Presented at
SignalComm
Europe 2000
16th - 17th May 2000
Birmingham, UK
UNDERTAKING A SAFETY CASE
IN A RAIL ENVIRONMENT
Odd
Nordland
SINTEF Telecom and Informatics
Systems Engineering and
Telematics
NO-7465 Trondheim
E-mail: Odd.Nordland@informatics.sintef.no
URL: www.informatics.sintef.no/~nordland
Abstract: The CENELEC standards for railway applications
require "safety cases" as a prerequisite for approval of such applications.
The paper explains the concept of a safety case and what it should contain,
discusses the various kinds of safety case that the standards identify
and illustrates the cost implications and benefits that can be achieved.
Keywords: CENELEC, Railway Applications, Safety
Case, Standards.
1. Introduction
Automatic train protection (ATP) and interlocking systems have become more
complex and rely heavily on microprocessors and software. Thus, the assessment
and certification of such new systems has changed and become correspondingly
more complex. In addition, European integration leads to a need for common
principles and procedures, so that assessments and certifications performed
in one country can be adopted by other countries.
One step in this direction is the adoption of European standards for
railway applications, notably EN 50126, EN 50128 and EN 501291.
They describe processes to be followed in order to be able to assure the
safety of a railway application. However, whilst they describe reasonably
completely what to do, they do not go into great detail on how to do it.
Now of the above-mentioned standards, EN 50126 is the top-level document
that covers the overall process for the total railway system. It defines
Safety Integrity Levels and sets the frame for the more detailed activities
that are described in EN 50128 and EN 50129. EN 50129 is the standard that
defines the activities for developers and manufacturers, but also describes
the requirements that a third party assessor shall verify. EN 50128 is
the software specific "subset" of EN 50129.
EN 50129 requires that a safety case shall be submitted by a manufacturer
and assessed by an independent third party before the safety authorities
should approve commissioning the system. The term "Safety Case" is perhaps
reasonably straight forward for people with English as their mother tongue,
but experience shows a large degree of confusion when non English speaking
Europeans use the expression! So the term requires some elaboration.
2. Safety Cases
The word "case" is used in a variety of contexts. We have special cases,
briefcases, suitcases, bookcases, court cases and - safety cases. The latter
is derived from the concept of a court case: the prosecutor and defendant
both "present their cases" to the court in order to convince the judge
that their proposition is right.
Now for our purposes, the proposition is that a new (or modified) system
is safe enough to use, and somebody has to present the case so that the
"judge" - the safety authority - can reach a decision. As in legal proceedings,
the "judge" will refer to an expert assessment by an independent party
before relying on his own personal impression. (The word "assessor" did,
in fact, originally mean "co-judge"!)
It is important to keep this analogy in mind. The safety case is a line
of argumentation, not just a collection of facts. It aims at convincing
a licensing authority, who does not necessarily have a deep knowledge of
the specific technological questions involved, that a given product or
system is safe enough to be taken into use. Or was safe enough: the safety
case could well be used as evidence in genuine legal proceedings years
after the system was authorised and commissioned.
One of the most common problems with safety cases is that they are too
concise. It simply isn’t sufficient to just refer to dozens of documents
and state that all the information is there, leaving it up to the assessor
to read through them all and extract the necessary facts. Imagine a court
case where the defendant claims that all the evidence of his client’s innocence
is in two meters of shelving in his bookcase, so the judge just has to
read through the documentation to convince himself. The judge would probably
advise the client to get a new lawyer!
The safety case must in itself contain enough information to give a
clear impression of the system’s safety properties and indicate where the
gory details can be found if this is desired. So the safety case sections
that are defined in EN 50129 must contain descriptive text rather than
just a more or less complete list of references.
2.1 The Contents of a Safety Case
EN 50129 gives a fairly extensive description of how safety cases should
be structured, but the information is distributed through the standard
and its annexes. Experience shows that there is a need for a briefer guideline.
In the following, the sections are briefly described. It should be noted
that it is perfectly legitimate to let each section be a free-standing
document, provided it has the right contents.
