Graded Scale of Events UN-DG Based Approach

Process Safety Performance
An EPSC view on the use of lagging indicators in reporting
process safety performance
June 2010
Process Safety Performance
EPSC encourages the use of performance indicators in process
safety amongst member and non-member companies. The use
of appropriate metrics serving both internal company and
industry-wide needs will stimulate learning, identify opportunities
for improvements and facilitate communication on
Process Safety is distinct to personal safety. Traditional safety
performance indicators such as ‘Lost Time Injuries’ or ‘Total
Recordable Injuries’ do not give sufficient insight in process
safety performance. When over-emphasized these might even
distract attention away from managing process safety
Process Safety Performance Reporting
In 2008 in an EPSC position paper, the problems of learning from
shared incident data were brought to the attention of CEFIC.
Lagging indicator definitions varied, making direct comparisons
between shared company performance data difficult.
The Previous best approaches for performance reporting were
accelerated by investigations such as The Baker Panel report...
– Company-internal reporting systems for incidents
– External sharing and reporting of incidents
– Reporting of process safety incidents under RC (US, 1995)
– OECD “Guidance on Safety Performance Indicators”(2003/2008)
– EPSC report “Leading Indicators for Process Safety” (2004)
– European legal framework (Seveso, MARS-database)
…and clearly placed ‘process safety’ (again) on the agenda.
Previous Best Approaches For Indicators
•Academic discussions in safety journals
•Shared lessons from other companies
•Guidance from agencies and industry bodies
Responsible Care® and Lagging Indicators
The principles of Responsible Care® require continual effort in
learning and improvement, and industry organizations are
developing ways to communicate on performance and share
EPSC values efforts from organizations like CCPS, ACC and
CEFIC in establishing common ground in reporting on process
safety performance through the development of lagging
indicators on process safety.
EPSC considers the internal company reporting on process safety
incidents a crucial condition for achieving continuous
improvements. Work on lagging process safety indicators
supports this reporting.
Lagging process safety performance indicators
Developments in the USA:
• CCPS developed an approach for lagging and leading indicators
for process safety (Guidelines for Process Safety Metrics)
• The ACC adopted the approach developed by the CCPS.
• API used this as a starting point and developed an API-standard
(API- RP754). Further referred to as ‘UN-DG based approach’.
Developments in Europe:
• EPSC installed a working party for sharing information,
comparing company approaches and developing joint vision.
• The CEFIC discussed an own approach and requested EPSC
technical advice on this.
• No recommended standard in Europe (yet).
(see Annex for more detail)
• The ideas for lagging process safety performance indicators
developed by CEFIC and the approach developed in the US by
CCPS have a lot of similarity and use the concept of ‘Process
Safety Incidents’.
– Both define so-called ‘process safety incidents’ based on releases of
substances above a certain threshold or when certain consequences occur
due to contact with released substances (e.g. fire, explosion, lost-time
injures, etc.
– Both follow a tiered approach.
• On the definition of release thresholds there are crucial
differences. In the API standard the classification based on UNDG is used (further referred to as “UN-DG based approach”), in
the CEFIC proposals GHS is used (“GHS-based approach”).
EPSC Working Group – A Comparison of Methods
Qualitative and Quantitative results
• EPSC has member companies in favor of either a UN-DG based
or a GHS-based approach. The pros and cons were extensively
discussed and several attempts to align on one global metrics
approach were made. Such alignment was not found.
• Starting in 2009 it was decided to run a comparison study of the
level of incident reporting which is offered by two reporting
systems, that of GHS-based and UN-DG based.
• Three companies provided historical incident data as input to
GHS-based and UN-DG based reporting systems. This provided
a sample of 221 incidents that were classified using the two
reporting systems.
• The objective of the exercise was to allow EPSC members to
make informed choices regarding the potential to report using
API-754 and to determine whether more efforts to compare the
approaches has added value.
A Comparison of Methods
Qualitative and Quantitative results
• Differences in the systems for handling location of release or
activity causing release do not greatly influence the results.
• Different classification systems for chemicals are in broad
agreement on hazard level of substances.
• Duration of release does not greatly influence the result.
• GHS and API-754 are qualitatively similar.
• For similar data sets, GHS produces approx. 4 times as many
PSI reports as API-754 produces Tier 1 reports.
• GHS produces approx. 2 times as many PSIs as API-754 Tier 1
and Tier 2 combined.
• Quantitative difference: The CEFIC GHS system defines more
incidents as reportable.
Observations – Ease of use
• When the classifications are known (UNDG and GHS), or when the
substance is in the PSI Evaluation tool (CCPS for UNDG or EU-lists
for GHS) the thresholds are easily found.
• When not: classification according to UNDG requires some toxicity
data, flash point, boiling point and sometimes volatility. Although
already requiring specialist knowledge, this is still very simple
compared to having to establish the GHS classification when it is
not yet available.
• In practice missing GHS data were compensated by the use of
present classification with hazard symbols.
• Releases of mixtures, intermediates and products are often easier
to classify in the GHS based approach when the current
classification with hazard symbols is used rather than a GHS
classification. GHS classification for mixtures are often not (yet)
Observations - Incident reporting systems
• Companies will likely have to expand the scope of current
internal reporting to record enough data to use either system
• A tiered approach has benefits: different levels of reporting,
different levels of scrutinizing data.
