University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
UWA
|
366
useful and safe operational life in order to define any actions required on
plant, based on the importance it presents to the operation and planned
enhancement of the system.
The 'Plan' is updated at least annually and addresses asset
replacements, refurbishments and modifications, and relevant maintenance
documentation (policies, servicing, instructions and contingency plans It also
shows asset long-term strategy for asset actions.
The objectives of the 'Asset Management Plan' are to ensure that:
• Plant performance will meet current customer requirements;
• Plant capacity and condition will enable customer requirements to be met
in the future;
• Faults can be repaired without jeopardising system security;
• Supply is restored quickly;
• World class industry practice levels are achieved for 'whole of life1 costs,
including long-term uprate and refurbishment programmes;
• The long-term business plan reflects necessary capital expenditures to
replace assets as their life is expended;
• The safety of employees, customers and the public is protected; and
• Environmental impact is acceptable.
3.1 Asset Management Plan Development
The principal steps in preparation of an 'Asset Management Plan' are:
a) The Assets Branch develops, on an annual basis, a 'first cut' list of plant
that is considered a possible candidate for update of its maintenance
documentation, or could be in need of modification, refurbishment and
replacement. The plant is identified by analysing statistics on plant age,
|
UWA
|
367
reliability, failure rates, and known or assessed condition. The outcomes of
these proceedings are contained in the part 'Assets Under Review'. The
relevant reports with their frequency of producing and way of use are given
in more details later.
b) Information is regularly gathered from other electricity utilities, industry
bodies and forums, and used to assist with constant improvement of criteria
used to arrive at the refurbish/repair/replace decisions.
c) The initial list is reviewed annually before capital budget is set in
conjunction with Planning, System Operations and Projects. It is necessary
to determine impact of new developments, planned upgrades, plant position
and criticality, plant current condition, and likely cost and benefits of
refurbishment/replacement programme.
d) The Assets Branch then finalises the list and includes it in the 'Asset
Management Plan', indicating expected actions and time frames. The 'Plan'
also specifies what additional work is necessary in the future where plant
condition, reliability and cost are not currently well defined or known.
e) Assets Branch prepares a detailed technical and financial business case
and a capital project approval submission where required in a standard
format for each individual plant item. The replacement projects are dealt
with in time for required action time frame.
f) Regular development and planning reports issued by the Planning Branch
are analysed and used to identify possible deletions or updates (in scope of
work and timing) for items listed in the 'Asset Management Plan'.
g) Plant maintenance and fault reports statistics from the relevant plant
databases should be produced and analysed on a regular basis to identify
possible additions to the 'Asset Management Plan'.
h) There should be an annual review of the 'Asset Management Plan' against
network development plans and of the status of work in progress to assist in
updating of the 'Plan'.
|
UWA
|
368
3.2 Asset Management Plan Structure
The 'Asset Management Plan' consists of the following main parts:
Part 1 - Assets Under Review- Plant Directory - This is a complete list of
the primary assets, consisting of primary equipment and transmission lines.
It represents the total population and is reviewed annually. This is a high
level review of plant performance, the outcome of which is a list of plant
below with possibly suspect performance that warrants closer scrutiny.
Plant Under Review - This is a list of plant determined from the
initial review and than subjected to more detailed analysis. This analysis
may lead to one of three outcomes:
1. Performance is acceptable and no further action is needed.
2. Further information is needed about the plant in question to reach a
conclusion and recommendation. This work is called 'Asset Investigations'.
3. The plant is assessed to need future action and is included in the plan of
'Asset Future Projects'. The projects are grouped into time frames of 5
years, 10 years, and over 10 years.
Part 2 - Asset Investigations- Before or during the assessment of items on
the 'Plant Under Review' list it may be necessary to obtain additional
information about the plant. This can include a number of inspections, tests
and measurements, and may include various trials and joint research
projects with other authorities, companies, universities and manufacturers.
|
UWA
|
370
The final outcome of these investigations can fall in two main groups:
1. No additional work required on the plant, but:
(a) units are to be reallocated for use at less important or demanding
locations in the network;
(b) it will be necessary to vary maintenance policy, maintenance servicing
or maintenance instructions (eg different servicing frequencies, changed
scope of work during service, etc);
(c) plant failures will be catered for by generic or specially developed
contingency plans for a specific period of time,
2. The plant is to remain in service and is added to the list of 'Asset Future
Projects' for:
(a) refurbishment or modification on site or in a workshop;
(b) replacement over a period of time, possibly used for spare parts.
Part 3 - Update of Documentation-As a result of 'Asset Investigations' or
'Asset Business Case Analysis' it is possible that an outcome would not
point to the need to perform a refurbishment, modification or replacement
of the plant. The identified problems could sometimes be better and more
efficiently addressed by altering the plant manuals for maintenance policy,
maintenance servicing or maintenance instructions, using an already
existing generic contingency plan, or by initiating a specially developed
contingency plan.
Part 4 - Asset Future Projects- The 'Asset Future Projects' represents the
potential work necessary to address plant refurbishment or replacement
issues. This may include projects for the installation of new technology
condition monitoring equipment onto existing plant.
• 0-5 years - Plant to be considered for action within the next five
years, with the recommended type of action and timing;
• 6-10 years - Plant to be considered for action within the following
five years, with the recommended type of action and timing;
|
UWA
|
371
• over 10 years - Plant to be considered for action after 10 years, with
the recommended type of action and timing.
Projects from this potential work program are subjected to a detailed
business case analysis as they mature to determine if the project should
proceed. If the outcome of the business case is positive the project will be
added to the list of 'Asset Planned Projects', and approval for capital or
operating expenditure for the recommended action and timing will be then
sought.
Part 5 - Asset Planned Projects- This is a list of the details of the
recommended projects arising from the detailed business case analysis of
projects taken from the list of 'Asset Future Projects'.
Part 6 - Asset Current Projects- This is a list of details of all current
projects, either approved or ongoing, in the areas of refurbishment and
replacement.
Part 7 -Asset Completed Projects-Lists of details of all projects completed
since the last 'Asset Management Plan' was issued.
3.3 Establishing Asset Review File
The initial work procedure to be undertaken for setting up each item is:
• Create a new file and obtain a transmission asset management (TAM)
file number,
• Run the latest plant current locations report from an appropriate
database,
• Prepare a population/age/dissipation report,
• Run the latest failure report from an appropriate database,
• Run the latest corrective maintenance (fault history) report,
• Obtain preventive maintenance criteria details and maintenance costs for
current assets,
|
UWA
|
372
• Propose and assess possible modification and replacement options, and
obtain information for their basic costing,
• Seek necessary input from Plant Section and Asset Services about the
above preliminary options,
• Indicate if further investigation will be required before the work on a
business case could proceed. This might include consulting other plant
specialists, manufacturers, authorities and industry bodies as required, or
conduct a series of site or indoor laboratory tests,
• Send the list of plant locations to the Planning Section and obtain their
comments from future development plans and impact on plant,
• When assessments involve instrument transformers request comments
from Protection Section with reference to future protection upgrade projects
that would affect instrument transformer projects,
• Send the list of locations for affected plant to the System Operations to
obtain their assessment of locations criticality, and possible impacts on
system security and reliability of supply,
• Make preliminary findings if this plant type is a candidate for eventual
replacement, refurbishment and modification, or update of maintenance
manuals and contingency plans, and include in list of 'Future Projects',
• If further action is recommended, obtain detailed cost estimates from the
Projects Group,
• A detailed business case and capital or operating project approval will be
prepared for each individual plant item of the replacement project,
according to its required timing.
4.0 ASSET BUSINESS CASE ANALYSIS
There are number of issues that need to be considered before a proposal is
presented for any action required of a particular item of plant. The individual
|
UWA
|
373
components are grouped by similarity of the area they address, and the
assessed ratings of a group are then used for an overall asset condition
assessment in the business case presentation.
Details of each group and its components are including their priority,
associated weighting factors, risk evaluation, and cost implication (life cycle
cost assessment) are given below.
Group 1 Age
The age of the units is given a low-weighting factor. The risk factor is
calculated from plant current age versus the total expected life of plant.
Group 2 Condition
The condition of plant is a significant factor in determining plant
performance, as it leads to failures and makes its future performance
uncertain. It is given a high-weighting factor, and is calculated from
frequency of failures and faults.
The frequency of failure risk factor is based on the number of failures per
unit per year, and the frequency of fault risk factor is based on the amount
of additional maintenance repair work required in excess of preventive
maintenance during regular maintenance intervals.
Group 3 Environmental/Regulatory Requirements
Plant is normally well designed to avoid any significant impact on the
environment, so the weighting factor is rated as low.
There might be special circumstances where a plan for an immediate action
could be considered, eg new legislation, certain areas proclaimed as
sensitive areas (eg underground water catchment), etc.
Group 4 Spare Parts and Services Availability
In assessing this group the following issues should be addressed:
a) Spare parts availability,
b) Competency of employees to perform work,
|
UWA
|
374
c) Availability and cost of external servicing.
A number of manufacturers have ceased to exist or have abandoned making
the particular model or spare components. In the case of a major fault
occurring, spare parts could be very difficult to obtain, which would result
in long waiting periods or very high costs to manufacture the parts on an
individual basis.
The skills to repair or completely refurbish the affected units within or
outside the company should be investigated and assessed. A combination of
factors listed below will be useful in the above assessment:
f) Difficulty in obtaining spare parts,
g) Remote locations of the units in question,
h) Difficulty in arranging access (eg feeders supplying mining sites),
i) Low success rate on a previously trialed refurbishment or
modification,
j) High cost of the rework required.
There might be special circumstances where a plan for an immediate action
should be considered where the risk factor 5 is assigned, ie all the above
factors are assessed as poor.
Group 5 Safety
In assessing this group the following issues should be addressed:
a) Safety of employees,
b) Safety of public,
c) Safety of adjacent equipment.
|
UWA
|
375
The units may pose a risk for the crews that perform work on adjacent
circuits, as the unit may operate under fault conditions with insufficient
security to guarantee its safe operation. Equipment maloperation may
present a direct risk to the public. That is a direct risk in the form of flying
debris or exposed live conductors. It may not always be possible on
occasions to mobilise a crew to site in sufficient time to quickly restore
customer supply, leading to hazards occurring in the general public domain.
There is also a real risk associated with suspected units having to operate
under fault conditions, as any adjacent equipment could be damaged by
flying debris, further exacerbating the consequences of failure, and could
cause a further release of any contaminated particles into the atmosphere.
There might be special circumstances where a plan for an immediate action
should be considered where the risk factor 5 is assigned, ie all the above
factors are assessed as poor.
Group 6 Obsolete Design Standards
Sometimes units are considered to be of obsolete design, with inherent
design imperfections, poor tolerances, etc, causing failures (eg jamming of
operating mechanisms), but generally perform their duties as required.
This may sometimes have influence on other groups, spare parts
availability, expensive maintenance, special skills required, safety problems
and operational restrictions, and must be then properly addressed.
Group 7 Impact on Customers and Company
In assessing this group the following issues should be addressed:
a) Impact on quality of supply,
b) Impact on customer,
c) Impact on company.
Any loss of supply from affected circuits has consequences on quality of
supply. It could cause complete loss of supply for a protracted period,
|
UWA
|
376
depending on the ability of the system to temporarily feed customers from
alternate sources. In a case of sensitive load areas there could be a
significant production time lost in mine site operations, disruption of supply
to general public, etc.
Such a loss of supply could also have a significant effect on the company's
business in a form of:
Loss of supply over a significant period of time would mean
significant loss of revenue,
Replacement and repair of equipment under emergency situations
would be very costly, particularly if other adjacent plant has been damaged,
Relationships with customers would suffer, an important
consideration in the environment of open access and increased competition.
There might be special circumstances where a plan for an immediate action
should be considered where the risk factor 5 is assigned, ie all the above
factors are assessed as poor.
Group 8 Costs
The maintenance costs can contribute significantly to the company's
operating expenditure, and they have to be assessed for the life of
equipment.
These costs should then be compared against the costs to repair, refurbish,
modify or replace the plant under consideration, using net present value
matrix (life cycle cost evaluation) to find the most economical solution. In
assessing this group the following issues should be addressed:
a) Maintenance costs,
b) Repair costs,
c) Modification costs,
d) Refurbishment costs,
|
UWA
|
377
e) Replacement costs.
Group 9 Ability to Manage the Replacement Process
It is important to realise the importance for the company to properly
manage its plant in the long run by initiating necessary action for equipment
that is assessed under risk even in the long term rather than to wait for
failures.
That might sometimes require plant replacements before they reach the end
of their individual life. This is undertaken by:
A suitable equipment long term condition and risk assessment,
The selection and review of all possible courses of action to deal with
the risks, and assessment of how each type of action can limit the identified
risks,
Identification of possible savings in long term operating costs
(tangible and intangible) by arranging suitable plant replacements, and how
will it lead to improvement of company's strategic result area indicators,
Implementing suitable interim arrangements to ensure secure supply
and the safety of staff,
A business long term financial approach to plan timely
implementation of repair, refurbishment, modification or replacement
programmes.
5.0 CONCLUSIONS
The papers 'Asset Management' and 'Infrastructure Replacement' have
outlined the development and implementation of a comprehensive asset
management model to support an electricity utility business in the
deregulated electricity market. The main purpose and characteristics of
main components of an asset management model have been described. We
|
UWA
|
379
14* ASSET M A N A G E M E NT PROCESS FOR TRANSMISSION
ASSETS
1. INTRODUCTION
This paper gives a description of an asset management process developed
and implemented by W PC to upkeep its transmission assets and to manage
the relevant documentation.
There are number of different maintenance, planning and
management systems that other businesses have in use, but this the first one
to connect all activities and groups involved in management of transmission
assets in single interconnected and mutually affected system. Assets
management process for transmission assets is defined through a number of
internal documents. Some of them give a high level overview of the asset
management process while others give more details for the relevant
particular areas.
The paper also provides some indication on how to collect and
prepare data, necessary reports to run, how to analyse their results, and how
to act on the relevant outcomes and their recommendations.
2. ASSET MANAGEMENT MODEL
2.1 Model Overview
An asset management model ensures that:
a) all assets receive adequate on-going care,
b) performance and condition of assets in service are continuously
monitored and reviewed,
c) plans are prepared for asset short and long term strategies.
|
UWA
|
380
The model specifies documentation that defines all necessary activities,
responsible groups to perform those activities, databases and reports they
use to perform their task successfully, and format and content of the
expected outcomes of their activities.
The model is based on the two main activity streams necessary to
cover all requirements for assets in service:
a) definition and execution of maintenance and repairs of assets, with a
broad term of asset maintenance management;
b) monitoring of asset performance, assessment of condition and location
criticality of assets, and planning of necessary actions to respond to the
identified asset problems (eg modification, refurbishment, replacement),
with a broad term of asset renewal management.
The model structure and its process, functions and procedures, required
base documentation, and results and output documentation is presented on
Figl.
2.2 Development of the Model
The model begins with an asset management strategy for the transmission
assets, presented in the document Asset Management Strategy. It follows
with a definition of aims and goals for the asset management process in the
transmission business, which is presented in the document Asset
Management Policy Manual.
The model is then further defined in two other base documents: Asset
Management Process and Asset Management Procedures. The first one
covers the process and main functions of asset management process. The
second one deals with documentation that is required to support the above
|
UWA
|
381
process, procedures to follow, databases to be used, reports to be prepared,
and format and contents of the required outcomes.
The Fig. 1 also shows a link to development plans for the transmission
system. They contain required future works in the transmission network to
cater for fault level increase, new connections and increase in load transfer.
It essential to correlate all asset works between management works and
development plans.
2.3 Asset Maintenance Management
Asset maintenance management is an organised process that makes sure all
installed equipment has adequate and up to date maintenance policies and
that necessary instructions how to do maintenance are in place for assets
inspection, checks, testing and adjustment. Dedicated and qualified
maintenance service providers apply work per these instructions and policies
on a regular basis.
It also defines procedures for effecting asset repairs, emergency response
and follow-up actions in the event of asset failures, contingency provisions
to assist in event of asset failures, and for proper handling of replaced assets
when failed or under performing assets require replacement.
The maintenance work is handed to internal and external service providers
through Service Level Agreements. They are monitored and assessed by
using various maintenance analysis techniques for effectiveness and
efficiency of their work.
Some documents prepared and issued to cover the above areas are
considered base documents:
- Maintenance Policy Manual,
- Maintenance Servicing Manual,
- Maintenance Instructions Manual,
|
UWA
|
384
3.4 Service Level Agreements
Service level agreements define relationships to service providers and cover
the following areas:
a) Service Level Agreement Purpose and Policy,
b) Scope of the Agreement that defines Assets, Services to be Provided,
Rework for Services Provided, and Qualification of Personnel,
c) Performance Management that establishes Performance Measures,
Budget and Cost Accounting, Acknowledgment for Services Delivered
Well, and Need for Improvement,
d) Continuous Improvement System that provides for Periodic Reviews,
SLA Revisions, and Work Performance Improvements.
3.5 Generic Contingency Plans
The generic contingency plan document provides planned responses to
possible asset failures, resulting from consequent or independent co
incident events, where such failures take the transmission system outside
the boundaries of planned operational risks.
The events can be caused by a variety of causes:
> Environmental sources such as lightning, fires, earthquakes,
cyclones, birds, trees;
> Human sources such as vandalism, accidents;
> Malfunctioning of other equipment, explosions of adjacent plant and
failure of protection systems.
These plans ensure that every transmission asset in the network can be
suitably replaced and power supply restored in the proper time.
|
UWA
|
385
The asset identified in need of further consideration beyond the scope of the
generic contingency plans will require a special contingency plan.
3.6 Business Case Analysis Manual
The main items used to assess asset performance and condition assessment
in a business case are:
(1) The age of plant,
(2) The frequency of failures,
(3) The plant condition,
(4) Reviewing impact of changes in environmental and regulatory
requirements,
(5) The current and future maintenance costs,
(6) The replacement costs for new solutions,
(7) The spare parts availability,
(8) The skills available internaly or externally to repair or completely
refurbish the affected units,
(9) Safety of employees, customers and public is part of a constant review
to ensure that equipment operates under fault conditions with paramount
security,
(10) Integrity of equipment adjacent to faulted one and risk of releasing
contaminated substances and particles into the ground or atmosphere,
(11) Obsolete design standards are reviewed for older units in service,
(12) Impact on quality of supply,
(13) Impact on customer,
(14) Impact on utility,
(15) Business ability for a proper and timely management of its asset
replacement programme.
|
UWA
|
386
4. ASSET M A N A G E M E NT FUNCTIONS
4.1 Asset Management
Asset management is responsible for management of asset management
process and documentation relevant to that process, and for overall
coordination of all other activities related to asset management that are
performed in other areas.
4.2 Network Planning
Network planning is responsible for setting network planning criteria and
plant performance standards to meet the system needs, and for preparing
network development plans to cater for future load growth and new
connections.
4.3 Maintenance Services
Maintenance services have the responsibility for construction,
commissioning and maintenance of transmission assets, including short-
term planning and scheduling services for the required work and local
maintenance work analysis and correction.
4.4 System Operations
System operations are responsible for the day to day operation of the
transmission system and use of power system assets to deliver the product to
the customer. That includes responsibility for operational reliability, security
and quality, and for the SCADA operations to support this activity.
4.5 Engineering
Engineering is responsible for providing engineering solutions and services
in the area of design, drafting and cost estimating for failed asset
replacements and new installations based on planned renewal projects.
|
UWA
|
387
4.6 Project Management
Project management is responsible for establishing and management of
planned asset replacement projects arising from approved capital project
recommendations to replace assets according to asset renewal programmes.
4.7 Finance
Finance is responsible to analyse requirements for operating and capital
expenditures and to determine responsible costs to ensure the ongoing
financial viability of the business.
5. ASSET MANAGEMENT DATABASES
Databases are required for recording of assets and asset management
activities in transmission business, and they can be divided into two areas:
5.1 Asset Information Registers
• Transmission Plant Management System
• Transmission Lines Management System
• Transmission Protection Equipment System
• Transmission Ratings Information System
• Transmission Geographical Information System.
5.2 Asset Activities Register
An integrated information management system is used throughout the
company to record all asset activities.