2.1.1 Definition of System
The first section in the safety case is the Definition of System. It shall
give a complete and detailed description of the system for which the safety
case is being presented. EN 50129 states:
"This shall
precisely define or reference the system/subsystem/equipment to which the
Safety Case refers, including version numbers and modification status of
all requirements, design and application documentation."
This means that the definition of system shall contain:
-
A description of what the system2 is,
its functionality and purpose, with references to the requirements’ specification
and other descriptive documents.
-
The product structure. This is more than just a parts list; it’s a document
that identifies the components of the system and the way they are related
to each other and the overall system.
-
Descriptions of all interfaces, both external and internal, with references
to the corresponding documentation. The interfaces should be traceable
to the product structure.
-
Issue, revision and date of all applicable documentation and/or versions
of components or modules.
It should be noted that this section is a definition of the system, not
just a description. Its purpose is to identify exactly which version of
the system is meant. Any modification of a component or module will necessitate
a new - or at least updated - safety case. The same applies if the documents
get updated.
The author recommends that a separate "List of Applicable Documents"
("LAD") should be used. Such a document list must identify all the
applicable documents, not just the directly referenced ones, by title,
document number etc. and the exact versions to be used. Similarly, a "Configuration
Item List" ("CIL") should be used to identify the exact versions of components
and modules.
It is then sufficient to explicitly identify the valid issue and revision
of the LAD and the CIL in the safety case - for all other documents and
components reference can be made to the LAD and/or CIL for the necessary
version data.
2.1.2 Quality Management Report
This section is a report that describes what has been done to ensure that
the system has the required quality throughout the entire life cycle. This
involves:
-
A description of the quality requirements with reference to the corresponding
"source" documents. These are more often than not generic, company internal
documents, but there can also be laws or regulations that define quality
requirements. Such generic company documents, laws and regulations must
be identified.
-
A description of the quality management system with references to the corresponding
plans and procedures. In other words, a description of what one intended
to do in order to ensure the necessary quality. This must also contain
a description of the project’s organisation and identify by name
the people in the various positions and describe their responsibilities
and qualifications!
-
A description of what actually was done, with references to e.g. audit
reports, minutes of meetings and any other documents concerning the performed
activities. In addition, any deviations from the plans and procedures shall
be described and justified. With deviation we mean any activities that
should have been performed according to the plans, but which either were
not performed, or for which no documentation or other evidence can be provided.
EN 50129, paragraph 5.2, contains examples of topics (based on ISO 9000)
that should be addressed, but the list should not be regarded as being
exhaustive. They are examples that need not necessarily all be relevant.
But in such cases it is wiser to state explicitly that a particular topic
is not applicable and give some kind of explanation rather than just not
mentioning it. Otherwise, there’s a fair chance somebody (the assessor!)
might think that it’s been forgotten. On the other hand, there may be other
aspects not mentioned in the list that are significant, so care should
be taken when using the list in EN 50129 as a check list!
2.1.3 Safety Management Report
This is the corresponding report for safety management. As with the quality
management report, the safety management report involves:
-
A brief summary of the safety requirements with reference to the safety
requirements’ specification. The safety requirements’ specification may
be, and often is, a subset of the requirements’ specification rather than
a separate document. In this case, the relevant parts of the requirements’
specification shall be identified. In addition, there will be standards,
laws and regulations defining general safety requirements. These must also
be identified.
-
A description of the safety management system with references to the corresponding
plans and procedures. In other words, a description of what one intended
to do in order to ensure safety.
-
A description of what actually was done, with references to e.g. the hazard
log, safety audit reports, test reports, analyses and any other documents
containing evidence of the activities that were performed. In particular,
the way the hazard log is handled must be described. In addition, any deviations
from the plans and procedures shall be described and justified. With deviation
we mean any activities that should have been performed according to the
plans, but which either were not performed, or for which no documentation
or other evidence can be provided.
