Lines solely used for freight transport         Staff
                                                Unauthorised persons on railway premises
                                                Level crossing users
                                                Others
                                                Societal/Collective risk




introduction to approaching CSTs for parts of the railway system_10                        Page: 7/9
Whether or not additional macro parts should be added or whether perhaps the list above
should be shortened should be subject to further analysis. A further issue to tackle is then also
to make clear definitions of the macro parts (eg how long is a “long” railway tunnel, what
lines should be regarded as high speed lines etc).

4. Some first conclusions
The experience so far from the CST work is that it seems difficult to foresee a way to derive
and monitor achievement of CSTs for parts of the system in a similar way as is being
developed at Member State level. However; a wish for harmonised defined risk acceptance
criteria for new lines has been expressed in the course of the work. It could therefore be a
feasible option to develop the CSTs for parts rather as target safety levels for the construction
of new lines. Whether this is a sensible way to approach the CSTs for parts would need some
more analysis.

In deciding the direction of work in developing CSTs for macro parts, it is important to
discuss the potential added values and the possible problems related to this issue: Some of the
most important ones are listed below:

“Pros”:
   + Developing “CSTs” that can be used as a “common safety requirement” for the CSM
   framework will actually help to satisfy the statement that it should be possible through
   CSMs to verify the achievement of CSTs. Doing this will help fill the gap between CSMs
   and CSTs and create a more direct link such as is implied in the Safety Directive.
   + Having a harmonized safety target as risk acceptance criteria for the construction of
   new lines will aid the harmonisation of the safety levels between Member States since all
   new lines will be built according to the same safety target. In a long time horizon, this
   means that the existing differences in safety performances will level out as new lines are
   built.
   + Within the CSM (and CST) work, the question has been raised by some NSAs and
   other organisations that it might be more important to be able to evaluate the overall
   safety of a new or significantly modified system rather than having evaluation criterias for
   technical details within the system. Adding macro level CSTs would satisfy this demand.

“Cons”/ potential problems
   - “CSTs” that are used as a design risk acceptance criteria for new lines can not meet
      the SMART criteria since they are inherently different from the CSTs at Member State
      level focusing on post evaluation of achieved safety rather on predictive evaluation.
      This means that; if it is an absolute requirement that CSTs by definition should meet
      the SMART criteria, this type of safety targets could actually not be considered CSTs
      but rather “common safety requirements” for the CSM framework.
   - It is difficult and maybe not desirable to base them on “current safety performances”
      (eg doubtful whether criteria for new and significantly modified systems should be
      based on current levels) as is stated in the Safety Directive.
   - Defining what the targets should be would require quite a lot of work, and is a project
      that would be most likely be difficult to finish within the first set of CSTs. The impact
      of implementing such criteria needs to be very thoroughly examined.




introduction to approaching CSTs for parts of the railway system_10                       Page: 8/9
The first steps in analysing the appropriateness and feasibility of using this approach would be
to define exactly what parts (including unambiguous definitions of these parts) and what
categories that should be subject to these CSTs. Following this an inventory of existing
criteria used on national levels and criteria used in recent projects should be performed to see
whether there are good examples to draw experience from. Using this as a basis it should be
possible to decide on how to progress on this issue, and also to decide on the time frame for
developing the CSTs for parts.




introduction to approaching CSTs for parts of the railway system_10                     Page: 9/9

								
To top