It contains details of location and plant fitted to that location, asset standard
type nominations, standard maintenance levels, required maintenance and
testing activities, frequencies (trigger dates), standard job templates and
their cost, maintenance history and repair activities, outstanding
maintenance work with its cost and planned future schedules.
|
UWA
|
388
6. ASSET M A N A G E M E NT PROCEDURES
There are number of procedures that are required to enable the asset
management work to succeed, and the main ones are briefly given below:
6.1 Recording of Assets Data
All assets need to be recorded to enable setting and planning of their
maintenance and monitoring of their performance and condition in service.
6.2 Recording of Maintenance Work
All maintenance activities need to be recorded against the relevant assets to
enable analysis of their system and maintenance performance.
6.3 Recording of Asset Failures
All corrective maintenance work done separately from asset preventive
outage programme is considered an asset failure, assigned proper failure
codes and recorded for future analysis.
6.4 Review of Maintenance/Failure Statistics
The above statistics are reviewed on regular basis to determine assets with
poor performance or high maintenance costs.
6.5 Review of Network Development Plans
This review needs to be done on a regular basis to ensure that assets
maintenance and renewal plans are timely adjusted against actions
foreshadowed in the network planning future projects.
6.6 Review of Asset Management Plan
The Plan needs review to incorporate any new non- performing assets,
assets dealt with from since the last review, and impacts of changes in
development plans.
6.7 Review of Asset Management Process
The asset management process and its success in managing performance of
assets in service need to be subject of regular reviews to confirm that the
|
UWA
|
389
model and its procedures are adequate. One of the tools that should be used
in that assessment is benchmarking vs its industry peers. The results of such
studies should then be used through development work to close the
identified gaps in performance or policies.
7. ASSET MANAGEMENT REPORTS
Continuous monitoring, review and analysis of assets performance reports
expose assets with deteriorating maintenance and system performance.
7.1 Asset Corrective Maintenance Work Report
Review of this report is used to decide if asset maintenance regime or
servicing instruction needs to be investigated to obtain more information
about its performance or condition.
7.2 Asset Failures Report
The review of report information will enable to determine if a further
analysis of asset failures for asset types represented in these failures is
warranted.
7.3 Asset Failures Summary Report
This report is used to highlight on a regular basis asset types that have fared
prominently throughout their total years in service with regard to their
failures when analysing the failures number, type and percentage to their
total population.
7.4 Asset Population/Attrition/Performance Summary Report
This report is used to highlight asset types that have fared prominently
throughout their total service years with regard to their terminal failures (ie
plant had to be removed after the failure and discarded).
|
UWA
|
390
The above and other reports results and reviews are used to show if and
what further action is required on assets, which are than selected for the
'Plant Under Review' part of the Asset Management Plan.
8. ASSET MANAGEMENT OUTPUT DOCUMENTATION
8.1 Maintenance Plan
Maintenance Plan is a complete list of maintenance work to be done on the
assets currently operating in the system to ensure that they continue to fulfil
their intended function in a cost-effective manner.
The maintenance work in the Plan has three main maintenance categories:
• Preventive maintenance work,
• Corrective maintenance work,
• Major works (after business case analysis for modification or
refurbishment and operating funds approval).
8.1.1 Preventive Maintenance Work
This work is planned in detail with its scope and frequency and the work
party, it is scheduled well in advance, and always catered for in the work
budget:
Routine maintenance: substation rounds, alarm checks; line aerial
and ground patrols:
Servicing maintenance: CB maintenance, injection and testing of
relays, line washing.
8.1.2 Corrective Maintenance Work
This category is subdivided into three groups:
Emergency (unplanned) maintenance. This type of work occurs with
no warning and at random incidence when equipment fails or needs to be
taken out of service immediately before it fails;
|
UWA
|
391
Deferred (1) (planned) maintenance if corrective action is taken
sometimes in same financial year;
Deferred (2) (planned) maintenance, if corrective action can be taken
in the later financial years.
8.1.3 Major Works
This type of work is always planned and budgeted for in operating budget.
It is classified as major work on plant types or line hardware that needs a
business case and operating funds project approval (involves technical and
financial business case analysis) before the work can proceed. It is divided
into three groups:
Overhaul (refurbishment);
Modification;
Remedial (repair).
8.1.4 Maintenance Codes
A number of responsibility centres, activity types and maintenance service
groups have been set up for maintenance activities to enable setting up
annual budgets and to monitor work and expenditure per asset and per
service provider throughout the year.
8.2 Asset Management Plan
The Asset Management Plan is the end result of asset management process
that aims to assess condition and remaining useful and safe operational life
for assets.
It identifies renewal actions that installed assets might need in the future to
ensure their effective, efficient and safe ongoing performance in system.
Asset renewal is therefore focused on managing the performance of existing
installed assets through major repairs, refurbishment, modification and
replacement of assets, or update of their maintenance policies.
|
UWA
|
392
The Plan contains details of all current and future, approved or planned,
activities for the installed assets based on the asset condition and
importance of the asset to the system operations and security.
8.2.1 Assets Under Review
An outcome of the review of all asset performance forms Assets Under
Review list, ie assets that warrant further analysis. That analysis may lead
to one of the outcomes:
(i) Performance is acceptable, no action;
(ii) Further information is needed about the asset in question. That work
is called Asset Investigations;
(iii) The asset is assessed to need future action and is included in the plan
of Asset Future Projects in 5-year timeframes with a 20-year horizon.
8.2.2 Asset Investigations
Asset Investigations lists assets requiring additional information about the
plant performance, condition and other details before decision can be made.
The outcome results of these investigations are then used to refine initial
assessment of asset condition and performance made in 8.2.1.
8.2.3 Update of Documentation
Update of Documentation is a result of Asset Investigations or Asset
Business Case Analysis, where identified problems can be corrected by
altering asset management documentation.
8.2.4 Asset Future Projects
Asset Future Projects lists the assets that have been assessed as requiring
some work in the future, but a deeper analysis of their condition will be
required to determine the best option to resolve the problems.
8.2.5 Asset Planned Projects
Asset Planned Projects is a list of the projects recommended to proceed
after a detailed business case analysis of project has been undertaken from
|
UWA
|
393
the list of Asset Future Projects, and have received an approval but have not
started yet.
8.2.6 Asset Current Projects
Asset Current Projects is a list of all current on-going approved projects in
the areas of asset renewal (repair, refurbishment, modification, and
replacement).
8.2.7 Asset Completed Projects
Asset Completed Projects is a list of all asset renewal projects completed
since the last Asset Management Plan was issued.
8.3 Business Case Studies
A detailed analysis and review of asset performance of assets with a
suspected condition, listed in Asset Future Projects with required time
frames for action, to determine required action.
When the time comes, asset management prepares a detailed technical and
financial business case for a particular asset project.
8.4 Special Contingency Plans
One of the outcomes of the asset business case studies for the suspect asset
is a special contingency plan for that group of assets, to cater for their
possible failures over specific period of time, or while some other planned
developments occur.
Typical examples of such plans are Metalclad Indoor Switchboards
Contingency Plan and Rapid Response Spare Transformers Plan.
There are also other special plans for response to particular catastrophic
events beyond normal operating risks, such as Catastrophic Failures in
Terminal Stations Plan.
|
UWA
|
22.4. Maintenance Policy Manual Structure
NETWORKS
TRANSMISSION MAINTENANCE
POLICY MANUAL
POLICY
The purpose of maintenance is to keep plant in an acceptable condition,
with adequate availability at minimum system cost over its life. This can
only be done if maintenance is done in a controlled way. It is Networks
policy to control maintenance by having a fully coordinated and
documented maintenance system. This ensures appropriate maintenance
practices are uniformly adopted across the whole of the state.
THIS MANUAL
This manual is issued by the Asset Strategy, and defines the maintenance
requirements for all the primary plant, lines and cables in the transmission
part of the networks.
The manual defines the responsibilities of all groups involved in the
maintenance activity, from planning through to actual performance.
In addition, the manual lays down the criteria on which the plant is to be
maintained. The criteria for some items of plant have not been agreed as yet
but as they are, they will be issued to you for inclusion as part of this
document.
It must be remembered that as the maintenance process is refined, revisions
to the various portions of this document will be made. As such, each
manual is issued to a specific position, and that position will receive all
subsequent updates and revisions.
Thus should the manual be forwarded to another position, please do advise
the Maintenance Development Engineer so that the circulation records can
be amended.
This manual is part of the Asset Strategy overall documentation system and
should be read especially in conjunction with Transmission Maintenance
Services and Maintenance Instructions Manuals.
|
UWA
|
402
No.l6Responsible Centre and Account Code Numbers (I)
(#366968) Rev 3
No.l7General Use Telephone Numbers (I) (#366940) Rev 1
N0.I8S & H Implementation Plan for WPC S & H Policy
and Plan (#407183) Rev 3
No.19 Standard Operating Procedure-Safe Handling and
Application Of Sylgard HV Insulator Coating
(#1012001) Rev 2
No.20 Procedures for KDR-FMMS Data Update
(\R\0998AN.DOC) Rev 0
No.21 Procedure for Return of Primary Plant from Substation
Site (#376298) Rev 0
No.22 Procedure for Creating, Maintaining and Disposing of
Facilities in MIMS (#385304) " RevO
No.23 Procedure for Fitting and Defining Plant Items to MEMS
Locations (#385307) Rev 0
No.24 Procedure for Creating and Closing Work Orders in MIMS
(#377740) ~ RevO
No.25 Procedure for Management of Operating Statistics
Related Work in MIMS (#1047058) Rev 0
No.26 Procedure for Recording Plant and Line Failures in TPMS
And TLMS (#1010348) Rev 0
No.27Procedure for Plant Arrival, Notification and Management
(#1038769) " RevO
No.28 Procedure for Preparation of Asset Maintenance Plan Rev 0
(#xxxx)
No.29 Procedure for Management of Primary Plant Purchase
Contracts (#xxxx) Rev 0
LAST UPDATE: 30 NOVEMBER 2001
|
UWA
|
403
22.6. Asset Maintenance Plan Budget Summary Sheet
Ace. Code Description Lab Lab$ Mat$ Equip Other $ Total $ Work
Hrs $ Group
20782000 Prev Maint - 2255 66625 0 0 590750 657375 ESP
Line Washing
20782002 Prev Maint - 2624 94950 0 182600 277550 TWSC
Underground
Cables
20782004 Prev Maint - 0 20000 20000 TAPA
Substations
Misc General
20782004 Prev Maint - 399 10594 289780 0 46803 347526 ESP
Substations
Misc General
20782005 Prev Maint - 1560 45000 295000 0 22000 362000 TAPA
Plant Test
Failure
Replacement
20782006 Prev Maint - 0 70000 70000 TAPA
Overhead
Lines
20782006 Prev Maint - 33452 845100 343900 0 1088375 2164875 ESP
Overhead
Lines
20782010 Prev Maint - 5395 164040 608500 0 300350 1072890 TSERVC
Easements S
20782020 Prev Maint - 262 9700 1600 0 7050 18350 ESP
Pole
Chemical
Treatment
20782022 Prev Maint - 2200 62500 461400 0 192350 716250 TAPA
Pole
Replacement
20782035 Prev Maint - 1000 5000 10000 0 0 15000 TAPA
Graffiti
Cleanup
20782036 Prev Maint - 500 25000 25000 0 0 50000 TAPA
TV
Interference
20782040 Prev Maint - 750 40000 0 0 83160 123160 TAPA
Switchgear
20782040 Prev Maint - 26805 767918 117448 83318 105007 1073759 ISPNSD
Switchgear
20782041 Prev Maint - 5220 131818 19600 0 71602 223020 TWSS
Substation
Insulator
Washing
20782052 Prev Maint - 3100 140000 1220000 0 510000 1870000 TAPA
Mods/Refurbi
shment
20782055 Prev Maint - 6546 214200 0 14400 228600 TWSS
Warranty
Testing/I nspe
ction
20782060 Prev Maint - 16639 358236 37696 0 52591 448450 ISPNSDN
Substations
Misc
Electrical
|
UWA
|
404
20782120 Prev Maint - 250 10000 2500 2500 2500 17500 TEMT
Substations
Equip HV
Testing
20782120 Prev Maint - 1781 48130 0 158820 206950 TWSST
Substations
Equip HV
Testing
20782125 Prev Maint - 0 38880 38880 ISPGEND
Trans Oil
Testing
20782135 Prev Maint - 50 1500 3000 0 21000 25500 ESP
Pole Testing,
Mechanical
20782205 Prev Maint - 4072 154525 0 3003 173651 571630 ESP
Line Patrols,
Ground &
Aerial
20782240 Prev Maint - 1276 37112 0 34396 71508 ESP
Substations
Routine
Inspections
20782250 Prev Maint - 0 0 0 16000 16000 TNSDG
Thermograph
ic Surveys
20782250 Prev Maint - 492 21220 7380 0 5400 34000 TWSS
Thermograph
ic Surveys
20782310 Maint Correct 1500 45000 20000 0 120000 185000 TASA
- O/H lines
20782315 Maint Correct 1500 45000 20000 0 10000 75000 TASA
- U/G cables
20782325 Maint Correct 13500 400000 150000 0 61000 611000 TAPA
- Substation
equipment
20782405 Maint 8200 246000 126000 0 150000 522000 TAMM
Emergency -
Substation
equipment
20782420 Maint Emerg 1000 30000 25000 0 65000 120000 TAMM
- O/H lines
20782425 Maint Emerg 2000 60000 30000 0 30000 120000 TAMM
- U/G cables
20792003 Prev Maint - 750 40000 0 0 10000 50000 TAMM
Protection
20792003 Prev Maint - 11087 339979 4580 0 0 344559 ISPND1
Protection
20792301 Maint Correct 0 0 0 0 45000 45000 TPLINES
- Protection
Equipment
20792325 Maint Correct 2000 100000 20000 0 10000 130000 TAMM
- Protection
Equipment
20792405 Maint 2000 100000 20000 0 10000 130000 TAMM
Emergency -
Protection
Equipment
13053332
|
UWA
|
00 CN 00 00 00 •stCD^-COCOCOr-NCOrMffl.CO C30 >—T» a—TS 00 LO
o 1- 1- 1- 1- I— I— I— I— •— •— I— I— I— •— I—I —I — •I——
la
%
o co LO CO 00 1^ CDOJOOOOCOOJ^-OOCDCMCMCOOOCO 1^ CM CO CO
o CO CO LO ^ S S S C B r O ^ C D ( D C O T f f N r- CO ^- LO ^ •—•
P Ho S . 0 L0 O CC TOO C CO O C LL OO M r> O iC M f C l3 C> MO CO MO oO O mL ^O fO NO CO M> I i O -L tO ^^ r- oL ^O • t _5 : r o -ro Lh -^O-
r
-oc
"
o r c •o o
*
(ocococo^^-^-^-^^fco^r^r^-t
O
T~ CM
i^ F^
CM 00 00 % T— •sT CO co LO
MM CO CM CO CO o
2 T— 5 CQ CQ •*r
'u o —- T™ o 01 00 CO CO LO C/3 CO CO < < LO
% o o o o o <*• o LO LO LO o
9 o 00 CO 00 o
S O) O) co co O)
n o LO CO CO o
§ _> HI LU LU _:
CL
O
JZ co _; 1 c a "CD o o CD > Q. Q. ro o c JD o o _ a, co c E E X) 'uu CD CD o o CO JOZ O LL V) (0 i_ M— f/1 0 -) la Z L. 0) M— 1 u CD o DC
__! co l_i O — •* o
o T— CO
£ T —— " L•* O
H h-'
kael liO kael liO !sageR
LO CO
co CM O o
C C 0O O
0
C T—O C C oO O
00
Cr O- oCO C oO T C—
O
0 C0
M LO
L LO
O
0 T~0
LO
C oM
>
o 005 00
CO
T C—
O 00
LO T— Tfr O) o ro •<* •<fr LO LO o T— •<1- —- CM ^ S T— •* CM
Co L OO C C'tM O f i £ o Ce O n C O CM O C o CO O o L LO O Lo L OO CL < OO J ) o CL OO o C O CM M C T •*— O o0 C0 O C CC O OO COC M O C >O - C >O - C CL M OO o Co O Cr •* M-
o CM T C— O o o _ 00. o o LO o o oo oo oo O o > oo o •^f o o 0r 0— Co O o CO O LO Co O O o >
o CO o CO o ^- o o o o 00 o t^~ o r>-C33 CJ5 CN 00
CO ^f CM •^r h- ^ O) oo O) O) CD 05 CM h- CO CO LO CO
^r o O s o o o o 00 O T- o h- cn 05 s O) O)
CO < o < < <- _: -> _• 1- _» CD < CO >- >- CO 1-
CO
_ CD CO co
CD :rr g o CO £Z
Cc C OOo CC CO OO r L L^ O
O
-< C C J Cr r C Q oC C o o
C
>OO 0Z O D _ D.o o D CC —Co U
O T
Cc c orC O O_ 1Q
—
M0Oo o g oJ J
o r
r- -C r C> r > C C E1 »Z D— O—O DD
o
o '»
-
o I
o
0J! J tC CC(o c r rO c Z ^c 5D L0— D D o o -o Jf
c
o" T h3 o O C oC C l C Q O O5 3
-
D D c o_ O O . oy . JT
c 0
oX or c r 5c o c D3-
0
_
o
o o o I
0 C
CC t o C o c
O
O5JD o o1 C
• •L*
*C c o oo _ Q OD o C' _ oC_ J J >r r C _c CC Z rD
O
O;O DO * o o o
oC
CJ J1 C r _ C o O3 D
O
O=D _ o O CC OCJ - • 3 O rCo o Ci Z O MQ
O
)o\ DP i o . . J (
C •
LJ '* rc oC (C C r C >Z D
M
O*O 5_0O O O - o o * O" T J C
c "
L- rC C oZ D Q 33Q
O
o
c f O
-
o " J _ oT
1
1X 1- C ^<co C QC Z Jo C 1 D i• 3 -
^
OD O— o— -D . » -Z. •
JC
^" J
L
o r oo r i0 r C C oL Z
OO
O— ~L— ED ) o o o C oo JQ C Dt • o o
f
(EX oo * O C 3a 3—o
- 0
« D C o_1
1 ^
0J *oC £ r o o CD £ ^O 3D O t- Z Z
-
o » r O* ^ —
C L
ICJ J C rcc E oM * -Z D M =
M O
o o o -— Cr C
c ^
cJC C X( 1 Q L o DDo DO D _0 ~—
o
o. J
c T
CJ J Z r o o coc Z EZ —D
O
i D
o
o o
c "<3o - c •*-o C LM O o L CO O C CM O C cM o o •< * L •^ O " •^•* ••*3 - C •SM T C •SM T • c* o " •*3 " 1 •*^ -c 'sn t • •*^ l- J ZD J o •<3 - L ,^O - o ,^ -C ,^O -
H H \- JrZ o h- h- I-"i -H I-'( -'H i-H I-'I -'I -' H CD H K 1^ h^
a o a o o U(_O 0 0 0 0 0 0 0 0 0 0 0 0 00 cc o o o o
C3
Q
-a t- q q q o q q q q q q q q o o o q
5 CM CM CM CM LO q q
o CM T- T- T~ T- O CM CM CM ?= ^ CM CM ^ ^ ^ CN O CM
CM CM LO j^ q j^ T- T- O ? ^ CO CO O
CO CO CO O LO CM ^ ^ CM CM CM •«- T- CM
CN T- CM CM CM
T- O) ^ N S C O C O C O T - C O L O O L O T I- C3> oo a> a> co oo o a> LO o
o CO CO CO CO CO rocM•^ Cr M L ^O r ro oo c D^ cj o- cc oo c oc ro o cococMcO o coO o orC oM ' ^C iD - rC oO ro CM s co
CO O) O) O) CN
CM CM T- T- CM
CCOO COMO CCMM CCOM CCO OCMCMCMCOCNCMCMCOrocOCOOCO
CO CO CO CO CO P £> P cp o o CoO COCOCOCOCOCOCMCMCOCOCOCO
O O CD O O
CQ >- > CQ
o o
fQ < CQ I- \- CO CD CQ CQ CQ CQ CO
CQ h- x x o c o o oo Q X O O O O x o CQ CO CQ CQ
o > O >< Q O O £_ Q^
£2 CO L —oO
I
M L —oo IaO o i OI C - Q
—
—C C Ct OD Q-
_
< -o L L _NO - _o XL
_
O
^
.° C O
C
M
O •
. -O LC
, O
M v^ .C CO O
O ^
° ~ J
O
y* O-
+C M
—L O L Co O O
O
LOo O
C
Oo Lo CC O OO L.Co O M , |__ Lo o
O
o o L aO
.