EN 50129, paragraph 5.3, contains examples of topics that shall be addressed:
5.3.1 Introduction
"In all
cases the hazard analysis and risk assessment processes ... are necessary,
in order to identify the required level of Safety Integrity ... This includes
those cases where the analysis and assessment reveal that a Safety Integrity
Level of zero may be assigned..."
5.3.2 Safety
life cycle
5.3.3 Safety
organisation
5.3.4 Safety
Plan
"A Safety
Plan shall be drawn up at the start of the life-cycle ... The Safety Plan
should include a Safety Case Plan, which identifies the
intended structure and principal components of the final Safety Case."
5.3.5
Hazard Log
"A
Hazard Log ... shall include a list of identified hazards, together with
associated risk classification and risk control information for each hazard."
5.3.6
Safety Requirements Specification
5.3.7
System/sub-system/equipment design
5.3.8
Safety Reviews
"Safety
reviews ... shall be specified in the Safety Plan, and their results shall
be fully documented."
5.3.9
Safety Verification and Validation
5.3.10
Safety Justification
"The
evidence that the system/sub-system/equipment meets the defined conditions
for safety acceptance shall be presented..."
5.3.11
System/subsystem/equipment Handover
5.3.12
Operation and Maintenance
5.3.13
Decommissioning and Disposal
Note that EN 50129 makes the use of the safety management process as sketched
above mandatory for safety integrity levels 1 to 4! In particular, this
means that a hazard log shall be maintained throughout the entire lifecycle
of such systems. EN 50129 defines the hazard log as:
"The document
in which all safety management activities, hazards identified, decisions
made and solutions adopted, are recorded or referenced"
So this is not just a list of hazardous events, it must identify their
probable consequences and frequencies, it must include a classification
(e.g. negligible, tolerable, intolerable etc.) and provide information
about how the hazard was tackled (e.g. technical modifications, administrative
procedures or other suitable measures). EN 50126, paragraph 6.3.3.3 defines
the hazard log in detail.
Obviously, there should be no unresolved, intolerable hazards left in
the log when a system is in operation.
2.1.4 Technical Safety Report
This is the section where the technical safety characteristics of the system
are described. It shall describe the underlying philosophy for achieving
safety and identify which safety standards have been applied. The safety
relevant properties of the system shall be presented and reference made
to the corresponding evidence, i.e. test and analysis results, verification
and validation reports, certificates and so on.
EN 50129 prescribes a structure for the technical safety report:
-
Introduction
This section
shall provide an overview ... of the design, including a summary of the
technical safety principles ... This section shall also indicate the standards
... used as a basis ...
-
Assurance of
correct functional operation
This section
shall contain all the evidence necessary to demonstrate correct operation
... under ... normal conditions ...
-
Effects of faults
This section
shall demonstrate that the system ... continues to meet its ... safety
requirements ... in the event of random hardware faults. In addition ...
which technical measures have been taken to reduce the ... risk ...
-
Operation with
external influences
This section
shall demonstrate that when subjected to ... external influences ... the
system ... continues to fulfil its ... requirements ...
-
Safety related
application conditions
This section
shall specify ... the ... constraints which shall be observed ...
-
Safety qualification
tests
This section
shall ... demonstrate successful completion ... of the Safety Qualification
Tests. ...
EN 50129, paragraph 5.4 and Annex B (which is normative) identify the topics
that shall be addressed, summing up to a total of 123 headings and sub-headings
for the technical safety report! Some of the topics may not be relevant
for a particular kind of safety case, but as in the quality management
report, it is wiser to state this explicitly: EN 50129, Annex B, requires
that the Technical Safety Report "shall be arranged" under the headings
given in the standard! On the other hand, additional topics may be necessary,
so the list is not necessarily exhaustive.
2.1.5 Related Safety Cases
If the system’s safety relies on the use of safe parts or components, the
corresponding safety cases shall be identified here. In such cases, any
restrictions or application conditions mentioned in those safety cases
shall be recapitulated here.