• API-754 Tier 1 alone does not provide enough incident data to
run improvement programs and be able to demonstrate
improvement. Internal to an organisation you want to get as
many incidents reported as possible.
• Quality and accuracy of data collected has a strong influence on
the number of results returned – good reporting habits are
• The set of substances used in a workplace influences the
coverage of specific types of incidents.
• Not one system catches all hazardous events. Some incidents
are reported only in GHS, other incidents are reported only in
API-754 T1 and T2.
• From a technical perspective either GHS or API-754 T1+T2 will
• Tier 1 alone of API-754 returns too few reports and should not
be used unless to complement a separate internal reporting
• With an extended dataset it is possible to extract reports
according to both approaches at the cost of additional effort.
• Distinction between the two systems is primarily in quantity of
reported data.
DRAFT Results sub-group PSI/March 2010
The Future
• EPSC will work together with volunteering companies develop a
tool to support using either one of the approaches for classifying
process safety incidents.
• Real world use of both systems will reveal areas of potential
improvement and may reveal opportunities to further harmonise
the systems for unified global reporting.
• Whilst it is crucial that external reporting systems for the sake of
transparency are carefully defined at their margins it is evident
that increasing the definition for internal reporting results in
decreasing returns – it would be better to focus on the promotion
of such a system across the business and encourage the full
and accurate workforce reporting of process safety incidents.
1. Characteristics of UN-DG based approach
2. Characteristics of GHS based approach
N.B. Characteristics are presented as needed for the purpose of comparing the
approaches and are as understood by the EPSC working party. They do not
contain the full scope of definitions and guidance on either one of the
approaches. For complete information on the approaches we refer to CCPS,
API RP754 and CEFIC publications.
Characteristics UN-DG Based Approach
• Tier 1 represents lost time injury, fatality, event
causing significant cost or with significant loss of
hazardous material.
• Tier 2 is a less significant but potentially dangerous
event. It uses the same definitions but with lower
• Tiers 3 and 4 represent deviations from ideal
procedure in equipment and human activities.
Graded Scale of Events
UN-DG Based Approach
Tier 1
LOPC Events of
Greater Consequence
ic a
Small loss of primary
containment event
in g
Large loss of primary
containment (LOPC) event
Tier 2
LOPC Events of
Lesser Consequence
In d
Challenges to Safety Systems
Tier 3
Challenges to the
safety system
Operating discipline &
management system
Tier 4
Operating Discipline & Management System
Performance Indicators
Thresholds for Reporting Event as Tier 1
UN-DG Based Approach
Threshold Release
UN Dangerous Goods Material Hazard Classification
Threshold Quantity
Threshold Quantity
TIH Hazard Zone A Materials
TIH Hazard Zone B Materials
5 kg
25 kg
2.5 kg
12.5 kg
TIH Hazard Zone C Materials
100 kg
50 kg
TIH Hazard Zone D Materials
200 kg
100 kg
Flammable Gases
500 kg
250 kg
1000 kg
500 kg
7 bbls
2000 kg
3.5 bbls
1000 kg
Liquids with Boiling Point < 35 °C (95 °F) and Flash Point <
23 °C (73 °F)
Other Packing Group I Materials
Liquids with Boiling Point > 35 oC (95 °F) and Flash Point <
23 oC (73°F)
Other Packing Group II Materials
Liquids with Flash Point ≥ 23 °C (73 °F) and < 60 °C (140
Liquids with Flash Point > 60 °C (140 °F) released at a
temperature at or above Flash Point
strong acids/bases
Other Packing Group III Materials
14 bbls
7 bbls
Inhalation toxicity and volatility
UN-DG based approach
In the UN-DG based approach inhalation toxicity and volatility is taken into
account in order to represent accurately the inherent human hazards of a toxic
Characteristics of GHS-Based Approach
• A binary system, divides events into Process Safety
Incidents (PSI) or not.
• Intended to be used alongside other reporting
systems for internal use and to complement existing
reporting systems.
• Based on Globally Harmonized System for
Classification and Labeling of Chemicals.
Schematic overview GHS-based Approach
Was a chemical substance or a chemical
process directly involved in the damage
Does not meet the
criteria for a Process
Safety incident
Did the incident occur in production,
distribution, storage, utilities, pilot plants
laboratories within the site boundaries of
a companies facility?
GHS classified
Health Hazards
Acute Toxic
all other
STOT single exposure
Injury resulting in a Lost Time Accident
or a Fatality or Hospitalisation (>24h) of
any people on or off site?
> € 25.000 of
direct costs to
the company
Physical Hazards
All categories
Environmental Hazards
All categories
Thresholds (8h rule applies)
Reportable Process
Safety Incident
5 kg
100 kg
Carcinogenic, Reproductive,
Not GHS classified
Was there a release of material or
energy (e.g. fire, explosion, implosion)
from a chemical process unit that
resulted in …
2000 kg
Thresholds for Loss of Containment becoming a PSI
GHS-Based Approach
GHS classified
Health Hazards
Acute Toxic
STOT single exposure
Physical Hazards
All categories
Environmental Hazards
All categories
Thresholds (8h rule applies)
5 kg
100 kg
Not GHS classified
all other
Carcinogenic, Reproductive,
2000 kg

similar documents