L r
P
-O • i ° Co Oo *. "o
C
C OMP c o • C oL "^ OO
o
_ L hO
-
<<<<CQCQCQCQCQCQuC_Q CQQ-C QaC.Q ICs-Q O•O- — ^ 0. « 2 z a. a.
>- >- ^ -I
^ z z a. - O a. o_ UJ UJ
o o o o o
|
UWA
|
ON o
o
© oo p P i- 00 5 CM hi CO oo r^ CM P CL CU 00 =l/ 3J O) P £ oo
CO
£ <
P8 k o z r z- p *: 2 i o- L z CL. J p < c I Clo - Oo u C£ O. C Z Q 1 CD- C ) o> hQ -. I - Z 0 h o-3 . 3- o 5a X oL U L C LULL z O C C C (5M C O - H- 2 Z3 C m LeCi UL On n 7 OLL m m z h Z Z 2- ) z 2 > <
I
OhC -M 5<z z CD
so
a
•*- o o o o o o o o o o o o o o o o o o o o o o o o oo
88
C•**O
o o o o g o o g o o o o o o o o o o o o o o o o o oo
< < < < < < < < < < < < < < < < < < < < < < < < <<
fligocjgcaogocioooopoooooooooooo
5^t^o3i^ooooc»d»oooooocnbocBc»o>53 o o S o S o o o oo
o
PlQa"j^C M^l OoOQ OoO CIMO CiOnO O5o5S oitnS cnT^CMtOOOCM
LJ O C M C M C M T - O 0 C M C M C M ' « - O C M T - ' < - - « - 0 ' «- C M C N C O C O O C M OO
013
a
H
IS
cocococococococococococococococo eo co co co co co co co co
h-r-i-l-f-r-l-l-l-i-i-l-HI-l-r- I— I— I— I— I— I—
W
88
t/3
O
a.
a
>-»
T - ^ T - T - T - T - ' . - N N N ' - ^ r - T - T - r - T - T - r - T - ^ T - ^ ^^
H > > > > > > > > > > > > > > > > > > > > > > > >>
O
PS
PH
I- (- I— I— I— h- H r-
O
CD
o
oCO
co 3
pc
O) • ca u
•C caO
u O cr i ( C CD DO ™
C
o
iZ rC oO
oG u X si c c o T a§ 0 Q3 . . - o C e QJ
.
)
O
T cc c >J ozo c ot < O coo / o- > r 0 oL_1 e j^ T coC Cc 33 zoD - C C oo o—C * O U
3
E T
0
JrC c Z3 )oD
1
e • O C O mc g D g 3 c > ou r oL _ C c <umIU L- c a Lo m m 3Ud acC C uLO .• &
D
Cr3
. D
e C c cu uO ^ .7 o- 23 = c§ Zc uu ) h cct -z~
c
c I O -S
o
Q o C ^2 Q > I l< - Q . C ta L: i o C * c C- D o » h,C -oU -To 5 a Q C C. O O u o QCO . i- CO1 nO3 e — O , c Cu- L o= Or e 'Qr o 7. 1e z D O r •Z > aJ r O _i- x ci u hCc Lo - Z N£J - Cc J) S erz> e p C OO Q Z ^ i tO O "3O I ° D "ZJ o
O < S
3
i»" 3£-5 5C C "CO O PLZ < 3C > CU D T r 33Q Q c C3 - O. . c± 5C3 o± D 8 c L 5c crio za 5C cD S_a c oCl O ) o 3> < I C D 5S~ „r Co CC OZD .e - S• —c ca £C I: " CuD. 5• JC 0o COO Z J.) ST O oJ ) a S• OQ QQ • a ) .) . 3hc Q Co D. 3< oB CDJ C 3au C D : .
3
- < Li L
3
r H O C C oe £ O SC oDc C o Z 5. ZC 0 CcL z Di —
5
C C> I CO O
D
.
3
• ^ CcO Co Og o 5CT . 5C CQ D9J D . - C - , DJCC oZ9 Co :L D -
D) CD !c CZ CZ
CC O CCZO C O C O C O C O C O W C O C O C O C O C O C O C O C O C O C O C O W C O C O C O O T C O CO
o
o ? 2 5 t N * 0 ( o ^ 0 ( | N ' - ( O c o o j O L O O ) a ) i , ) ( M i > ) c o o ) N co
h- CO CM CM CO O) •*mcnmh~co'<*-mLOCDCJ) co in CM CM r~- m oo oo
o f N S S r r Q S S ^ ^ ^ ^ O J L O i n i n c n L o i n o o oo "" CM CM CO o> o o
h - 0 0 0 0 0 > C 3 > 0 ) 0 ) 0 ) 0 ) C A 0 > O O O O O OO T- CM CN
S S i E i D i H y N C M r M N C M c o n n no cocococococococococo
2 S 9 S o c > c > o o o o o oo c o c p c p c p g o o o oo
|
UWA
|
Cu o
UJ
1—I
00 •a to «* cza
c3 e K o to s
•a 33 «
13 S" 3 a. eS Q
c t- c 60 fe ro"
o .« .£ ft
CN
o> ^
•o
c —
"S 8
<3 8o1 -
cu ° cu 3 c
c PQ
oo S W e UJ « SJ
Cu a to . • uo c a o e. L 'cL- u. , : 2 3 § •*c . C & |'£ ° .2 a u-
00 c 2 o> ? ^ *o o uu X u - » " ^w 3 _
2
& c£ ' U3 5 ~ fJ t tC >0
^
0 t . o s o or o\o
3 VJ •<3-
Di tu, T3 3 < o
C~- © g cS to (_ oo
— t« a. u I «> o
53 •- U CO c
•a * C °C _ s. to • - o
J3 g^2
2 '•% ~ C u
c "• g 4) « 3 ft Q
o a. ° 2 o ft ^ o
00 3 (j) * u C o o
c o to 60 O o T3 O O ^5 S
£ •a IS 5 o c 2 C/3 «R
n. ±s <= ^ C to « oo ft Q uu
Jo Z Sen 1 2 o -^ B ft. 5 -^ CO
CM o5 ca o "^ ^! -S o O t_ - • T" 3" CNo
a o a o- - > 2 ° ° CM tL> . e
l3g \0 <u •- to "a JS «! o
£ .S2 uS Q' 3 «5a: c Oo Ut o
a
-a o
J 3 -w * „-CN g ft •- « 2 -= CB
£ Co D Oo t ^u Cu 3 ft, "43 o M!
W t«7 B
o 2 •& a 2^g3 | o oo
G co E c
U c cu cd t. a a ^ °
•* =3 <° E « «
-*- s -s *~ o e ao xo : E BO
"o aim
3
E ECU o So I o S • • • c toaJa 2 < o 3 5£ s OU & at j n -cS t D «LSfs3 -sO t O 2 o C^ « •~ U o£ 1 a . —0 u" ~- °S0 t a 3 ^ a2 S 2 -* 5 ' «• .S t-c f o = o t 2 ts ES f _ o «S3& M " ~ Iu to c ai. sov —o o c . »C u So ~o S 3o o • 5 oH a o
«
£ ! J a 2 «O«-. 3 y M fT U ^f 3 t t Cc SL >
o to 5 =f t x2> •SS° c"uo "s o > c tu a* 3
to U f- tR S- ; Zw t o o o> — D' O t S II U S3
"3 tu o o o —
D. o o O o
CU TT
Cs U H 3> > o CD • sa o o o• O* o o o o C 3 OO ~ ^ CN
« Q CQ Q r- C oN o CN o CN o CN CN
Q Ce Nn o C N o
fa.
K
O Ci
H 33
' sJ*^» "B O Q tu
CU
2 ••-»
3
L chU - o o fa- ^ S a tc uu 33 1 o oco u O U
O C zO a CU O 3 2 Z Z t u 3 E-3 Z tu X
*<, o C/5 OS z CO 00 -J c to oCC oj oa o H r~ -
O W 5 f• C3 + aJ S n VO ^o V ^O o t t to o o tt too o ^ tr o o^ o t Vo o O
CO
z
2 S
h- 1
o
c c c c c c c
'si" o o o o o o o
Z Z Z
Z Z
|
UWA
|
o
?tc oz 8mu i * ct =o
a
4
7c
33
a C EO
1
C
C3
D D
>CD •C aM roS - .o
2
•a p CO a ca c «, CL
CD V o tC a
C
oO
J 5
co aC < cDO u r 7 CC Zi 3 OL J " C2 —c 2O 7E C 33O 7 oa C 33 Z C ED 73 ™ .- ? 8n . 2
I 5 1TO C ZL
-
C CC 5' O DO j 5 j j ~C C g O O 2 C«
N ^.
-J • c cC Cc aZ O D a o 7 TC C C C3 3D U O D o "o • 7LcC Cc Lo 3O Duu
IS
f 5 Po t - 3c o>O" S fTX C To C C a C> J
J
OOD J U £C O CD
D
JC c C CcO D U C cO a o cc o c z
o CD
7o CC3 LO
CD
33 o!7 c C3 o
1
r -—« o cO
C
< O1 CC2 - DD ;
3 c UC C x t > _ 0 o o — c| r cO CQ -O o> l O o- 8 D w. T b x iJ* Z C OCO CE c$ 2 t * J ; O = Oo i ' < C - T r<tUC C C OC J" *~ -™ D=Do OZ U a « a O 0 f r-^ C ^oO o ) Q z tr I > o o v «M C- - " CO 3 :MC a, asC L C i -„ D7 IC s a £ ?3 — -Oa S = S C P7 D « C a CC f = cC3 U L O DD) >t o J I mD 5 J 7-O < ,3 3 Cc °C oC O Q SS ™ 3D D O >2 2 J iC CS L7C =QC or OO OE £oU^O : "O o C Q § 1 ,- c I S 2 O 1 *o D 0 J> a 2 S *^ 2 ,C c ° Lc Q C _ oO 2c o 2• UOo o £ C C 7 CCC Co D 3 DO DOD >C 7,p7 T. 3C .Ot 3 £ = - Oo >_ mQ i • E8 — a.< : S ,. n = aj I iC sa 3oO . f c £-o o C eCC C CC ^ nt u Dc D D Z OD - £ c 5 o o tCoc DC Oo -D a ) " o CCm •o c 3 occ 3 Cc CcCc C C DOc 3 aoDO oD O a o a i J " X H Co i c cE Cs 3 r C J o ai: C" n7± f o C 32c o f C< C3 -C * O3i 1 O D 0 - t t - 2 c&c C 7 * ccC 3 o- oc Q oO u r t X C3 5 7 r i.B 3gt o a C3 C CC I 3 o o O ^2z Z O D O «*C T 7 o= oo cc cC 1o ^3L J Z a : o 0 - ij-• C XceS 5E o£o S^ " O sun 1u C7E c c oi " JJ Sroc Z3 » = Oo o Su C '" o M t i "c r o^ L- §3 ^ 8 3| :.co g t - C o " J>I. &M- C 3 50 o Z->£ - D o
o E E o O= P 5LO - ? x • C i£ D o S' 2Cr 3S O o* .x Ca *O : pCaS O s o * >5< = , 7 cCc c cC 3I uOO u z 2 Q <f OC t o O * f -< tOaD • C 2 C CJ O D U L> S o -c S ( C o 0 a 3 0> C Di T -CCCC C IJ I DLJ ) LC ac 3o CLO o O r C.t 3o a Oo r• •Ctc c S_ -Ozf o ^ <C3 >1 oU IW£ E -/ O' T C £ — OD o C Om u ci o O _ u • ccoc C CO aD = oo 7m L* C C Cc 3 L- D D D a c cO ozo z C* c CC C— Q^ U Da D ^ L IJ U >i SC C> U< -J OZ T ?acJ * CI O Oo- 3 OZ • Co Q" •p ,O£ 2 >3= 5 O l^) p S 7C— ^ cS . aO 3g > o o -' * ,Cc pC oa0 ;>O > J ,o i - . i fCQC — OD t
o CU CD DQ < O o o
Q ® en
o 0) < a. co
o CD
CO
E to
O CJ) CD
o CD CD
" s 7C 0I ) O CD a a .E 2 - c . t0 o
r V) o o O 3 o U U U m L.
0o S ^*
)
«- =« °S% 1
f
ts r L L
a
o32 L Ua . Zt s ^, CC3 ! ir DD ="
.
< 7C E1 ma £3O_ i Q a
C
<o2 a 1— L. j 3 a O C <_t cio OC L. D C> 5O OL tc C E
o
3c 1U o— Do r C <£ < os LJ L C u= zU O
i
L fC
a
oC jLD O
tz
C
O
aL oc fsDL 2 o
t
a
C
O
Q fL 'C Q
< aO
Jc L tZD o C
:
D fC L 'CD tLO
J
Si £
Q
C
<a
_
.' £ io o > •
L
2
_+ • ? 0 _ .U iOo l - _ . 2f '* C Q <_ 0£ .* ioo L -
.
2
£
.2
cO C o3C f aD oU t
5s. oo o 5S a3 C D f 3 Ct D -t » C- D •C £O fc3 o f 3
CD
t £3 2- ta5 zj tt a z) C a3 1 D— £3 £3 £3 £f 3t 3f £t 3f £3t f £t £f 3 t
o o o C o o o • oa a a a 0 t 0z t 0z t 0z a 0
s0
T 2 0J 73a
3
a 7"xa
3
:
I
• £
3
3£
3
3 7J 3Z ~3 a £
3
7c
3
'ac
1 7
c 3
3
3 3 J 'a =Z IJ 'a «Z Iaa
3
J 3a Z Ja 3Z
R Q c E oCU Lc3 Lo OCD co OC D Lc La LC3 LO OC U 1— D 7 C3 D OCD L3 t L o L3 C L O L3 C L O L3 C L O L3 C L O
LL O
LL
. OR^5 G3 °C O C r n z h, O -— r C OC c h <Oo -M C | L z LsM L U- Y I 0
a
_C lD T . _0 C > z
Q i
0 O ?: ^ C C 2r Qi-OC >| 0
h
>s -1 1. 1. _CC > <
Q J
O O : ^T h L— - U c C h c/r -J c: <
H
n C
Q L
I L.c S2
U
-- t ost Cr Q^
Q
Q: - V C CL
Q Q
LC 0 C
Q _
J0 O
oh-
o
o
5 m CQ CQ DC m
cc
IV co LO CJ) CD co co ^_ 1^ |s- CD O T- T- in r^ LO CD CJ)
E CN t- CM O •* tt CM T— t CN CO co ^— t- •t T— CO co
1- L *O ~ c 0v i 0r o d * -L ^O — L T—O CM OLO c 0 o Tir -i C T~M •^ OT— r•^ —t | Os - 0C O r 0» . co d
5> 0 O OT— T 0— OT- 0T- ^ 0 ~ TO— * O — T0— TO— Or r- 0t— ~3 3 OT— T O— T0— 0^ — T0— 0T_
* (- 0» ~3 ~> "3 O —3 3 —3 3 -3 3 -3 5 S3 -3 3 —3 > —3 > —3 } CM —3 > ~3 > —3 > -3 a -3 3 —3 3
Q O C OO c 0 o O 0) v- TT~ — C ""O "C T D — C' NJ Cr Mt Ci Mt • CG M Ct Mt Ct Mt t Ct M t Ct M I CT M) h CM~
(Q 0 ~ ZO 9 cC COO o c• C C* O Oo c cL co o oO c C CC O OD o cr c cs oo o. cc cc oo oo C c c cJ o o o) O c c^1 o o- C t C cM O o C CC •*J O O ) c C0 •sl O0 o- cr T c- f oo - C t C CO OtD L t C CO t O O C C ctO Ot o t C Ct O O c Ct •* Oo O L CCO OO cCt L Ot O o
|
UWA
|
412
22.12. Asset Renewal-Network Development Plans Review
To: Engineer Network Planning From: Primary Assets Engineer
Thank you for the opportunity to comment on your long-term development
plans in the South Fremantle and Cannington Terminal load area 66kV
substations in the light of ageing equipment still in service in those
substations.
Please find a general overview for the requested sites for the time being,
bearing in mind that a more detailed and conclusive review would take
more time.
That work will be organised over the next few months, and another
feedback will be supplied later in the year.
South Fremantle Load Area Substations:
North Fremantle: Most equipment (CBs, CAPs, DES and some VTs) will
require replacement in 8-12 years, with the indoor
switchboard probably requiring replacement of the indoor
oil filled circuit breakers over the next 5 years.
Edmund Street: Some CBs, CTs, DES, SAs and VTs will require
replacement in 7-10 years.
O'Connor: Most equipment (CAP, CBs, CTs, DIS, DES and VTs)
will require replacement in 8-10 years.
APM: Some equipment (CBs, CTs, DIS, SAs and VTs) will
require replacement in 5-8 years.
Myaree: Some equipment (CBs, CTs, DIS, SAs and VTs) will
require replacement in 5-10 years.
Cannington Terminal Load Area Substations:
Victoria Park: Most equipment (CA, CBs, CTs, DIS and VTs) will
require replacement in 5-10 years.
|
UWA
|
414
22.13. Benchmarking ITOMS1999 Report Review
To: GM Networks From: Primary Assets Engineer
BACKGROUND
Western Power (WPC) has participated in the ITOMS 99 benchmarking
survey. The survey has used 1998 data.
INTRODUCTION
UMS has presented WPC with a summary of end results from the above
survey. The results have highlighted the need to analyse the outcome for
some items where the results indicate poor/bad performance by W PC in
comparison to world and A NZ averages.
There are eight such areas, four line areas and four substation areas, and a
summary of the analysis outcome is given below.
SUMMARY
Overhead Lines
132kV Line Maintenance
Maintenance Costs-15 % of costs reported incorrectly (cost of new
replacement poles $300K),
- an abnormal year with 22 % of reported costs being for pole testing
($500K),
- an abnormal year with 12 % of reported costs being for line maintenance
work to catch up with old problems ($250K),
- a reduction of 9 % of 1998 reported costs has now been achieved (from
$400K to 200K) due to installation of leakage current (pollution)
monitors on lines, and using their readings to reduce line washing costs.
Cost Conclusion: The total reported 1998 cost would be reduced by 58 %,
or approx. 2.5 times, bringing it above ITOMS and A NZ averages.
Further savings are expected with implementation of improved
assessment procedures for new work with our current service providers.
Service Levels-5 % of reported fault outages reported incorrectly (actually
forced outages),
|
UWA
|
415
- 44 % of reported fault outages are successful re-close operations, but
performed manually. As their duration is therefore more than 1 minute
they are deemed fault outages and not transient faults. They almost all
originate due to lightning strikes during the summer season,
- 6 % of reported fault outages are caused by others (eg customer trips, car
vs pole, vandalism etc),
- 16 % of all reported fault outages line returned to service with no cause
found,
- 5 % of all reported fault outages caused by external bushfires,
- 24 % only of all reported fault outages caused by actual line equipment
faults.
Service Levels Conclusion: The service level could be reduced by 52 %,
or approx. 2 times, bringing it to E U RO and A NZ averages.
That would be achieved by installation of auto re-close equipment or
construction of overhead wires to prevent impact of lightning strikes on all
lines seasonally affected by lightning during summer season (ie remove the
affect of the fault quickly or avoid the fault altogether).
It is doubtful we would justify the expense against the performance
benefit gained. Some improvements could be made by reviewing the
practice of not re-closing on some lines, as no permanent faults have been
discovered.
220-330kV Line Maintenance
Maintenance Costs-an abnormal year with 45 % of reported costs being for
one cyclone event ($182K),
- a reduction of 25 % of 1998 reported costs has now been achieved (from
$125K to $25K) due to installation of leakage current (pollution)
monitors on lines, and using their readings to reduce line washing costs.
Cost Conclusion: The total reported 1998 cost would be reduced by 70%,
or approx. 3 times, bringing it to the ITOMS best performance.