Note that this also applies when there is a previous version of the
system for which a safety case already exists. In this way, producing a
safety case for an upgraded system can be considerably simplified. Since
the previous version’s safety case contains all the relevant information
for that version, only the changes from the previous to the new version
need to be described. In addition, evidence must be provided that the modifications
have not adversely affected the safety properties of the unmodified rest
of the system.
"Related safety cases" can also refer to certificates for parts or components,
since such certificates will themselves be based on documentary evidence
of the relevant safety properties.
2.1.6 The Conclusion
This is more than just the statement that the product is sufficiently safe.
EN 50129 states:
"This shall
summarise the evidence presented in the previous parts of the Safety Case,
and argue that the relevant system/subsystem/equipment is adequately safe,
subject to compliance with the specified application conditions."
This section is the plea that recapitulates the evidence presented in the
previous sections and the argumentation for the system’s safety. This is
where the justification shall be given that the quality and safety management
activities were adequate, that the technical properties satisfy the safety
requirements and that the conditions imposed by the related safety cases
have been adequately taken into account. Any restrictions, limitations
or assumptions that were made shall be stated here as "application conditions".
2.3 Different kinds of safety cases
EN 50129 identifies three different categories for Safety Cases:
Generic
product Safety Case (independent of application)
A
generic product can be re-used for different independent applications.
Generic
application Safety Case (for a class of application)
A
generic application can be re-used for a class/type of application with
common functions.
Specific
application Safety Case (for a specific application)
A
specific application is used for only one particular installation.
The underlying idea is that a Generic product Safety Case (GPSC) will present
the safety case for a product, regardless of how it is used, so that it
can be deployed in a variety of different safety related applications.
The Generic application Safety Case (GASC) will present the safety case
for an application without specifying the actual products to be used as
components. It will simply refer to the generic product properties that
the components should have.
Unfortunately, this makes the boundary between GPSC and GASC rather
fuzzy, because a complex system that can use "generic" components (and
therefore has a GASC) can itself be deployed as a component in a greater,
more complex system. Then the "generic application" becomes a "generic
product" within the more complex application.
For example, a controller for wayside objects could be a generic application
that can control a variety of different objects without specifying exactly
which "brands" must be used. Using a particular kind of controller in an
interlocking system would make it a generic product within that system.
However, this should not influence the evidence for the controller’s safety
properties.
Finally, the Specific application Safety Case (SASC) presents the safety
properties of a particular combination of products in a given application.
It will, of course, draw on the underlying GPSCs and GASC, but in particular,
the details of planning and operation will be relevant here. In fact, EN
50129 prescribes "separate
Safety approval ... for the application design of the system and for its
physical implementation... ", so there must be
two SASCs:
-
A "SASC - Design" that presents the safety evidence for the theoretical
design of the specific application.
-
A "SASC - Physical implementation" that presents the safety evidence for
"e.g.
manufacture, installation, test, and facilities for operation and maintenance".
EN 50129 requires the same structure for all the above kinds of safety
case, although the contents of the various sections will depend on the
kind of safety case that is involved.
3. Cost Implications
3.1 Getting started
The amount of documentary evidence that the CENELEC standard requires is
formidable, and producing a safety case looks like an intimidating task.
But it is much simpler - and cheaper - that it would appear to be at first
glance. Remember that a safety case is really a line of argumentation,
and most of the necessary information will already exist.
Let’s look at the various sections of the safety case. The Definition
of System contains the kind of information that you have to provide to
your engineers so that they will know what to build. A brief description
of the concept has probably been produced for presentation to management,
in order to get the budget authorised. Specifications, interface descriptions
and the like are part of normal engineering work, and producing and maintaining
the list of applicable documents is a normal configuration management task.
So the task of producing section one of the safety case boils down to a
cut-and-paste exercise with a bit of editorial polishing afterwards.
For the quality management report the situation is even easier. The
Quality Manager will normally be expected to report to e.g. the board of
directors at least once a year, explaining what he has done, what he hasn’t
done and why. He will be well advised to identify the documents that prove
that he really did his job and that the results were acceptable. So the
Quality Management Report already exists!