Service Levels-50 % of reported fault outages (3) are successful re-close
operations, but performed manually. As their duration is therefore more than
1 minute they are deemed fault outages and not transient faults. They almost
all originate due to lightning strikes during the summer season.
- 17 % of reported fault outages (1) are caused by external bushfires,
- 33 % only of all reported fault outages (2) caused by actual line
equipment failure, and these 2 outages were actually the same incident
(HIW micro burst).
|
UWA
|
416
Service Levels Conclusion: Firstly, the extremely small sample size (6)
makes the statistical analysis meaningless and the numbers volatile.
Secondly, the service level could be reduced by 50%, or approx. 2 times,
bringing it to ITOMS and ANZ averages.
That would be achieved by installation of auto re-close equipment or
construction of overhead wires to prevent impact of lightning strikes on all
lines seasonally affected by lightning during summer season.
The abnormal event (microburst) is not expected to re-occur which
will further improve reported service level to ITOMS best performance.
132kV and 220-330kV Line Patrol/Inspection
Maintenance Costs - Best performer.
Cost Conclusion: No comment, as the cost is very low.
Service Levels-Calculated from the above line service levels (number of line
fault outages).
Service Levels Conclusion: Comments from maintenance items above are
applicable here as well. With the changes implemented as indicated, line
patrols and inspection service levels would be at or above the A NZ and
ITOMS averages.
Overall Conclusions: We can expect significant improvement in the
results from the next ITOMS, especially in cost. We may have to accept the
limited lightning performance of our lines.
No major differences or deficiencies have been identified in our
maintenance and asset practices in comparison to the best practice themes
listed by UMS.
SUBSTATIONS
132kV Transformer Maintenance
Maintenance Costs - Best performer.
Cost Conclusion: No comment, as the cost is very low. Some increase in
cost possible with the introduction of the dielectric testing during routine
maintenance, but detecting problems early should reduce repair costs.
Service Levels-24 % of reported fault outages are switching errors,
|
UWA
|
417
- 18 % of reported fault outages are distribution feeder re-close operations
on very close in faults that the LV breaker can not handle,
- 18 % of reported fault outages are caused by other plant failures,
- 40 % of reported fault outages are caused by non-operational reasons
(vandalism, birds, lightning, and carvs pole).
Service Levels Conclusion: The service level could be reduced by 42 %,
or approx. 2 times, bringing it to the A NZ average by eliminating switching
errors and by installing fault distance sensitive o/c relays on feeder circuit
breakers to avoid re-close on very close faults.
Further improvements could be made by reviewing the non-
operational fault outages to determine if better access restrictions, phase
clearances/insulation and lightning protection practices should be
introduced.
220-330kV Circuit Breakers Maintenance
Maintenance Costs-90 % of costs are caused by three units of one type of
breaker, 330kV Brown Boveri.
Cost Conclusion: The cost area remains uncertain even though we are
replacing the three problem units through a capital project. Seven units of
the same type remain in service.
They can have similar type failures at any time and, depending on the
estimated costs, a decision will have to be made on a case by case basis to
repair them or to replace them.
Service Levels-50 % of reported fault outages (2) are 330kV Brown Boveri
type,
- 25 % of reported forced outages (1) is caused by a gas loss of G EC FE2
type,
- 25 % of reported forced outages (1) is caused by a gas loss of A S EA HPL
type.
Service Levels Conclusion: Firstly, the extremely small sample size (4)
makes statistical analysis meaningless and the numbers erratic.
Secondly, the service level could only be significantly reduced from
the 4 faults experienced by removing all outstanding Brown Boveri units
and a full refurbishment of all FE2 and HPL units. This would obviously
incur a very high capital cost to remove a very small number of events.
Our present policy is to monitor the performance of the breakers and
repair the faults as they occur. We will make a decision to replace the
breakers as/if we find their actual costs too high compared to expected life
|
UWA
|
418
cycle costs. Operations have not put a high risk priority on the outages, as
they are happy with available redundancy in the switchyards.
Therefore we do not expect improvement in the service levels for the
next ITOMS.
132-220-330kV Disconnector & Earth Switch Maintenance
Maintenance Costs - Best performer.
Cost Conclusion: No comments, as cost are very low.
Service Levels-20 % of reported fault outages (1) reported incorrectly
(multiple operation),
- 40 % of reported fault outages (2) are on one faulty unit,
- 40 % of outages (2) were reported as forced because a repair crew
happened to be on site and an outage arrange immediately. Otherwise
these would have been done as planned work and not counted in the
ITOMS service levels.
Service Levels Conclusion: Firstly, the extremely small sample size (4)
makes statistical analysis meaningless and the numbers erratic.
The service level would be very difficult to predict in view of a very
few outages. Further inquires will be made with some of the best
performers to confirm that they reported only one or zero outages and their
practices to achieve this.
132-220-330kV Instrument Transformers & Other Equipment
Maintenance
Maintenance Costs - Best performer.
Cost Conclusion: No comment, as the cost is very low. Some increase in
cost possible with the introduction of the dielectric testing during routine
maintenance, but detecting problems early should reduce repair costs.
Service Levels-40 % of reported fault outages (3) reported incorrectly,
- 15 % of reported fault outages (1) due to lightning strike,
- 15 % of reported forced outages (1) due to detected low oil to prevent
explosion,
- 15 % of reported fault outages (1) due to vandalism,
- 15 % of reported fault outages (1) due to plant failure before testing.
|
UWA
|
420
22.14 Review and Instruction to Improve Line Maintenance
To: Manager Maintenance From: Primary Assets Engineer
BACKGROUND
Western Power (WPC) participated in the ITOMS 1999 benchmarking
survey. The survey used 1998 data.
INTRODUCTION
UMS has presented WPC with a summary of end results from the above
survey.
The results have been analysed, and the areas highlighted where there was a
need to further analyse poor/bad performance by W PC in comparison to
world and A NZ averages.
The analysis is presented in the memo to General Manager dated 16 March
2000 (#383416).
One of the items is line maintenance cost, and it has been accepted that the
maintenance policy & instruction defined process for the line maintenance
has not been fully implemented.
SUMMARY
It is now the right time to discuss the best way of to implement, an also
improve if possible, the procedure for line maintenance work with our main
service provider, Line & Cable Maintenance Section, Westpower Services
Branch, by the end of this financial year.
This review should involve Line&Cable Maintenance Section, Maintenance
Management Section, and Primary Assets Section.
CONCLUSION
The Primary Assets Engineer will call an initial meeting to set up scope of
work and working arrangements in early May 2000.
|
UWA
|
421
22.15 Asset Management Model and Process Audit
REVIEW OF WPC ASSET MANAGEMENT PROCESS-PB POWER
Asset management in WPC is described in the following series of
documents:
- Asset Management Policy for Transmission Assets
- Asset Management Process for Transmission Division
- Asset Management Procedures for Transmission Division
- Business Case Analysis for Transmission Division
The Policy identifies that the long-term asset management plans will be
prepared annually based on the following data:
- Age
- Condition service requirement
- Future role
- Consequence and probability of failure
- Environment
- Decay predictions
- Life cycle costs
- Potential for obsolescence
- Business needs
All proposals for major expenditure on assets are prepared in accordance
with an approved process. Critical assets are ranked according to a standard
risk matrix process and expenditure prioritised accordingly.
The asset management process has been in place for a limited time and
therefore the success of the process in terms of the achievement of set
reliability targets and optimisation of maintenance and capital expenditure
cannot be assessed at this time.
The process also does not as yet incorporate the creation of new assets or
the augmentation of existing assets required meeting future load
requirements.
This integration of the asset maintenance/replacement process and the
planning process is to be undertaken in the near future.
|
UWA
|
422
The asset management process combines condition based maintenance and
reliability centred maintenance in order to achieve optimum performance of
the assets at least cost. In accordance with modern practice, time based
maintenance activities are programmed as a result of the consideration of
the condition of the assets and the risk of failure.
Replacement of assets is based on a business case evaluation of the least
cost option including the risk of failure over the life of the asset.
The asset management process also incorporates a review of maintenance
procedures where this is identified as necessary as a result of the detailed
investigation of the condition of a particular asset class or type.
Maintenance periods are lengthened or shortened depending on the specific
requirement or recommendation.
The asset management process described in the WPC documents is
comprehensive and should ensure that expenditure is focussed in the areas
identified as requiring attention.
Provided that the process is followed and adequately monitored and that the
information systems are put in place and resources are made available to
provide the data and reviews that are required, PB Power believe that the
resulting expenditure forecasts will be reasonable.
The asset management model used by WPC in this asset management
process allows objective data on the condition of each group of assets, in
the form of the expected remaining life of that group of assets, to be used to
generate the replacement requirements.
This process is very similar to the process followed by the asset
replacement and capital forecast model of the PB Power
|
UWA
|
423
22.16 Working Group for Substation Maintenance Review
INTRODUCTION
Ever since the Transmission Maintenance Branch has been created, and a
decision to give it full responsibility for the Transmission Division's
maintenance function, it has been our clear understanding that management
of maintenance practices and costs must be our number one priority.
There are a number of reasons for it, but the two main ones will certainly
have to be:
• Our desire to achieve optimum maintenance practice with minimum
costs in order to maximise our contribution to the transmission business.
• With the introduction of an open access regime to the transmission
system, our maintenance practices and costs will come under the close
scrutiny of all future users.
They will have to be convinced that the procedures and associated costs are
achieving the world's best practice results, or that procedures are in place
how to get to the required targets in a reasonable time period.
All this must be subject to the overriding priority of always maintaining the
safety of public and employees.
In order to answer the above challenges, the Maintenance and Power
Services Branches have initiated a review of some areas by a dedicated
working group (WG) of our Maintenance Policy and Maintenance Services
Manuals, as documented below.
Additionally, the WG is to take into consideration obtained results of the
recent U MS ITOMS 1997 benchmarking survey of international companies
involved in the electricity supply industry and the switchgear maintenance.
This report provides the findings and produced recommendations of the
WG on how to further improve the procedures and practices as defined in
the above two Maintenance Manuals.
The review took place over a nine-month period. The items, chosen to be
reviewed first, had been initially assessed as the ones representing maximum
possible savings as far as resources and overall costs were concerned.
|
UWA
|
424
EXECUTIVE SUMMARY
The Working Group finds that:
1. The electrical inspection rounds of metro terminal stations can easily be
incorporated in their regular instrument rounds, already performed by
the PSB Controls Section. Details provided later in the document;
2. There is a possibility to significantly reduce regular intrusive time based
disconnector and earthing switch maintenance activities. Ditto;
3. There is a possibility to reduce the frequency of regular time based
instrument and electrical rounds of zone substations following a
comprehensive testing programme undertaken and completed during the
last three years. Ditto;
4. There is probably no need to continue with manual washing and
cleaning of switchyard insulators when the live switchyard washing is
introduced. Ditto;
5. There is a need to introduce live washing of substations on a regular basis
with the choice of substations and their washing frequency based on the
critical position of the site and pollution type and severity. The need for
washing of other sites can be determined annually, using pollution
assessments during regular site inspection rounds, and installed pollution
monitor readings;
6. Thermographic survey of all our substations needs to be introduced on an
annual basis timed to line up with the particular site maximum loading
period for the high voltage side. For feeders, switching programmes to
ensure maximum feeder current is available during the survey;
7. The Galileo Circuit Breaker Types ORE20, ORE30, and ORE36 could
have their frequency of maintenance increased from every one to every
two years. The maintenance interval triggered by the numbers of faults be
increased from two to five breaker fault operations;
8. The Metro Vickers Circuit Breakers Bulk Oil Type existing maintenance
procedure could be changed by introducing three different levels of
maintenance instead of the existing one level only. From now on only the
lowest level of maintenance will be done regularly every four years or
after a prescribed number of fault operations. The results of that level of
activity will then be used to determine if level two or level three are
required, instead of the current practice to do just level three every time;
|
UWA
|
426
22.17 Plant and Line Failures Report
Introduction
The attached tables present you with the fault statistics for our transmission
primary plant & lines, with a summary of the main issues outlined below.
Primary Plant
There were 151 faults in the period July - December 2001. For main
contributing plant items see page 5:
• Circuit Breakers had 72 faults.
• Current & Voltage Transformers had 15 faults.
• Power Transformers had 31 faults.
• Disconnectors & Earthswitches combined had 7 faults.
• Circuit Breakers with CTs had 21 faults.
• Others had 5 faults (capacitors, surge arresters, reactors and busbars).
There were five fault outages during this period, as described below:
1. Kalamunda Substation-T3 Cct-internal fault in the transformer;
2. Hay Street Substation-Tl Cct-internal flashover in 317.0 circuit breaker;
3. Cannington Terminal-66kV capacitor (Cap72)-loose connection;
4. Cunderdin Substation-T2 Cct-Tap changer failure in the transformer;
5. Margaret River Substation-Tl Cct-W phase fuse failure.
The percentage of failures for the above period is 2.5%, down from 2.93%
in July 2001, and is based on the current primary plant population of 12072.
(The target for percentage of plant failures for 2001/2002 is 2.1%. This target
decreases annually by 0.2%.)
There have been three failures of plant still under warranty during this period:
1. Merredin Substation-22kV CB (516.0)-Gas leak;
2. Picton Substation-22kV CB (503.0B)-Yph insulator broken at base;
3. Hay Street Substation-1 lkV CB (317.0)-Flash over.
|
UWA
|
oo
CN
*sf
-1 o
00 o
o O CM t- CM CO "sf CO IO IO IO
> o o o
X
> o
o > o o o
_i
X
1- 5 o o o o
X
UI
>
T_ o T- V
I
CO - •a
tu
Hi > © o o o>
_i
c
> re
T— CM T" CM CO
Q. I L.
< *• i r- e
5 o ^ T- cu
-Q
C
> o o o re
X u
< o
tu
CO > o o o O)
re
3
> O
T_ o T- T"
CO X c
UI T- 0 roe tU
Q > o © o l_ E i—
o O
••—< a)
10 tu to
> CM o CM CM o (A o CO
C QO X
> CO o CO CO
IO o
<D
'tf El o
03
• t >4— u •
_1 *-» 3 to o
to to o O JD
X > X
>
T- T- T- 0 C0 M C oO 0 C o0 M C oO
CO
V re -o tta
uu
7 • c5 o T tct3
u
u
J
cC t
Z
O u
-J •o cu X3 re •d £ o
> LO o IO IO o is ra > C; L 8tu "^3 a $c
> X IO & =- 3 i c To 3 a > C O 3C L to a c) a to o
> o o o iS -o a (O0 (U T3 Q- o
_l ai <u i_ c tu uW -" 25 Tc3 tu c
>
T- t- •sf o to to
W
CJ )
W3 =3 tt uu toc o to* - O(0 O CT tC uO " t-
u
t o1—o
1- X o •D o £ » S E — o "to 'to o
O O » -n E E DcCa
> _l CM CM o ** •* o w 10 t° u TJ T3 T3 w r e ac )u TP 3 To 3 C ht -Ou
m CD CU tu tu
O Q. o ao . > >
co > X o •s> t o en CA CM 3 C O — a > ~> -< E ctu a > tg u n c> u i •o a it I c— au O E aj O E a ) c o
o > CO o CO CO |s» w 3 t >u ai t— o . iEt o i C O to to L t_ o o
_1 CM CM CM re co cz cz cz L c a * c- z. * c- « tu
> o o 3 3 3 to a. 3 3 T3
h O- X o o o u £ o Q x. t rU 0 r) tU r 3 r e tt uu tu tu t JU Z
C O+ Q > ,- T- °> •• o CM CM CM C oO 5~ 7t zu " _ ^ DC C DC C CC cC cC 4 oQ -) 1
LL LL LL
(0 ©
u_ s Q ±i o z
LL LL £ LL
LL
oo
' >
LL LL
<sf
|
UWA
|
431
FUTURE WORK ASSOCIATED WITH MAJOR DEFECT
CONTRIBUTORS
1. Galileo 22kV Circuit Breakers - Type ORE.20
26 of these units will be replaced under the project 57211 before the end of
2003 and the remaining 28 units will be replaced before the end of 2004
under the project T0034382
2. Alstom 22kV Circuit Breakers - Type GL107
A gas leaking problem and an unsuitable part in the mechanism have been
identified at the early stages. The unsuitable parts are being replaced at the
moment at manufacturer's cost.
3. REYROLLE 66kV Circuit Breakers - Type 66.OSM. 10X1
Three complete spare breaker poles have been retained for replacements in
major incidents. Individual planned CB replacements will continue if the cost
for repairing leaking poles is too high.
4. GEC 132kV Circuit Breakers - Type FG1B
The worst performing breaker will be replaced and refurbished to use as the
replacement for the next worst performing breaker. This process will be
repeated as necessary.
5. Yorkshire 22kVCircuit Breakers - Type YSF6
16 units are recommended for refurbishment.
6. G EC 33kV Circuit Breakers - Type FK1 - SF6
Three of these units will be replaced under project T0028388 in 2001 and
another four units under project T0056700 in 2002. The performance of the
remaining units is to be continuously monitored.
7. Endurance Electric Voltage Transformers - Type DK.SVT.132
30 of these units are being replaced under the project T0032684. The
remaining 21 units will be replaced under the project T0045902 in
July/August 2002.