There may still be a bit of editorial polishing necessary, because
the Quality Manager might have reported for all projects in one go, so
the non applicable sections must of course be removed.
For the Safety Management Report we have the same situation: the Safety
Manager must report to somebody, so the report already exists. It, too,
may need some editorial polishing.
The Technical Safety Report is a bit more work. However, it contains
basically the information that a manufacturer would supply to his customers
in e.g. data sheets etc. Identifying where the proof can be found that
the product really does have the technical properties that are claimed
is something that the legal department should be very interested in. It’s
what will be needed to ward off any liability claims that might arise.
So if the data hasn’t already been compiled, it’s high time it was!
As can be seen from the description of the technical safety report,
this could be more than a cut-and-paste exercise. But it doesn’t require
producing documentation that wouldn’t be produced otherwise.
The section on related safety cases is basically part of the design
documentation. If "safe" components are used, the corresponding documentation
can be identified by the configuration manager. And summarising the conditions
under which the components remain safe is a necessary input for the design
engineers.
Finally, the conclusion is something "new". But we have seen that most
of the safety case already exists, so that producing one is mostly an editorial
exercise with a very limited amount of new text!
3.2 Re-use
Much of the work invested in generating one safety case can be re-used
in another. Certainly the descriptions of what was intended for quality
and safety management will not change substantially, so only the corresponding
evidence of what was actually done needs to be updated. But the potential
for re-use can be much greater than that.
Complex technical systems are made up of less complex components, which
in turn will consist of parts which are even less complex. This inevitably
leads to a hierarchical structure of products. When safety cases have been
produced for the bits and pieces in the bottom row of such a hierarchy,
they can be used over and over again, in all higher level combinations.
The same will apply to the products on the next hierarchical level, all
the way up to the final products. In other words, for each product the
safety case has to be produced exactly once. And the further up you get,
the less you have to prove, because most of the proof is already contained
in the lower level safety cases.
Similarly, if a product, for which a safety case already exists, is
modified, the new safety case can be based on the already existing one.
Only the changes and their effects need to be argued for. This is evidently
considerably less work than producing a whole safety case every time.
4. The benefits of the CENELEC approach
4.1 Cross acceptability
The CENELEC standards foresee a very clear and uniform way of presenting
the evidence that a given product or application is safe to use. So at
least for the Generic product Safety Case and the Generic application Safety
Case, one safety authority will be able to accept a safety case that has
been accepted by another authority. The only additional requirement could
be to assure that the safety requirements and application conditions are
still applicable or compatible.
For specific applications it is reasonable to expect that only local
authorities will be responsible for granting approval, so the question
of cross acceptance should not arise.
4.2 Efficiency of the approval process
The approval process will require at least the presentation of a safety
case and an assessment report. For the assessor, a well-structured and
complete safety case will greatly simplify his task, shortening the time
he needs to produce his report.
For the authorities matters will also be simplified. When the safety
case is clear and easy to understand, all they need to look for is an assessment
report confirming that what they have read in the safety case is, in fact,
true. It is unlikely that they will require further documentation in order
to become convinced of the product’s safety. And if the hierarchical structure
described above is used, the safety cases will become simpler and simpler
as time goes by.
5. Conclusion
A safety case is a line of argumentation and not just a collection of documents.
The CENELEC standards prescribe a clear structure for the contents that
can draw on reports and other documentation that are produced as a part
of normal work.
The CENELEC approach enforces a transparent and effective way of producing
and documenting safety related products. In return, it leads to efficient
approval processes that can often be quite widely applicable.
In the course of time it can be expected that safety cases will become
simpler and simpler because they will be based on already existing and
approved safety cases. This will facilitate and accelerate both the production
and the authorisation processes.
1 At the time of writing, only
EN 50126 has been adopted as a standard. The others are still pre-standards.
For simplicity, they are also referenced as standards (EN instead of prEN
resp. ENV).
[back to text]
2 For simplicity, the term "system"
will be used in a very generic sense that includes subsystem, equipment,
product etc.
[back to text]