|
UWA
|
432
22.18 Asset Failures Report
TRANSMISSION ASSETS BRANCH
UNITS DEFECT
REPORT
Sorted by Equipment Type and Defects
Spec No. Spec Order Date Notn. Manufacturer Model Equip. Total Total Total Defects Age
Item Volts Type Order in Defects YTD
No. (kV) Use
99/SS1 10 1/01/1986 66 ASEA RACK CAP 2 2 11 2 14
99/SS1 07A 1/01/1986 22 ASEA RACK CAP 14 14 10 2 14
99/SS1 4 1/01/1986 11 BICC CUBICLE CAP 4 4 9 1 14
2/SS2 1 1/01/1979 66 TYREE RACK CAP 1 1 8 1 21
99/SS1 1 1/01/1986 7 ASEA CUBICLE CAP 12 12 5 14
99/SS1 9 1/01/1986 33 ASEA RACK CAP 2 2 5 14
97 6 1/01/1963 11 DUCON OFT.6411 CAP 3 3 3 1 37
TD6/01/82.1 3 1/01/1983 33 ASEA RACK CAP 3 3 3 17
0
102/SS1 2 1/01/1987 0 ASEA RACK CAP 2 2 2 13
102/SS1 1 2/09/1987 0 ASEA RACK CAP 3 3 2 13
TD6/03/83 2 6/07/1983 22 BICC RACK CAP 3 3 2 17
99/SS1 8 1/01/1986 22 ASEA RACK CAP 12 12 2 14
99/SS1 7 1/01/1986 22 ASEA RACK CAP 2 2 2 14
157/SS1 6 10/05/1989 33 ABB RACK CAP 2 2 2 11
97 7 1/01/1961 7 DUCON OFT.6411 CAP 2 2 39
99/SS1 2 1/01/1986 11 ASEA CUBICLE CAP 2 2 14
TD6/05/81 1 6/11/1981 11 BICC CUBICLE CAP 4 4 19
99/SS1 3 1/01/1986 11 BICC CUBICLE CAP 11 11 14
214/SS1 03A 8/02/1994 22 ABB POWER CAP 1 1 6
TRANS
214/SS1 01A 30/03/1993 22 GEC 5TH CAP 1 1 7
ALSTHOM HARM.FILTER
194/SS1 04A 8/05/1992 22 ABB RACK CAP 1 1 8
194/SS1 02A 12/04/1991 22 ABB RACK CAP 2 2 1 9
157/SS1 3 10/05/1989 22 ABB RACK CAP 2 2 11
10/K 1 1/01/1967 132 DELLE OR2K-12-16 CB 38 13 197 6 33
12/MU 1 1/01/1963 132 DELLE OR2K-12-16 CB 14 2 61 37
13/MU 1 1/01/1963 66 DELLE OR1K-9-14 CB 10 8 54 5 37
11/SS1 1 12/09/1972 330 BROWN FT13/24C1 CB 8 8 45 4 28
BOVERI
011/YYY 1 1/01/1958 22 GALILEO ORE.20 CB 27 25 43 13 42
0107YYY 1 1/01/1958 22 GALILEO ORE.20 CB 31 29 40 9 42
CTS176 1 11/12/1974 33 ASEA HLC36- CB 75 72 35 1 26
52/1250A
TD6/01/82.0 4 1/01/1983 220 GEC FE2 CB 8 8 31 4 17
2
TD6/01/82.0 02B 1/01/1983 132 GEC FG1B CB 10 10 30 1 17
2
116 1 29/09/1966 66 REYROLLE 66.OSM.10X1 CB 36 27 26 1 34
114 1 18/08/1966 22 MAGRINI MMS.24C CB 18 13 24 34
GALILEO
DIS162 1 2/02/1973 132 ASEA HLD145/1250 CB 48 47 21 2 27
B
12/K 01A 1/01/1967 66 ASEA HLC72.5/1250 CB 43 41 17 2 33
A
CTS111 4 1/01/1974 33 MAGRINI 38MGE-750 CB 29 27 16 2 26
GALILEO
DIS138 01A 6/05/1970 132 GALILEO OCERF.145 CB 13 12 16 1 30
|
UWA
|
433
22.19 Asset Performance/Population/Attrition Report
TRANSMISSION ASSETS BRANCH
EQUIPMENT POPULATION / AGE /
DEFECTREPORT
|
Sorted by Equipment Type and Age
Spec No. Spec Order Date Nom. Manufacturer Model Equip Total Total Total Age
Item Volts Type Order in Use Defect
No. (kV) s
97 7 1/01/1961 7 DUCON OFT.6411 CAP 2 2 1 39
97 3 1/01/1961 11 DUCON OF.6412 CAP 2 2 39
97 5 1/01/1961 11 DUCON OFT.6411 CAP 1 1 39
97 1 1/01/1961 11 DUCON OF.6412 CAP 1 1 39
97 6 1/01/1963 11 DUCON OFT.6411 CAP 3 3 3 37
97 4 1/01/1963 22 DUCON OF.6412 CAP 1 1 37
97 2 1/01/1963 22 DUCON OF.6412 CAP 1 1 37
2/SS2 1 1/01/1979 66 TYREE RACK CAP 1 1 8 21
TD6/05/81 1 6/11/1981 11 BICC CUBICLE CAP 4 4 1 19
TD6/05/81 3 6/11/1981 22 BICC RACK CAP 5 5 19
TD6/05/81 2 6/11/1981 22 BICC RACK CAP 3 3 19
TD6/01/82.17 02D 1/01/1983 0 NISSEN CTYPE CAP 12 12 17
DAMPED
TD6/01/82.17 02C 1/01/1983 0 NISSEN CTYPE CAP 12 12 17
DAMPED
TD6/01/82.17 02A 1/01/1983 0 NISSEN SERIES CAP. CAP 12 12 17
TD6/01/82.17 02E 1/01/1983 0 NISSEN 3RD TUNED CAP 12 12 17
TD6/01/82.17 02F 1/01/1983 0 NISSEN 2ND ORDER CAP 12 12 17
DAMPD
TD6/01/82.17 02B 1/01/1983 0 NISSEN BYPASS CAP. CAP 12 12 17
TD6/01/82.17 02G 1/01/1983 0 NISSEN DETUNED CAP 4 4 17
PFC
TD6/03/83 3 6/07/1983 7 BICC CUBICLE CAP 2 2 17
TD6/01/82.10 1 1/01/1983 11 ASEA RACK CAP 4 4 17
TD6/03/83 , 1 6/07/1983 11 BICC CUBICLE CAP 3 3 17
TD6/01/82.10 01A 1/01/1983 11 ASEA RACK CAP 1 1 17
TD6/01/82.10 4 1/01/1983 22 ASEA RACK CAP 4 4 17
TD6/03/83 2 6/07/1983 22 BICC RACK CAP 3 3 2 17
TD6/01/82.10 3 1/01/1983 33 ASEA RACK CAP 3 3 3 17
99/SS1 1 1/01/1986 7 ASEA CUBICLE CAP 12 12 5 14
99/SS1 3 1/01/1986 11 BICC CUBICLE CAP 11 11 1 14
99/SS1 4 1/01/1986 11 BICC CUBICLE CAP 4 4 9 14
99/SS1 2 1/01/1986 11 ASEA CUBICLE CAP 2 2 1 14
99/SS1 11 1/01/1986 11 BICC CUBICLE CAP 1 1 14
99/SS1 07A 1/01/1986 22 ASEA RACK CAP 14 14 10 14
99/SS1 8 1/01/1986 22 ASEA RACK CAP 12 12 2 14
99/SS1 6 1/01/1986 22 ASEA RACK CAP 3 3 14
99/SS1 5 1/01/1986 22 ASEA RACK CAP 2 2 14
99/SS1 7 1/01/1986 22 ASEA RACK CAP 2 2 2 14
99/SS1 9 1/01/1986 33 ASEA RACK CAP 2 2 5 14
99/SS1 10 1/01/1986 66 ASEA RACK CAP 2 2 11 14
102/SS1 1 2/09/1987 0 ASEA RACK CAP 3 3 2 13
102/SS1 2 1/01/1987 0 ASEA RACK CAP 2 2 2 13
102/SS1 1 12/07/1988 0 ASEA RACK CAP 1 1 12
102/SS1 3 6/07/1988 0 ASEA RACK CAP 1 1 12
157/SS1 1 I 10/05/1989 11 ABB CUBICLE CAP 1 1 11
|
UWA
|
434
22.20 Outline of Special Contingency Plan for Indoor Metal Clad
Switchboards
MANAGEMENT OF THE FAILURE OF LOW VOLTAGE
INDOOR SWITCHBOARDS
INTRODUCTION
There are a number of older substations that have indoor bulk oil/compound
filled switchboards. Over a period of years deterioration of the insulation can
lead to failure of these switchboards. In addition the onerous switching duty
required of some switchgear (in particular those on capacitor circuits) can
lead to mechanical failure of the switchgear. Failure of these switchboards
can result in major fire and smoke damage to equipment in the building in
which the switchboard is housed. In the event that a catastrophic failure of a
switchboard occurs there needs to be a plan to restore supply to customers
within an agreed period of time.
Switchboard failures have already occurred at three substations and the
switchgear at a fourth was consider to be in such a poor condition that it was
taken out of service. It is noted that in two of the cases mechanical failure of
the switchgear is strongly suspected. Tests on the switchgear have shown
some to have insulation in a very poor condition. This effectively means that
any of the indoor switchboards can be regarded as being at risk of failure.
This report describes the aims and scope of the project, and provides
recommendations for short and longer-term contingency plans. Detailed
contingency plans will be provided in separate documents.
TERMS OF REFERENCE
The terms of reference for the project team were generally as follows: -
• Identify the "high risk" indoor switchboards/substations;
• Identify substations at risk of rotation for extended periods;
• Agree with Network Services Division the extent of Distribution Transfer
Capacity (DTC) available;
• Develop an outline of a contingency plan for these substations;
• Establish the level of service required by Network Services Division for
supply restoration;
• Review/enhance stocks of strategic spares, includingreclosers;
|
UWA
|
435
• Develop an operational contingency plan to quickly achieve maximum
D TC and restore supply to critical feeders, including use the possible use
of mobile generators;
• Review the benefits of purchasing a mobile substation.
It should be noted that this report does not consider the requirements for
managing the failure of the substation high voltage switchgear nor the failure
of the substation transformers. These are the subjects of separate
contingency/response plans.
In most cases the substation high voltage switchgear and transformers are
outside the switchgear building and the risk of the catastrophic failure of the
low voltage switchboard also involving them is extremely low.
In the case of a substation transformer, if a failure does occur it is expected
that the appropriate Rapid Response Spare Transformers would be mobilised
as part of the response plan.
SUBSTATIONS WITH INDOOR SWITCHGEAR
A list of the substations that have indoor switchgear is given in Appendix.
A program of testing for the switchboards listed in Appendix is currently
being carried out using the initial assigned priority order specified.
The purpose of this program is to provide an ongoing assessment of
the condition of the switchboards. Based on the results of the tests the
switchboard is assigned a condition category that indicates the level of risk of
failure.
The current category assignment based on the tests is indicated in the table.
Where no category has been assigned the switchboard has not been tested.
The tests are based on similar tests carried out by power utilities in Australia
and Europe and are considered to represent world best practice.
The tests are ongoing and the interval to the nex ttest is based on the
assessed condition of the switchboard.
The substations in the first three initial assigned priority groups have been
identified as being at greatest risk of fire and smoke damage due to their bulk
oil and/or compound content in their plant.
|
UWA
|
436
The substations listed in the fourth group have more modern switchboards
(some using SF6 gas), and are considered to have a very low risk of fire and
smoke damage.
The Transmission Assets should continue to monitor the state of the indoor
switchboards and update the assignment of the risk category based on the
results of the ongoing tests.
Note that it is not intended to make provision for switchboard failure at
Central Business District substations, as these are designed with design
criteria that allow for the complete loss of supply to the substation.
LEVEL OF SERVICE
The required time for full restoration of supply is critical to the extent of the
contingency plan and to the level of resources needed.
The requirements for restoration of supply will have a major bearing
on the manner in which restoration is carried out and the resources needed.
A Service Level Agreement has been signed between the Marketing and
Sales Division and Transmission Division.
Amongst other items this agreement requires that, in the event of a
catastrophic failure of a switchboard, full supply must be restored to the
substation within two weeks.
However, this report focuses on strategies designed to meet two-week,
one-week and one-day restoration times.
SHORT TERM ACTION
6.1 Catastrophic Failure
Action must be taken immediately the catastrophic failure of a switchboard
occurs to restore supply. The most immediate action that can be taken is to
utilise the Distribution Transfer Capacity (DTC) between the substation
concerned and adjacent substations.
The Network Operations Controllers must take this action as soon as
the System Operations Controllers confirm the switchboard failure. It is
expected that the utilisation of D TC will take up to 2 hours to implement.
The available D TC between substations may be limited by the loading
of the transformers and feeders at the adjacent substations. Where this is the
case it may be possible to provide additional D TC by offloading adjacent
substations using the D TC between them and other substations adjacent to
|
UWA
|
437
them. It is expected that the Network Operations will be able to provide any
additional D TC within one day.
Where the available DTC cannot match the expected load demand, it is
expected that the Network Operations will manage the situation by
implementing planned rotation of load together with other means of reducing
load.
Due to the way substations are now being operated it is expected that the
levels of D TC available will reduce. The new Normal Cyclic Rating (NCR)
criteria allows the transformers at substations to be loaded up to about 9 0%
of the Long Term Emergency Rating (LTER).
The consequence of this is that as more substations reach this limit
both the amount of D TC available and the ability to offload to adjacent
substations will be limited by the loading of the transformers and not by the
distribution network.
A number of the substations listed in Appendix have DTC limited due to the
fact that the distribution voltage level is not the same as that used in adjacent
substations. In some cases there is no D TC available.
6.2 High Risk Switchboards
If switchboard testing reveals that a substation switchboard is at risk of
failure it will be necessary to have a contingency plan to deal with any failure
in the short term.
When a switchboard has been identified as being at high risk of failure,
System Operations will liaise with the Network Operations and other
Transmission areas and a decision will be made on what action is to be taken
to reduce the impact of a failure.
It is expected that in these cases temporary arrangements will be put in place
to enable the switchboard to be removed from service and in due course
replaced by a new switchboard.
This is expected to be a planned process not requiring immediate
action.
|
UWA
|
438
R E C O M M E N D A T I O NS
It is recommended that:
- A contingency plan for the catastrophic failure of LV indoor
switchboards is prepared based on the use of an emergency switchboard
consisting of a pre-built, trailer mounted, metal frame complete with
reclosers, auxiliary transformers and protection, control and
communications facilities.
Action: Manager Engineering
- The Transmission Engineering proceed with the design of an emergency
switchboard consisting of a pre-built, trailer mounted, metal frame
complete with reclosers, auxiliary transformers and protection, control
and communications facilities.
Action: Manager Engineering
- The Transmission Assets should continue to monitor the state of the
indoor switchboards and update the assignment of the risk category based
on the results of the ongoing tests.
Action: Manager Transmission Assets
- The Substation Design should review the layout of each of the critical
substations included in Appendix to confirm that there is sufficient space
for the emergency switchboard.
Action: Manager Engineering
- The Protection Design should review the protection arrangements at the
remote ends of all transmission lines supplying substations that have
indoor switchboards and determine the most suitable arrangements for
providing temporary radial protection for inclusion in the contingency
plan.
Action: Manager Engineering
|
UWA
|
DECLARATION FOR THESIS CONTAINING PUBLISHED WORK AND/OR WORK
PREPARED FOR PUBLICATION
This thesis contains published work and/or work prepared for publication, which has been co-
authored. The bibliographical details of the work and where it appears in the thesis are outlined
below.
Yimiao Huang, Guowei Ma, Jingde Li, Hong Hao. (2015). Confidence-based quantitative risk
analysis for offshore accidental hydrocarbon release events. Journal of Loss Prevention in the
Process Industries. 35, 117-124. (Chapter 6)
The estimated percentage contribution of the candidate is 60%.
Yimiao Huang, Guowei Ma, Jingde Li. (2017). Multi-level explosion risk analysis (MLERA) for
accidental gas explosion events in super-large FLNG facilities. Journal of Loss Prevention in the
Process Industries. 45, 242-254. (Chapter 5)
The estimated percentage contribution of the candidate is 65%.
Yimiao Huang, Guowei Ma, Jingde Li. Grid-based Risk Mapping for Gas Explosion Accidents
by Using Bayesian Network Method, Journal of Loss Prevention in the Process Industries, 48,
223-232. (Chapter 4)
The estimated percentage contribution of the candidate is 70%.
Guowei Ma, Yimiao Huang. A Risk-based Human Safety Assessment Method for Explosion
Accidents at Petrol Stations. Submitted to Safety Science. (Chapter 3)
The estimated percentage contribution of the candidate is 60%.
Yimiao Huang, Guowei Ma. Quantitative Risk Analysis method for Domino Effects- induced Fire
Accidents at Petrol Stations. Submitted to Fire Safety Journal. (Chapter 3)
The estimated percentage contribution of the candidate is 70%.
I
|
UWA
|
ABSTRACT
In the oil and gas industry, a vapor cloud explosion (VCE) induced by oil and gas leaks is one of
the major threats to process safety because it may lead to catastrophic consequences. Risk
evaluation of such gas explosions is complicated when multiple factors (e.g., leak severity, vent
condition, structural complexity), multiple consequences (e.g., building damage, human loss,
environmental effect), and complex interrelationships need to be considered. Meanwhile, for
various process facility types such as small petrol stations, large onshore process factories, highly
congested offshore platforms, and others, specific site characteristics also need to be involved in
explosion risk analyses. However, most traditional risk analysis methods are too simple to reveal
the complex mechanism of gas explosions while some advanced computational methods are too
time-consuming. Therefore, this research carries out a comprehensive study on the gas explosion
risk analysis approaches of different oil and gas facilities and aims at developing more accurate,
detailed, efficient, or reliable risk analysis methods for VCEs under different conditions.
First, to enable a more accurate explosion risk analysis when multiple factors and multiple
consequences are involved, an advanced Bayesian network–based quantitative explosion risk
analysis method is proposed to model risks from an initial release to vapor cloud explosions and
further consequences because of the ability of the Bayesian network to reveal complicated
mechanisms with complex interrelationships between parameters. Meanwhile, since fire accidents
frequently occur at process facilities and may also result in significant consequences, a risk analysis
of fire accidents using the Bayesian network is conducted as well.
Second, for a risk analysis of gas explosion in process facilities close to residential areas, a grid-
based risk-mapping method is developed to provide a more detailed explosion risk analysis for
large areas under complicated circumstances. A target area is divided into a number of grids of
appropriate size and with simplified conditions, and risk analysis is conducted at each grid.
Compared with traditional explosion risk analysis, the proposed grid-based method simplifies
complicated conditions throughout the gridding process, and therefore, a more precise risk analysis
is enabled.
Third, a multi-level explosion risk analysis method is established to offer a more efficient explosion
risk assessment of super-large oil and gas facilities with highly congested environments such as
floating liquefied natural gas (FLNG) platforms. When computational fluid dynamic (CFD)
software is used to calculate overpressures, a large amount of computational time is required for
such structures because of its enormous size and highly complicated topside structures. The
II
|
UWA
|
ACKNOWLEDGEMENTS
I would like to extend my fondest appreciations and thanks to the following people for their
advice and support, both technical and emotional, throughout the development of this research
project.
First and foremost, I would like to express my heartfelt appreciation and respect to my
supervisor, Professor Guowei Ma for his invaluable guidance, strong support, big patience and
encouragement over the years of my doctoral study. His rigorous attitude and eternal
enthusiasm in research have imposed a profound influence on me, and will benefit me in my
future career and life.
I am also grateful to Professor Hong Hao from Curtin University for sharing his academic
experience and rewarding advice. Thank another team member Jingde Li for sharing the
experience and knowledge of the numerical modelling.
I would like to express my special appreciation to my wife Sika Fang for her unconditional
love and support. Her companionship during the PhD study fuelled me with motivation and
confidence. I could never achieve my goal without her support.
I also would like to thank my parents overseas for their continual support, encouragement and
understating during the process of performing the tasks required for the research undertaken. I
could not have performed well without their continuous support.
Finally yet importantly, I greatly appreciate the financial support provided by Australian
Government Research Training Program (RTP) Scholarship, Australian Postgraduate Award,
UWA Top-up Scholarship and University Travel Award.
Yimiao Huang
March 2017
VI
|
UWA
|
PUBLICATIONS ARISING FROM THIS THESIS
Journal papers
1. Yimiao Huang, Guowei Ma, Jingde Li, Hong Hao. (2015). Confidence-based
quantitative risk analysis for offshore accidental hydrocarbon release events. Journal
of Loss Prevention in the Process Industries. 35, 117-124.
2. Yimiao Huang, Guowei Ma, Jingde Li. (2017). Multi-level explosion risk analysis
(MLERA) for accidental gas explosion events in super-large FLNG facilities. Journal
of Loss Prevention in the Process Industries. 45, 242-254.
3. Yimiao Huang, Guowei Ma, Jingde Li. Grid-based Risk Mapping for Gas Explosion
Accidents by Using Bayesian Network Method, Journal of Loss Prevention in the
Process Industries, 48, 223-232.
4. Guowei Ma, Yimiao Huang. A Bayesian Network based Quantitative Risk Analysis
method for Explosion Accidents at Petrol Stations. Submitted to Safety Science.
5. Yimiao Huang, Guowei Ma. Quantitative Risk Analysis method for Domino Effects
induced Fire Accidents at Petrol Stations. Submitted to Fire Safety Journal.
6. Jingde Li, Guowei Ma, Madhat Abdel-jawad, Yimiao Huang. (2016). Gas dispersion
risk analysis of safety gap effect on the innovating FLNG vessel with a cylindrical
platform. Journal of Loss Prevention in the Process Industries. 40, 304-316.
7. Jingde Li, Guowei Ma, Hong Hao, Yimiao Huang. (2016). Gas explosion analysis of
safety gap effect on the innovating FLNG vessel with a cylindrical platform. Journal
of Loss Prevention in the Process Industries. 44, 263-274.
8. Jingde Li, Guowei Ma, Hong Hao, Yimiao Huang. (2017). Optimal blast wall layout
design to mitigate gas dispersion and explosion on a cylindrical FLNG platform.
Journal of Loss Prevention in the Process Industries.
9. Jingde Li, Hong Hao, Yimiao Huang. Internal and external pressure prediction of
vented gas explosion in large rooms by using analytical and CFD methods. Submitted
to Journal of Loss Prevention in the Process Industries.
VIII
|
UWA
|
The Piper Alpha explosion caused the deaths of 167 men with only 61 survivors (Hendershot,
2013). As one of the worst offshore oil disasters in history, this accident ceased approximately
10% of North Sea oil and gas production during the 1980s, and the economic loss was more
than $3 billion (Patecornell, 1993). Another catastrophic fire and explosion event occurred at
a petrol station as a result of a flood incident in Accra, Ghana, on June 4, 2015, and caused
over 152 fatalities (Asumadu-Sarkodie et al., 2015). For the Guadalajara explosion, 206 people
were killed, nearly 500 to 600 were missing, and 1,800 were injured (ARIA, 2007). The
estimated monetary damage ranges between $300 million and $1 billion. On July 31, 2014, a
series of gas explosions occurred in Kaohsiung, Taiwan, which caused 32 fatalities and 321
injuries. More than four main roads with a total length of approximately 6 kilometers were
damaged, and traffic was blocked for several months (Liaw, 2016).
Vapor cloud explosions (VCE) are one of the most serious hazards to occur on oil and gas
facilities. VCE is defined as “an explosion resulting from an ignition of a premixed cloud of
flammable vapor, gas or spray with air, in which flames accelerate to sufficiently high
velocities to produce significant overpressure” (Mercx & van den Berg, 2005). To evaluate the
risks of vapor cloud explosions, it is essential to consider a large array of random variables,
such as gas property, wind conditions, congestion scenarios, and others. Meanwhile, the risk
analysis of VCEs can be more complicated when multiple consequences such as structural
damage, human loss, environmental effects, and so on, need to be considered. Furthermore, for
various types of process facilities, different environments and structures will be presented, and
therefore, appropriate methods should be applied and specific site characteristics should be
taken into account to ensure the reliability and efficiency of explosion risk assessments.
There are empirical and numerical methods used to quantify the overpressures of explosions.
For example, the empirical TNT-equivalency method (Baker et al. 2012) has proved to be a
popular overpressure calculation tool in the past. A multienergy method (Vandenberg, 1985),
which is based on experimental database for approximating explosion overpressures, has
become a fast and more reliable prediction solution. Later, advanced computational fluid
dynamic (CFD) programs are able to use the three-dimensional Cartesian Navier-Stokes flow
solver to predict cloud propagation, ignition probabilities, leak probabilities, and overpressures.
By considering more fundamental physics, such as the complicated congestion and
confinement in the geometry, the CFD-based approach provides more accurate overpressure
results than empirical methods. However, when CFD analysis is conducted in a large process
14
|
UWA
|
facility with complicated environments and high congestions, an unacceptable computational
force and calculation time may be required.
So far in the oil and gas industry, a wide range of risk analysis methods, from qualitative to
quantitative, are available. For qualitative methods, the analysis is normally done by team
studies based on the generic experience of knowledgeable personnel and does not involve any
mathematical estimation. Overall qualitative evaluations are essentially checklist reviews in
which questions or process parameters are used to prompt discussions of the process design
and operations that would develop into an incident scenario due to an identified risk. The
qualitative risk analysis of process safety involves HAZOP (CIA, 1992; CCPS, 2011; Kletz,
1999), what-if analysis (CCPS, 2011), preliminary hazard analysis (PHA) (Vincoli, 2006),
layers of protection analysis (LOPA) (Summers, 2003), and the like. Qualitative risk analysis
is an effective tool for risk screening or routine inspections. However, if detailed risk
assessment is required, such methods may not be appropriate.
Quantitative risk analyses (QRA) are mathematical estimations based on historical data to
estimate the probability of failure or to predict the consequence of an event or incident. The
event tree is the most used logical model to mathematically and graphically describe the
combination of failures of event and circumstances in an incident sequence, expressed in an
annual estimation. Fault tree analysis (FTA) is another important quantitative risk analysis
method. It is a deductive method that focuses on one particular incident, often called a top
event, and then constructs a logic diagram of all conceivable event sequences that could lead
to that incident. It is normally a logic model that mathematically and graphically depicts various
combinations of equipment faults, failures, and human errors that could lead to an incident of
interest, presented in an annual express. Quantitative risk analysis methods are widely applied
to explosion risk analysis of oil and gas facilities. They normally focus on macroscale
evaluation, which provides an overall statistical result of risk for a target area, such as a fatality
accident rate (FAR) and potential loss of life (PLL) for human loss (Vinnem, 2014). However,
for risk analysis of a large area under complex circumstances, it is difficult for such macroscale
analysis to consider all specific local details and deal with complicated conditions.
To conclude, qualitative risk analysis is widely used to assess the explosion risks of process
facilities, and it is the most efficient method when a rough assessment of explosion risks is
accepted. Traditional quantitative analysis methods such as the event tree and the fault tree has
15
|
UWA
|
also been increasingly used to evaluate the risks in the oil and gas industry. However,
conventional QRAs have difficulty revealing the complicated mechanisms of interrelationships
between risk factors because they only have simple Boolean functions and sequentially
dependent failures. Therefore, an advanced quantitative analysis method such as Bayesian
network (BN) modelling is implemented for a more accurate risk assessment of explosion
accidents when multiple consequences and complex interrelationships are required to be
considered.
Meanwhile, in this research, a multi-level explosion risk analysis (ERA) method is developed
to improve the efficiency of traditional CFD-based explosion risk analysis of super-large
offshore structures such as floating liquefied natural gas (FLNG) facilities, as one critical issue
in applying such CFD-based ERA to FLNGs is that an unacceptable computational time is
normally required because of the large size and complex structures of FLNGs. Moreover, a
grid-based method is proposed to conduct a detailed explosion risk analysis when the process
factory is located close to residential areas, and detailed risk influences on both industrial and
residential areas need to be considered. Last but not least, during the quantitative risk analysis
process of explosion accidents at process facilities, subjective judgment–related uncertainties
are unavoidable. Therefore, in the present study, a fuzzy set theory–based confidence level
method is proposed to deal with the uncertainties in accordance with experts’ subjective
judgments by incorporating confidence levels into the traditional QRA framework.
1.2 OBJECTIVE OF THE STUDY
The aims of the present study include:
Developing a more accurate Bayesian network–based quantitative risk analysis method to
model complicated mechanisms of explosion accidents and further consequences such as
building damages and human losses occurring at oil and gas facilities because of the ability
of BN to deal with complex interrelationships between risk factors.
Developing a more detailed grid-based risk mapping method to enable a detailed explosion
risk analysis for large areas with complicated environments involving not only process
factories but also residential areas; a target area is divided into a number of grids of
appropriate size and with simplified conditions, and risk analysis is conducted at each grid,
from which total risk mapping can be depicted.
16
|
UWA
|
Establishing a more efficient multi-level explosion risk analysis (MLERA) method for
super-large oil and gas facilities with complicated structures, which divides the whole
structure into subsections and applies detailed CFD calculations only to the areas with the
highest risk levels so that the computational time can be reduced to a realistic and acceptable
level.
Developing a more reliable confidence level–based risk analysis method to reduce the
uncertainties of QRAs caused by subjective judgments and consequently to provide a more
robust evaluation of explosion risks of oil and gas facilities; the proposed method applies
fuzzy set theory in dealing with the uncertainties.
1.3 THESIS ORGANISATION
This thesis is composed of seven chapters. Six chapters following the introductory chapter are
arranged as follows. Chapter 3 to Chapter 5 are discussed by different level of structural
complexity of the target sites from petrol station to super complicated offshore FNLG
platforms. For different kind of oil and gas facilities, we suggested different kind of risk
evaluation methods based on their characteristics.
Chapter 2 presents a broad literature review on the state-of-the-art explosion risk analysis
methods including both qualitative and quantitative approaches. Meanwhile, factors
influencing explosion loads and overpressure calculation methods regarding gas explosions are
also reviewed.
In Chapter 3, a BN-based QRA is developed to model explosion risks of petrol stations from
initial release to consequent explosions and human losses because service stations are normally
located close to largely populated residential areas and explosion accidents may lead to
significant human losses. Meanwhile, fire accidents, which occur more frequently than
explosions at petrol stations, may also cause severe consequences. Therefore, a risk analysis of
fire accidents based on BN is also conducted.
Chapter 4 presents a grid-based mapping method used to assess explosion risks of a refinery
factory close to residential areas. A BN model is implemented to consider multiple
consequences and the complex interrelationships between consequences and basic factors. A
mesh convergence of different grid sizes is conducted to determine an optimal balance between
accuracy and computational time.
17
|
UWA
|
CHAPTER 2. LITERATURE REVIEW
2.1 OVERVIEW
This chapter presents a literature review on the existing explosion risk analysis methods
including both qualitative and quantitative approaches. In the meantime, some factors affecting
the severity of the gas explosion loads are introduced to enable a better understanding of the
risk analysis. Moreover, empirical and numerical methods used to estimate explosion loads are
also reviewed. The literature review covers: 1) introduction of factors affecting gas explosion
loads; 2) existing and new approaches in gas explosion loads estimation and prediction; 3)
discussion of current risk analysis methods of gas explosion accidents.
2.2 MAIN FACTORS AFFECTING GAS EXPLOSION LOADS
A gas explosion is defined as a process where combustion of a premixed flammable fuel-air
cloud which is causing rapid increase of pressure. Figure 2.1 illustrates the process from gas
releases to gas explosions. Both release and ignition must be present to result in fire or
explosion accidents. (Bjerketvedt et al., 1997).
No ignition
Immediate
Fire
Ignition
No damage
Release of Gas
and/or Liquid
Damage to
personnel
Formation of and materials
Ignition
Combustible Gas explosion
(delayed)
Fuel-Air Cloud
Fire
Fire
and
BLEVE
Figure 2.1 Typical process of gas explosion (Bjerketvedt et al., 1997)
19
|
UWA
|
The severities of overpressures caused by gas explosions depend on various factors and several
important ones are briefly reviewed here:
The confinement and venting surrounding the gas cloud;
The congestion or obstacles within the cloud;
The fuel properties and concentrations;
The ignition type and location;
Gas explosions can occur in confined areas, such as tanks, pipes or channels, partly confined
offshore modules or buildings and unconfined process plants or other open areas. In a confined
situation, a high flame velocity is not required to generate pressure and the turbulent
combustion process causes a more dramatic increase in overpressure. For example, the
overpressures and impulses of explosions in a confined chamber can be enlarged by 2-3 times
if the confinement volume increases (Kuhl & Reichenbach, 2009). Therefore, it is of great
importance to investigate the flame propagation for reliable design of structures in such
confined explosions (Sauvan et al., 2012; Shi et al., 2009; Tang et al., 2014). Partly confined
explosions occur in structures that are partly open such as offshore modules or the production
or process areas within buildings. In this situation, the size and location of the vent area play
significant roles in building overpressures. Attentions have been increasingly paid to the
partially confined overpressures. (Pedersen et al., 2013; Woolley et al., 2013). If the cloud is
truly unconfined and unobstructed, the flame is not likely to accelerate to velocities of more
than 20 – 25 m/s, and the over pressure can be neglected. In this case, the explosion normally
turns to be a flash fires (Bjerketvedt et al., 1997).
Obstacles is another critical factor that may cause significant influence on gas explosion loads.
The expansion-generated flow created by combustion will generate turbulence when the fluid
flows past the obstacles (Dorofeev, 2007; Kim et al., 2014; Na'inna et al., 2013). The newly
generated turbulence will increase the burning velocity by expanding the flame area and
increasing the molecular diffusion and conduction processes, and consequently increase the
expansion flow and boost the turbulence. The generation of increasingly higher burning
velocities and overpressures in this cycle due to the obstacles is called Schelkchkin mechanism
(Lea, 2002). Figure 2.2 shows the Schelkchkin mechanism of flame acceleration caused by
obstructions constitutes a strong positive feedback loop (Bjerketvedt et al., 1997).
20
|
UWA
|
Increased
pressure
Combustion
Flow interacts Turbulence is
of premixed Expansion
with obstacles generated
gas cloud
Turbulence
enhances the
combustion
Figure 2.2 Positive flame acceleration feedback loop due to turbulence (Bjerketvedt et
al., 1997)
A blockage ratio of obstacles is an important factor that influences the flame propagation and
the explosion overpressures (Oh et al., 2001). The blockage ratio of obstacle is used to describe
the degree of obstruction. Generally, the maximum overpressure increases when blockage ratio
increases. However, the rate of increase depends on the geometry of obstruction (Ibrahim &
Masri, 2001). Normally, higher overpressures will be produced with smaller diameter objects
if the blockage ratio is given. Moreover, more tortuous route (flame in baffle-type obstacles)
results in greater explosion overpressures compared to round obstacles. This occurs due to the
fact that the turbulence enhancement of the burning velocity is higher in the shear layer of the
sharp obstacle (Bjorkhaug, 1986).
The fuel properties strongly affect the flame speeds for stoichiometric fuel–air mixtures
(Dorofeev, 2011). Acetylene, ethylene-oxide and ethylene are most likely to cause significant
overpressures (Dorofeev et al., 1994; Matsui & Lee, 1979). For other fuels, such as butane and
propane, a strong deflagration is required to initiate the detonation (Bjerketvedt et al., 1997).
Moreover, the detonation triggered by methane could be more complicated (Boni et al., 1978;
Wolański et al., 1981). Meanwhile, the fuel concentration also has significant effects on the
21
|
UWA
|
flame region distribution and the explosion behaviours (Halter et al., 2005; Q. J. Ma et al.,
2015). A premixed gas cloud below the lower flammability limit (LFL) and above the upper
flammability limit (UFL) will not be burnt, and very low burning rate will be produced if the
fuel mixtures near the flammability limits. The maximum explosion overpressure is normally
generated within stoichiometric composition or slightly rich premixed gas cloud (Bjerketvedt
et al., 1997).
Additionally, the consequences of gas explosions can be significantly influenced by types of
ignition sources. If the ignition sources are jet-type, or bang-box-type other than a planar or
point source, high explosion overpressures can be observed. In addition, the location of ignition
is another critical factor affecting VCEs. The maximum overpressure can be considerably
increased by an order of magnitude if the location of ignition is placed at less vented or more
confined areas (Babrauskas, 2003; Bartknecht, 2012). Meanwhile, edge ignition may produce
greater overpressures than central ignition as the edge ignition has longer flame propagation
distance for flame acceleration (Zeldovich & Barenblatt, 1959).
Gas explosions are very sensitive to these parameters mentioned above, and therefore, it is
important to carefully consider those factors when risk analysis of a gas explosion is conducted
for any oil and gas facility.
2.3 LOAD PREDICTION METHODOLOGIES
Prediction of gas explosion overpressures is the most important part of explosion risk analysis
as the severities of the other consequences such as building damage, human losses, business
losses etc. are mainly dependent on blast loads. Therefore, some of the most widely used
deterministic explosion load prediction methods from the traditional ones to the most state-of-
the-art ones are reviewed. The explosion risk analysis can be conducted by combining risk
analysis methods and detailed load prediction approaches.
2.3.1 Empirical models in DNV PHAST
DNV PHAST (DNV GL, 2016) is a leading consequence analysis tool using empirical models
to simulate hazardous gas releases, gas dispersions, fires, and explosions. It is fairly simple and
is typically used as a screening tool for rapid indication of physical effects and consequences.
The empirical models used by PHAST include simplified TNT equivalency model (Mannan,
2012), TNO Multi-energy model (Vandenberg, 1985) and Baker-Strehlow-Tang model (Baker
22
|
UWA
|
et al., 1998). These models significantly simplified the physics based on correlations derived
from assessments of experimental data.
The TNT equivalency method is based on the assumption of equivalency between the
flammable material and TNT. It has been extensively studied by Baker et al. (2012), Mannan
(2012) and Stull (1977). A yield factor plays a critical role in converting the energy of gas
explosion into the same explosive charge of TNT. The relationship between the gas explosion
and TNT explosion is seen below:
𝑊 = 10𝜂𝑊
𝑇𝑁𝑇 𝐻𝐶 (2.1)
where 𝑊 is the equivalent mass of TNT, 𝜂 is an empirical explosion efficiency, 𝑊 is the
𝑇𝑁𝑇 𝐻𝐶
mass of hydrocarbon.
The TNT equivalency has been widely used in the simplified models. For example, the Health
and Safety Executive (HSE) evaluated the TNT Equivalence method for both near and far field
range of explosives and energetic materials in a simplistic way (Formby & Wharton, 1996).
Rui et al. (2002) used the TNT equivalency method to evaluate the distributed blast of fuel-air
detonation. Skacel et al. (2013) also discussed the applicability of the TNT equivalency method
for calculations of blast wave characteristics after vessel rupture in 1-D geometry detonation.
Another empirical approach is the TNO Multi-energy method (Vandenberg, 1985). It is based
on the assumption that only confined or congested gas clouds contribute to the overpressure
built-up, and the flame velocity is assumed to be constant for the explosion when the gas cloud
is ignited from centre. Two parameters are vital in the overpressure calculation. Firstly, the
combustion-energy scaled distance R , should be determined, it is defined as:
ce
𝑅
0
𝑅 =
𝑐𝑒
𝐸 (2.2)
3
√
𝑃
0
where E is the combustion energy, 𝑅 is the distance from the explosion centre to the target,
0
and 𝑃 is the atmospheric pressure.
0
The second important parameter is the strength of the explosion, which is classified from a
number between 1 and 10 to represent the level of explosion, as seen on the left hand side in
Figure 2.3. The choice of the explosion strength level from 1 to 10 depends on a conservative
23
|
UWA
|
assumption or other simulations. Therefore, the accuracy of Multi-energy method is highly
dependent on subjective judgments.
Figure 2.3 Hemispherical fuel-air charge blast for the multi-energy method
(Bjerketvedt et al., 1997)
The simplicity and fast estimation speed of the TNO multi-energy method was proved by
Alonso, et al. (Alonso et al., 2006; Alonso et al., 2008). This fast method is used to evaluate
the characteristic overpressure-impulse-distance curves and to analyse the consequence of
damage to humans from VCEs. Pitblado et al (2014) also applied the TNO multi-energy model
in predicting overpressures of the facility siting hazard distance of VCEs. In addition, an
explicit implementation guidance was proposed in their research to improve the consistency in
TNO MEM blast load predictions.
Moreover, Baker, et al. (1994) developed the Baker-Strehlow-Tang model for estimating the
overpressure of VCEs. The calculation of this method relies on finding appropriate Mach
number (M) by assessing fuel reactivity, flame speed and confinement. Then, the
f
overpressures can be determined by reading a range of curves as shown in Figure 2.4 based on
the value of M. This model was further revised and extended by a new set of blast curves from
f
VCE calculations (Baker et al., 1998; Tang & Baker, 1999).
24
|
UWA
|
Figure 2.4 Scaled peak side-on overpressure vs. scaled distance for use in BST model
(Tang & Baker, 1999)
Overall, the Baker-Strehlow-Tang method provides conservative prediction of flame speed,
and it considers some geometrical factors such as the confinement. However, the flame speed
estimation may not be conservative for the unconfined 3D flame expansion scenarios.
Therefore, Pierorazio, et al. (2005) updated the flame speed table in the Baker–Strehlow-Tang
methodology for these exceptional cases and Worthington, et al. (2009) provided a correction
method to the Baker-Strehlow-Tang model for the ground effect of vapour cloud explosions.
2.3.2 Computational Fluid Dynamic (CFD) Analysis
The numerical approaches use the Computational Fluid Dynamic (CFD) codes. The
fundamental partial differential equations, which govern the fluid flow and other explosion
processes, are employed in most of the numerical models during the calculation of VCEs.
When comparing the numerical models with empirical models, the numerical models offer
greater accuracy and flexibility, and by discretising the solution domain in both space and time,
a wide range of geometrical arrangements and conditions in the VCEs can be considered in the
numerical simulations.
One of the most popular numerical model is the FLame ACceleration Simulator (FLACS) code.
The FLACS code has been developed for over two decades at Christian Michelsen Research
Institute in Norway (Bjerketvedt et al., 1997). It has been widely used in the onshore/offshore
25
|
UWA
|
explosion analysis, and now extensive validations have been accumulated (Bakke & Hjertager,
1986; Hansen, et al., 2010; Hjertager et al., 1988; Middha & Hansen, 2009). Ma et al. (2014)
applied the FLACS to predicting the consequences of explosion events at large-scale oil and
gas facilities with complicated and highly congested environments. Li et al. (2016a; 2016b)
used FLACS to evaluate safety gap effect on both gas dispersion and explosion risk for super-
large and highly congested offshore floating liquefied natural gas (FLNG) platforms. Gavelli,
et al. (2011) applied FLACS to evaluate the consequences of the ignition of a flammable vapour
cloud from an LNG spill during the LNG carrier offloading process. Middha, et al. (2011)
analysed the safety benefits of hythane by using FLACS regarding the flame speeds and
flammability limits. Moreover, Bakke et al. (2010) carried out a study on the effect of trees on
gas explosion and Yet-Pole, et al. (2009) employed FLACS to evaluate the possible hazards of
different worst-case scenarios within a naphtha-cracking plant.
Another CFD coded model still under development is the EXSIM model which was initially
created at Telemark Institute of Technology and Telemark Technological R&D centre in 1989
(Hjertager et al., 1992). The EXSIM code is similar to FLACS in the aspect of numerical
modelling, namely, the Cartesian grid and finite volume code are used to represent small-scale
objects in EXSIM (Hjertager, 1997). Therefore, some previously investigated projects for
FLACS had also been validated by EXSIM. For example, the Buncefield explosion was
investigated by both FLACS and EXSIM for explosion simulations (Taveau, 2012). Moreover,
the detailed flame behaviour assessment can be found in the work of Johnson et al. (2010).
Saeter (1998) conducted modelling and simulation of gas explosion in some other complex
geometries. Høiset et al. (2000) implemented the EXSIM model in the investigation of
Flixborough accident.
The third reviewed finite volume computational code for fluid dynamics is the AutoReaGas
model. It is developed by TNO – Prins Maurits Laboratory and allows a detailed simulation of
different aspects of gas explosion (Van Den Berg et al., 1995). The AutoReaGas model
integrates features of the REAGAS and BLAST codes as solvers to deal with gas explosion
and blast waves respectively. The gas explosion solver in AutoReaGas uses the Flux Corrected
Transport technique to cope with the blast wave propagation by applying the 3D Euler
equations (Hjertager & Solberg, 1999). A range of user experience of AutoReaGas also exists
in the industry. For instance, Janovsky et al. (2006) performed the computational simulations
of vented confined explosions by using AutoReaGas and compared the CFD results with the
26
|
UWA
|
Stramberk experimental data. Pang et al. (2014) used AutoReaGas to carry out numerical
simulations of a series of methane–air explosion processes in a full-scale coal tunnel, and the
flame propagation mechanism beyond the initial premixed methane–air region was analysed
by comparing the numerical and experimental results. Jiang et al. (2012) provided a theoretical
guidance for gas explosion disaster relief and treatment in underground coal mines by
performing the AutoReaGas simulation, the propagation characteristics of VCES, and the safe
distance for various initial temperatures had been investigated.
In conclusion, for overpressure prediction of gas explosions, using simple empirical methods
is the most efficient way and the results have a certain level of accuracy when environments
are not complicated. However, if the structure of an oil and gas facility is highly confined and
congested, such simple methods may not be appropriate. The advanced CFD methods have the
ability to deal with the complex conditions and to enable a much more accurate and reliable
load prediction. Nevertheless, one of the critical issue of using CFD calculations is time
constraint. Normally a large amount of computational time that may be unacceptable is
required when CFD is applied to simulate gas explosions occurring at super-large and highly
congested facilities.
2.4 EXPLOSION RISK ANALYSIS METHODS
Both qualitative and quantitative analysis methods may be used to evaluate the explosion risks
of process facilities. The following is a brief description of typical risk analyses undertaken in
the process industry.
2.4.1 Qualitative Risk Analysis
Qualitative risk analysis is normally done by team studies based on the generic experience of
knowledgeable personnel and does not involve any mathematical estimation. Overall
qualitative evaluations are essentially checklist reviews in which questions or process
parameters are used to prompt discussions of the process design and operations that would
develop into an incident scenario due to an identified risk. Table 2.1 indicates an example of
qualitative explosion risk analysis based on API (2006) and UKOOA (2003). When levels of
consequence and likelihood are decided, a risk matrix as shown in Table 2.2 can be used to
determine the total risk level.
27
|
UWA
|
Table 2.1 Qualitative explosion risk analysis check list
Consequence:
Low congestion level due to the low equipment count, being limited
to wellheads and manifold with no vessels (i.e., no associated process
Low pipework)
consequence: No more than two solid boundaries, including solid decks
Unattended facilities with low maintenance frequency, less frequent
than 6-weekly
Medium congestion level due to the greater amount of equipment
installed compared to the low case
Higher confinement level than that for the low case
Medium Unattended facilities with a moderate maintenance frequency, more
consequence: frequent than 6-weekly
A processing platform necessitating permanent manning but with low
escalation potential to quarters, utilities, and control areas located on
a separate structure
High congestion level due to the significant processing on board,
which leads to a high equipment count
High
High confinement level of the potential gas release point
consequence:
Permanent manning with populated areas within the consequence
range of escalation scenarios
Likelihood:
Low equipment and inventory count, which align closely with the
Low consequence scenarios
likelihood: Low frequency of intervention, less frequent than 6-weekly
No ignition sources within the potential gas cloud
Greater amount of equipment installed than for the low likelihood
Medium Medium frequency of intervention, more frequent than 6-weekly
likelihood: Weak ignition sources, such as a hot surface, exist within the potential
gas cloud.
A high equipment and inventory count
High
Permanently manned installations with frequent processing on board
likelihood:
Strong ignition sources exist within the potential gas cloud.
Table 2.2 Risk matrix for risk level determination
Consequence of Failure
Likelihood of Failure Moderate Major Catastrophic
A B C
Almost certain 1 Medium risk High risk High risk
Likely 2 Low risk Medium risk High risk
Possible 3 Low risk Low risk Medium risk
28
|
UWA
|
Some of the widely used qualitative risk analysis methods are reviewed below. A checklist or
a worksheet is a standard list which identifies common protection features for typical facility
design and operations. Risks are expressed by the omission of safety systems or system
features. API RP14J (1993) gives a suitable basic checklist for offshore installations. A hazard
and operability (HAZOP) study is a formal systematic qualitative investigative safety review
method. It is used to perform a systematic critical examination of the process and engineering
intentions of new or existing facilities. Guidance on HAZOP is illustrated by CIA (1992),
CCPS (2011), and Kletz (1999). The HAZOP method is deeply reviewed by Dunjó et al.
(2010). Another qualitative method is called What-If analysis which originally introduced by
CCPS (2011). It is a safety review method using “what-if” investigative questions which are
asked by an experienced team of the system or component under review where there are
concerns about possible undesired events. HAZID is a particular form of hazard identification
commonly applied to offshore installations (Spouge, 1999). It is a systematic review of the
possible causes and consequences of hazardous events. Other common qualitative risk methods
in process industry includes preliminary hazard analysis (PHA) (Vincoli, 2006), failure mode
and effect analysis (FEMA) (Stamatis, 2003), layers of protection analysis (LOPA) (Summers,
2003), etc. Qualitative risk analysis is an effective tool for risk screening or routine inspection.
However, if detailed risk assessment is required, such methods may not be appropriate.
2.4.2 Quantitative Risk Analysis
Quantitative risk analyses are mathematical estimations based on historical data or estimates
of failures to predict the occurrence of an event or incident. An event tree analysis (ETA) and
a fault tree analysis (FTA) are the most widely used method to model explosion risks for
process safety and therefore this two methods are briefly reviewed.
An event tree is a visual model describing probable event sequences which may be developed
from a hazardous situation. (Vinnem, 2007). It uses branches to show the various possibilities
that may arise at each step and is often used to relate a failure event to various consequence
models (Spouge, 1999). A detailed procedure for constructing and analysing the ETA for a
process system can be found in Mannan (2012). For explosion events, the event tree
construction starts from a hydrocarbon release event and works through each branch in turn as
shown in the Figure 2.5 below.
29
|
UWA
|
𝑁
OR gate: 𝑃(𝐴) = ∑𝑃(𝐵 ) (2.5)
𝑖
𝑖=1
𝑁
AND gate: 𝑃(𝐴) = ∏𝑃(𝐵 ) (2.6)
𝑖
𝑖=1
where 𝑃(𝐴) is output event probability; 𝑃(𝐵 ) is input event probabilities; N is the number of
𝑖
input events.
If the probability are larger than 0.1, gates with two independent inputs should be calculated
as:
OR gate: 𝑃(𝐴) = 𝑃(𝐵 )+𝑃(𝐵 )−𝑃(𝐵 )𝑃(𝐵 ) (2.7)
1 2 1 2
AND gate: 𝑃(𝐴) = 𝑃(𝐵 )×𝑃(𝐵 ) (2.8)
1 2
The two quantitative risk analysis methods reviewed are widely applied to risk analysis of oil
and gas facilities. Aven et al. developed a barrier and operational risk analysis (BORA) method
which used event tree and fault tree to assess leak frequency of offshore platforms. Huang et
al. (2001) provided a formal procedure for the application of fuzzy theories to evaluate human
errors and integrate them into event tree analysis. Dong & Yu (2005) used fuzzy fault tree
analysis to assess the failure of oil and gas transmission pipelines and a weighting factor was
introduced to represent experts’ elicitations based on their different backgrounds of experience
and knowledge. Ferdous et al. (2011) implemented fuzzy set theory and evidence theory into
traditional event tree and fault tree analysis in order to provide a more robust method to handle
the uncertainty in QRA for the process systems. Wang et al. (2013) proposed a hybrid method
of fuzzy set theory and fault tree analysis to quantify the crude oil tank fire and explosion
occurrence probability.
However, since ETA and FTA only have simple Boolean functions and sequentially dependent
failures (Khakzad et al. 2011), it is difficult for them to reveal the complicated mechanisms of
interrelationships between risk factors. Therefore, an advanced quantitative analysis method,
the Bayesian network (BN), is implemented for risk assessments of explosion accidents when
multi-consequences and complex interrelationships are required to be considered.
32
|
UWA
|
P(A ∩ B)
P(A | B) = (2.9)
P(B)
If A and B don’t have any influence on each other, A and B are defined as independent events.
The multiplication rule can be written as:
𝑃(𝐴 | 𝐵) = 𝑃(𝐴)∗𝑃(𝐵) (2.10)
If there are mutually exclusive events𝐵 ,𝐵 ,… 𝐵 , then one of the events must occur.
1 2 𝑛
According to Bayes’ theorem (Johnson, Freund & Miller 2011), the rule of total probability
equation can be written as:
𝑛
𝑃(𝐴) = ∑𝑃(𝐵 )∗𝑃(𝐴 | 𝐵 ) (2.11)
𝑖 𝑖
𝑖=1
The probability of a particular state can be obtained using the following equation:
P(A ∩𝐵 )
𝑖
𝑃(𝐵 | 𝐴) = (2.12)
𝑖 P(A)
Substitute 𝑃(𝐵 )∗𝑃(𝐴 |𝐵 ) for 𝑃(A∩𝐵 ) and ∑𝑛 𝑃(𝐵 )∗𝑃(𝐴 | 𝐵 ) for 𝑃(𝐴)
𝑖 𝑖 𝑖 𝑖=1 𝑖 𝑖
𝑃(𝐵 )∗𝑃(𝐴 |𝐵 )
𝑖 𝑖
𝑃(𝐵 | 𝐴) = (2.13)
𝑖 ∑𝑛 𝑃(𝐵 )∗𝑃(𝐴 | 𝐵 ) for 𝑃(𝐴)
𝑖=1 𝑖 𝑖
where r = 1, 2, …, n.
BNs have been extensively applied to risk assessments in engineering aspects. Peng and Zhang
(2012) modelled a BN to assess human risks due to dam-break floods. An application of a BN
for earthquake risk management is considered by Bayraktarli et al. (2005). Zhang et al. (2014)
introduced a BN-based risk analysis method in construction projects. Lee et al. (2009)
presented risk management for large engineering project by using a BN and applying it to the
Korean shipbuilding industry. Meanwhile, the BN is increasingly used for risk analysis of
process facilities as it is flexible and well suited to taking the performance of human and
organisational factors in to consideration, and it offers a more precise quantitative link between
risk factors (Vinnem, 2007). Khakzad et al. (2011) used BN to conduct safety analysis of a
feeding control system that transfers propane from a propane evaporator to a scrubbing column,
and results proved that a BN is superior to a traditional FT model for complicated systems.
34
|
UWA
|
Haugom and Friis-Hansen (2011) built a BN of gas risks at a hydrogen refuelling station that
considered gas leak, jet fire, and loss of life. They concluded that a BN has greater freedom
and flexibility to analyse the dependence between the different variables than a standard ET.
Netherton and Stewart (2016) developed a risk-based blast-load modelling and used
conditional probabilities as an analysis tool for likelihood evaluation. Pasman and Rogers
(2013) implemented a BN to layer of protection analysis (LOPA) for gas risk analysis at a
hydrogen tank station and found that the BN had great potential in describing scenarios, dealing
with uncertainties, and supporting decision-making.
Furthermore, the Bayesian network is used in some parts of studies in this thesis when it is
appropriate and it has more advantages than traditional event tree or fault tree method when
evaluating vapour cloud explosion risks.
Firstly, evaluation of gas explosion risks is complicated when multi-factors (e.g. leak
severity, vent condition, structural complexity), multi-consequences (e.g. building
damage, human loss, environmental effect) and complex inter-relationships are
considered. Instead of traditional event tree and fault tree analysis, a Bayesian-network-
based quantitative risk analysis method has better ability to deal with complicated
mechanism and enables a more reliable risk analysis.
Secondly, the BN allows a quick and simple evaluation of each risk factors which offers
a clearer review of the criticality of each risk factor. Based on this kind of review,
decisions of further detailed assessments and risk mitigation measures can be made more
easily.
Thirdly, it is actually easier and more intuitive for engineers to understand a BN than
conventional event tree or fault tree. Meanwhile, assessments based on BN are also more
efficient than that based on traditional methods.
2.5 SUMMARY
This chapter provides a detailed literature review of factors influencing explosion loads,
prediction methods for explosion loads, and qualitative and quantitative analysis methods of
explosion risks.
The review of risk factors which may affect the explosion severity includes the confinement
and venting surrounding the gas cloud, the congestion or obstacles within the cloud, the fuel
35
|
UWA
|
properties and concentrations, and the ignition type and location. It is important to carefully
consider those factors when gas explosion risk analysis is conducted for oil and gas facilities.
Three empirical models, the TNT equivalency model, the TNO Multi-energy model, and the
Baker-Strehlow-Tang model, and three CFD codes, the FLACS model, the EXSIM model, and
the AutoReaGas model, are reviewed in section 2.3. The empirical methods are able to provide
an efficient overpressure prediction when the structures are not too complicated. CFD models
has the capacity of dealing with complex environments. However, the time consumption is
normally very large.
The risk analysis methods reviewed in this study includes both qualitative and quantitative
approaches. For qualitative methods, Checklist, HAZOP, What-If analysis, HAZID, PHA,
FEMA and LOPA are briefly introduced. In terms of quantitative risk analysis methods, event
tree and fault tree are reviewed and a BN model is explained in detail. In this study, BN is
applied more frequently than traditional qualitative and quantitative methods because BN is
more suitable for dealing with the complicated inter-relationships among explosion risk factors
when multi-consequences are considered.
In conclusion, there are some problems of current risk analysis to evaluate gas explosions
accurately and efficiently. First, multi-factors are involved in the gas explosion risk analysis
and inter-relationships among those factors are complicated. Thus, it is difficult for current risk
analysis methods to analyse gas explosion risk accurately as traditional methods such as event
tree or fault tree can only handle sequences with independent nodes. Meanwhile, load
prediction methods are either too simplified to provide a reliable risk evaluation when complex
conditions are involved, or too time consuming to enable an efficient risk analysis. Therefore,
it is important for the current study to develop risk analysis methods for more accurate,
efficient, detailed or reliable explosion risk evaluations.
36
|
UWA
|
CHAPTER 3. A BAYESIAN NETWORK-BASED QUANTITATIVE RISK
ANALYSIS METHOD FOR EXPLOSION AND FIRE ACCIDENTS AT PETROL
STATIONS
3.1 INTRODUCTION
This chapter introduces a more accurate explosion risk analysis method by using Bayesian
network (BN) when multi-factors, multi-consequences and complex inter-relationships are
considered. Traditional risk analysis methods such as event tree and fault tree can hardly be
applied when complicated relationship are involved. Therefore, a Bayesian-network-based
quantitative risk analysis method is developed to model explosion accidents and further
consequences. Petrol station is selected as an analysis model used to illustrate the proposed
Bayesian-network-based quantitative explosion risk evaluation.
For large-scale oil and gas facilities, a large amount of research has been conducted. Pula et al.
(2006) suggested a grid-based approach for fire and explosion consequence analysis as well as
an enhanced on-site ignition model to obtain better results in the consequence assessment
process. Suardin et al. (2009) proposed a fire and explosion assessment framework (FEAF) to
conduct risk screenings, options evaluations and assessment quality checks. Huang et al. (2017)
developed a multilevel explosion risk analysis method for super large offshore floating
liquefied natural gas facilities. However, very little research on explosion risk assessments has
been conducted for small process facilities such as petrol stations probably because of the fact
that consequences of accidents at service stations seem to be insignificant compared with those
at large-scale oil and gas facilities.
Nevertheless, since petrol stations are located close to residential areas, not only would process
facilities be damaged during an explosion event, but severe human loss may also occur because
of the large population of residential areas. For example, a catastrophic fire and explosion event
occurred at a petrol station as a result of a flood incident in Accra, Ghana, on June 4, 2015, and
caused over 152 fatalities (Asumadu-Sarkodie et al., 2015). Another recent accident happened
at Port-au-Prince, Haiti, on March 17, 2016. The fire and explosion killed 7 people and severely
burned about 30 others. Therefore, developing a risk analysis method for explosion accidents
at service stations is necessary. Meanwhile, fire accidents occur more frequently than
explosions at petrol stations and may also cause severe consequences. For instance, an
overfilling-induced spill happened in Mississippi, the USA, in August 1998, and led to a large
37
|
UWA
|
fire on an adjacent road and caused 5 deaths (Evarts, 2011). Therefore, a risk analysis of fire
accidents based on BN is also conducted.
Explosions and fires normally happen as a consequence of a domino effect from initial
accidents. Domino effect is determined to be a factor in which a sequence of events from the
initial accident may occur and lead to more significant consequences. The domino effect occurs
frequently in the process industry because escalations easily occur from oil and gas releases to
catastrophic fire or explosion events. Abdolhamidzadeh et al. (2011) summarised 224 major
domino events in the process industry, and they also found that vapour cloud fires or explosions
are the most common events causing domino effects. Drabra et al. (2010) selected 225
accidents with domino effects from a wide range of data sources. Their analysis indicated that
other than fire and explosion, loading/unloading operations also cause a significant number of
domino accidents, and human error has been proved to be one of the main reasons of accidents.
Kourniotis et al. (2000) examined a set of 207 major chemical accidents with domino effects,
and they concluded that accidents involving vapour hydrocarbons are the most likely to cause
domino effects, and liquid-fuel-induced domino accidents could cause the most severe
fatalities.
Meanwhile, studies have been conducted to analyse the domino effects of accidents in the
process industry. Abdolhamidzadeh et al. (2010) developed an algorithm named FREEDOM
(FREquency Estimation of DOMino accidents) to evaluate the domino effects of highly
complex and nonlinear systems. Cozzani et al. (2005) proposed a quantitative analysis for
domino accidents to estimate individual and social risks from domino scenarios. Khan and
Abbasi (2001) discussed the likelihood of domino events and developed a domino effect
analysis method for risk assessments of domino accidents.
To model complicated mechanisms caused by domino effects, a Bayesian-network-based (BN)
QRA is developed in this chapter because of the ability of the BN to deal with complex
interrelationships between risk factors. The BN is a probabilistic graphical model which
represents a group of random variables and conditional dependencies between them. It can deal
with multistate variables with different causal relationships, while the traditional event-tree and
fault-tree approaches only have simple Boolean functions and sequentially dependent failures.
However, the accuracy of BN modelling is limited by the difficulty of finding sufficient data.
Thus, three kinds of data, namely, practical information, computational simulations, and
38
|
UWA
|
subjective logical judgments, are included in the proposed study to improve the reliability and
accuracy of the proposed method. Practical information includes historical data regarding basic
risk factors, for instance, leak frequencies, ignition sources, and specific site information. The
numerical software, PHAST (DNV GL, 2016), is used to simulate explosions and jet fires with
different leak scenarios and output thermal radiations as input data for the BN analysis of fire
risks. Subjective logical judgments are applied when no data can be found. Such judgments are
useful for deciding conditional dependencies when the logic between nodes is easy to define.
Moreover, the current study focuses on the risk assessment of explosion and fire accidents
during the refuelling process from a fuel tanker to a petrol station. This scenario is selected
because a tanker stores a large quantity of flammable materials, which may cause significant
consequences if a second tanker fire or explosion occurs as a domino effect of an initial
accident. Both BN analyses of explosion and fire risks are conducted in this chapter.
3.2 BAYESIAN NETWORK ANALYSIS OF EXPLOSION RISKS
The proposed method consists of the following steps:
Modelling: Model BN based on risk factors and their interrelationships.
Quantification: Find data to quantify the established BN.
Calculation: Calculate the probabilities of target nodes of BN.
3.2.1 Bayesian network model of explosion events
A Bayesian network is an illustrative diagram which contains nodes and links with conditional
probabilities. As shown in Figure 3.1, a Bayesian network is built to evaluate risks of
explosions and loss of life when a leak occurs at petrol stations. The network consists of 14
nodes and 18 links which describe risks of leaks, explosions and further consequences. Nodes
and states of each node are listed in Table 3.1.
39
|
UWA
|
3.2.2 Quantification of Bayesian network
The quantification of the BN has two parts: finding the probabilities of the basic nodes and
determining the conditional probabilities of the interrelationship between nodes. Quantification
based on historical statistical data is the most convenient and reliable way. A total of 27 cases
of explosion accidents at service stations are selected and recorded as quantification data, as
listed in Table 3.2.
Table 3.2 List of recorded explosion accidents
No. Year Time Location DESCRIPTION Death Injury Reference
Explosion of gas
Riverton,
1 2015 1800 cylinders caused major ABC News, 2015
Australia
fire
Flood-swept fuel to
Accra, Asumadu-Sarkodie et
2 2015 2200 nearby fire caused 152
Ghana al., 2015
explosion
Fire ignited and escalated
Kaduna,
3 2016 1751 over 3 hours while tanker 3 Daily Post, 2016
Nigeria
discharged
Al-Ghubra, Car caught fire while
4 2015 1240 2 Times of Oman, 2015
Oman fuelling
Tanker caught fire and
Maddington, Department of Mines
5 2009 1345 exploded while
Australia and Petroleum, 2009
discharged
Electric fan heater in
Birmingham,
6 2015 1100 retail area heated up a gas 4 BBC News, 2015
UK
cylinder
Explosion occurred when
Kizlyar,
7 2016 1825 fuel truck discharged 40 RT News, 2016
Russia
LPG into tank
Port-au- Fuel tanker caught fire
8 2016 1700 7 30 Yahoo News, 2016
Prince, Haiti and exploded
Kuala Explosion caused by cell
9 2016 1330 Lumpur, phone while filling the 1 FMT News, 2016
Malaysia car
Truck explosion,
10 2015 2100 Cobar, NSW Daily Liberal, 2015
unknown cause
Southern
11 2016 1540 Khatlon, Explosion 1 17 Asia-Plus, 2016
Tajikistan
Vienna, Explosion caused by leak
12 2015 1205 WJLA, 2015
Austria of acetylene
41
|
UWA
|
Laval, UST exploded when ARIA Technologies,
13 1989 1200 2 1
France degassing and cleaning 2009
La Explosion while
ARIA Technologies,
14 1993 Gueriniere, discharging caused by an 2
2009
France electrical switch ignition
Annecy, Explosion caused by ARIA Technologies,
15 1997 1 1
France welding in tank manhole 2009
Compiegne, Explosion caused by ARIA Technologies,
16 1985 1
France worker-lit cigarette 2009
Les Cheres, Car crashed into ARIA Technologies,
17 2003 1 2
France dispenser 2009
Montlucon, New UST exploded ARIA Technologies,
18 2004 1200 1
France while filling 2009
Valleiry, Flash occurred while ARIA Technologies,
19 2004 1600 1
France filling car 2009
Aubigny-
Explosion caused by ARIA Technologies,
20 2004 1500 sur-Nere, 1
vapour leak from UST 2009
France
UST leaked and exploded
a few hours later, caused ARIA Technologies,
21 1958 1100 Paris, France 17 30
by spark from electrical 2009
switch
Car fire triggered an
Sotteville-
explosion of 16 LPG ARIA Technologies,
22 2007 0200 les-Rouen,
cylinders, fire at the 2009
France
service station for 4 hours
Tanker truck overfills
and spilled 2,839 litres of
fuel. The spill spread
Mississippi, outside the catchment
23 1999 0100 5 Evarts, 2011
USA and reached an adjacent
road and was then ignited
by unknown ignition
sources
Car crashed into pipes of
South
two 4,000 gallons above
24 1991 Carolina, 1 Evarts, 2011
ground fuel tanks; fire
USA
was put out after an hour
8,500 gallons of tanker
New
collided with car and
25 2014 0230 Orleans, 1 FOX8, 2014
crashed into petrol station
USA
concrete pillar
Kuala
Exploded during Channel News Asia,
26 2016 1515 Lumpur, 1
maintenance work 2016
Malaysia
Shanghai,
27 2007 0800 4 40 ABC News, 2015
China
42
|
UWA
|
As can be seen from Table 3.2, for most of available cases, only fatalities, injuries or estimated
economic losses are reported. Therefore, quantifying the interrelationships between nodes by
using historical data only is difficult. Thus, numerical simulations and logical judgements are
also applied in this study to deal with the limitation of statistic data.
DNV PHAST is used to quantify the interrelationships between leaks and consequent
explosions. Conducting PHAST analysis has four steps: input data, build model, calculate and
output result. These four steps are briefly introduced below, and more details about how to use
PHAST can be found in the PHAST manual (DNV GL, 2016).
Input data: including site map, weather conditions and data for explosion analysis
Build model: select the analysis method and determine explosion scenarios
Calculate: determine calculation scenarios and run the simulation
Output result: which can be GIS outputs, result diagrams and reports
Besides numerical simulations, logical judgements are also applied to quantifications of
interrelationships of the proposed BN. If the logical relationship between nodes is obvious and
easy to determine, subjective judgements are able to provide a certain level of accuracy and
reliability. However, such quantification requires regular examination, and if the site condition
changes, adjustment is required to ensure that the logical relationship is updated. Meanwhile,
if logical relationships are complicated and uncertain, a confidence-based method can be used
to reduce the uncertainties of subjective judgements (Huang et al., 2015).
3.2.3 Calculation of Bayesian network
The subnetwork of human loss is taken as a simple illustrative example to explain the BN
calculation. As shown in Figure 3.2, this subnetwork contains four nodes and three links. The
severity of human loss (node N) is decided by the severity of the initial explosion (node E) and
the number of people inside the petrol station (node M), which varies at different times of day
(node L).
43
|
UWA
|
Figure 3.2 An example of a Bayesian network
The prior probability of explosion loads can be calculated by Equation 3.1.
4 5 3
P(N = major) = ∑∑∑P(N = major,E = E , L = 𝐿 ,M = 𝑀 ) (3.1)
𝑖 𝑗 𝑘
i=1 j=1k=1
where P = probability, N = human loss, E = initial explosion, E = states of node E, L = time of
i
day, L = states of node L, M = number of people in station, M = states of node M (see Table
j k
3.1). Based on the theorem of the Bayesian network (Nielsen & Jensen, 2009), the joint
probability can be decided by Equation 3.2.
𝑛
𝑃(𝑥 ,…,𝑥 ) = ∏𝑃(𝑥 |𝑃𝑎(𝑥 )) (3.2)
1 𝑛 𝑖 𝑖
𝑖=1
where 𝑃𝑎(𝑥 ) is the parent set of 𝑥 . The function remains an unconditional probability of
𝑖 𝑖
𝑃(𝑥 ) if there are no parents of 𝑥 . In this subnetwork, the node of human loss has parents of
𝑖 𝑖
evacuation and jet fire, the node of evacuation has a parent of evacuation time, and the node
evacuation time has a parent of jet fire. Therefore, the following equation can be decided:
P(N = major,E = 𝐸,L = 𝐿 ,M = 𝑀 )
i j k
= P(M = major|L = 𝐿 ,E = 𝐸 )×P(M = 𝑀 |L = 𝐿 )×P(L = 𝐿 ) (3.3)
𝑘 𝑖 𝑘 𝑗 𝑗
×P(E = 𝐸 )
𝑖
where the conditional probabilities are decided by the interrelationship quantification.
44
|
UWA
|
3.2.4 Case Study – Quantifications
A case study is conducted to illustrate the proposed method and explain the quantification
process in detail. Figure 3.3 shows an example GIS map of a petrol station in Australia. An
explosion accident occurred in this petrol station when a tanker was refuelling the station.
Therefore, this site is selected as an example to demonstrate the proposed method. The
refuelling area is shaded in Figure 3.3, and the bund size is about 4 m × 6 m. In the case study,
each node of the proposed BN will be described and quantification will be introduced.
Figure 3.3 Target petrol station and shaded refuelling area
Quantification of Release Scenario
Table 3.3 shows 18 cases of spill incidents during the refuelling processes from fuel tankers to
petrol stations. These cases were collected from dangerous goods incident reports of Western
Australia between 1996 and 2008 (Department of Minerals and Energy, 1996;1998–2000;
Department of Consumer and Employment Protection, 2000; 2002; 2004; 2006; Department
of Mines and Petroleum, 2008). The probability of each scenario can be decided according to
the 18 cases as shown in Table 3.4. This case study is an illustrative example of the proposed
method. Thus, only 18 cases are selected to quantify leak scenarios. The more data are acquired,
the more accurate and reliable evaluation the BN is able to provide.
45
|
UWA
|
Table 3.3 Cases of spills
No. Year Location Volume of Spill Cause
1 1996 Dardanup 40 Misconnection
2 1998 Coolgardie 200 Overfilling
3 1998 Bassendean 22 Misconnection
4 1998 Kalgoorlie 315 Coupling failure
5 1998 Jarrahdale 70 Coupling failure
6 1998 Moora 80 Misconnection
7 1999 Swanbourne 50 Vapour recovery
8 1999 Upper Swan 50 Misconnection
9 2000 Australind 50 Overfilling
10 2000 Geraldton 300 Hose rupture
11 2002 Mt. Pleasant No record Overfilling
12 2002 Dampier Port No record Overfilling
13 2002 North Dandalup 750 Misconnection
14 2004 Canning Vale 20 Misconnection
15 2004 Kwinana 5000 Hose rupture
16 2006 Rivervale No record Overfilling
17 2006 Christmas Island 400 Overfilling
18 2008 Collie 8400 Hose rupture
Table 3.4 Probabilities of release scenarios
Scenario Description Number Probability
Gauging error, driver over fill
Overfill 6 33.33%
underground storage tank
Misconnection Driver error 6 33.33%
Hose rupture Mechanical failure of the unloading hose 3 16.67%
Coupling failure Result in disconnection of unloading hose 2 11.11%
Vapour Stage 1 vapour recovery connection
1 5.56%
recovery propped open
Quantification of Leak Severity
The probabilities of leak severities are classified into three categories: outside catchment, inside
catchment and inside cesspit. Figure 3.4 describes an example of a catchment of fill points.
Major spill is determined when the spill reaches outside the catchment, while minor spill is
determined when the spill can be held inside the cesspit.
46
|
UWA
|
Figure 3.4 An example of a catchment at a petrol station
The size of the catchment at the target site is approximately 6 m × 4 m based on the
measurement from the GIS map, and the height of the bund is about 50 mm. The size of a
typical cesspit is about 200 L, and the specific size of the cesspit at the target site cannot be
determined. Therefore, for a conservative estimate, the cesspit is assumed to be able to contain
120 litres of spill, which is about 10% of the total volume of the catchment. Consequently, the
severity of the 15 cases of spills can be determined as shown in Table 3.5. In Table 3.5, if the
height is lower than the height of the cesspit, the spill is defined as minor. When the height is
lower or higher than the height of the catchment, the spill is then to be decided as medium or
major respectively.
Table 3.5 Classification of severity of spills
Volume of Spill Scenario Height of Spill (mm) Major Medium Minor
40 Misconnection 1.67
200 Overfilling 8.33
22 Misconnection 0.92
315 Coupling failure 13.13
70 Coupling failure 2.92
80 Misconnection 3.33
50 Vapour recovery 2.08
50 Misconnection 2.08
50 Overfilling 2.08
300 Hose rupture 12.50
750 Misconnection 31.25
20 Misconnection 0.83
5000 Hose rupture 208.33
400 Overfilling 16.67
8400 Hose rupture 350.00
47
|
UWA
|
Quantification of Ignition Sources
As for the basic node of ignition sources, the most common heat sources ignited at the service
stations in the USA from 2004 to 2008 are displayed in Table 3.6 (Evarts, 2011). The result
indicates that smoking and heat generated from power equipment have the highest probability.
Table 3.6 Heat sources at a service station ignited from 2004 to 2008
Ignition Source Abbreviation Cases Probability
Smoking S 160 21.3%
Arcing A 90 12%
Hot ember or ash HE 140 18.7%
Spark or flame from
operating equipment SF 70 9.3%
Unclassified heat from
powered equipment UH 180 24%
Static discharge SD 40 5.3%
Heat or spark from
friction F 60 8%
Lighting L 10 1.3%
Quantification of Ignition
The number of ignition sources is used to determine the probability of ignitions, and it depends
on the size of spills. If the spill spreads outside the catchment, the chance of reaching ignition
sources is higher. Therefore, the incident of the spill spreading outside the catchment is
assigned with all the possible ignition sources, while the ignition sources of the spill incident
inside the catchment may be ignited by S, A, HE, SD or L. Meanwhile, only S, SD and L are
considered for minor spills because only they can reach spills inside the cesspit. The probability
of each source being ignited is determined to be 90% when it presents inside the spill range
and 10% when it is located outside.
Quantification of Initial Explosion
Release-related initial explosions are simulated by DNV PHAST under 15 leak severities. In
this study, wind effects such as wind directions and wind speeds are not considered. Therefore,
the explosive cloud is assumed to spread from the leak point. The Baker-Strehlow-Tang model
is selected in conducting the explosion analysis. A medium level of obstacle density and fuel
reactivity is determined based on the specific condition of the target station. Table 3.7 lists the
48
|
UWA
|
results of PHAST analysis, and Figure 3.5 shows an example of a PHAST output of an
explosion range of 0.689 bar when 400 litres of petrol spills.
To determine the severity of an explosion, an overpressure of 0.689 bar is used as an indicator
because such overpressure can cause direct death or severe injury because of the blast loads
and complete structure destructions (Lobato et al., 2009). If an area with overpressure higher
than 0.689 bar reaches the nearest dispenser (about 10 m) and the store area (20 m), the severity
is decided as medium and major, respectively. The severities of explosions are classified in
Table 3.7.
Table 3.7 Classification of severities of initial explosion
Explosion Severity
Spill Severity (L) Distance of 0.689 Bar (m)
Major Medium Minor
40 7.8
200 13.33
22 6.39
315 15.51
70 9.4
80 9.82
50 8.4
50 8.4
50 8.4
300 15.14
750 20.55
20 6.19
5000 38.82
400 16.67
8400 46.35
Figure 3.5 Circled area of an explosion range of 0.689 bar when 400 litres of petrol spills
49
|
UWA
|
Quantification of Leak Rate
Leak rate is necessary for jet fire analysis, and it can be calculated based on the leak volume
and time. Since no specific data of leak time for each case can be found, leak times of 60 and
300 s are used to calculate the leak rate according to the assumption that the driver has a range
of react time of 60 to 300 s to stop the spill because the driver may not realise the leak
immediately. For the 15 cases with recorded spill volumes, 300 s is a more reasonable
assumption because, normally, only a severe spill with a longer leak time will be recorded.
However, 60 s is also applied to provide another set of data of leak rates to conduct a more
conservative evaluation.
Based on the assumption of 60 and 300 s of leak time, 30 leak rates can be calculated as listed
in Table 3.8. According to HSE (2015), the severity of the leak rate can be considered major
when the leak rate for a liquid is more than 10 kg/s and minor when the leak rate is less than
0.2 kg/s. The density of normal gasoline is approximately 0.71 kg/l. Thus, Table 3.9 indicates
the severities of leak rates after calculation.
Table 3.8 List of leak rates
Leak Rate (litre/s)
Volume of Spill (litre)
60 s 300 s
40 0.67 0.13
200 3.33 0.67
22 0.37 0.07
315 5.25 1.05
70 1.17 0.23
80 1.33 0.27
50 0.83 0.17
50 0.83 0.17
50 0.83 0.17
300 5.00 1.00
750 12.50 2.50
20 0.33 0.07
5000 83.33 16.67
400 6.67 1.33
8400 140.00 28.00
50
|
UWA
|
Table 3.9 Severities of leak rates
Leak Rate Severity Leak Rate (l/s) Number of Cases
Major 14.08– 4
Moderate 0.28–14.08 18
Minor 0–0.28 8
Quantification of Jet Fire
In this study, fire severity is assumed to only affect the probabilities of tanker explosion, and
evacuation conditions and fire-induced fatality or injury are not considered. For the
interrelationship between leak rates and jet fires, a numerical simulation using PHAST is also
applied. Jet fires are calculated and quantified based on 30 leak rates as listed in Table 3.8.
Figure 3.6 shows the PHAST GIS output of the effect zone of a thermal radiation of 2 kW/m2
when the total volume of release is 20 litres and the leak time is 60 s. Minor injury, major injury
or fatality may occur based on the distance from the fire centre to the edge of the circled region
as shown in Figure 3.6. According to the Federal Emergency Management Agency (1990),
people may suffer severe pain or second-degree burn if they are exposed to a thermal radiation
of 2 kW/m2 for over 45 or 187 s, respectively. People who stay outside the circled region may
have around 1 minute of evacuation time, which is assumed to be sufficient for people to seek
shelter or escape. Therefore, the region outside the circled area is considered a relatively ‘safety
zone. Details of distances of jet fires with a 2 kW/m2 thermal radiation based on 30 release
rates is listed in Table 3.10.
Figure 3.6 Circled area of a thermal radiation of 2 kW/m2
51
|
UWA
|
Table 3.10 Severities of jet fires
Jet Fire Distance
Volume of Spill
60 s 300 s
40 23.49 11.27
200 48.88 23.49
22 17.89 8.58
315 60.1 28.89
70 30.31 14.56
80 32.22 15.47
50 26.01 12.48
50 26.01 12.48
50 26.01 12.48
300 58.78 28.26
750 89.11 42.89
20 17.13 8.21
5000 210.28 101.52
400 66.99 32.22
8400 265.71 128.43
The severities of fire are determined based on the sizes of specific sites. In this case study, the
size of this site is 40 m × 32 m, and the refuel points are located at the edge of the site.
Therefore, the classification of the severity of the jet fire is shown in Table 3.11. Major fire is
determined when fire covers the retail shop area, while moderate fire is considered when more
than half of the dispenser area is affected. Minor fire is determined when less than half of the
dispenser area is influenced. According to this classification, 4 of the 15 cases are major fire,
7 are moderate and 4 are minor.
Table 3.11 Classifying the severity of the tanker fire distance
Jet fire severity Fire Distance (m) No. of cases
Major > 34 10
Moderate >16 12
Minor < 16 8
Quantification of Tanker Explosion
For a fire-induced tanker explosion, the proposed BN considers the effectiveness of a safety
barrier in preventing or delaying the escalation of fire to the fuel tanker. The activation of the
safety barrier relies on the tanker outlet valves as shown in Figure 3.7. In the case of major jet
52
|
UWA
|
fire, the effectiveness of the safety barrier is limited even if the barrier is activated successfully
because of the high thermal radiation, which can cause rupture of the tanker’s vessels in a very
short period of time. Meanwhile, accessing and activating the lever valve may not be safe for
the driver if moderate or major jet fires occur. The barrier has a high chance of being activated
in case of minor fire because the driver has more time to turn off the lever valve safely. Table
3.12 indicates the conditional probabilities of the tanker driver activating the lever valve
successfully.
Figure 3.7 Typical tanker outlet valve in an open position (Department of Mines and
Petroleum, 2009b)
Table 3.12 Conditional probabilities of barrier
Active Safety Barrier Tanker Explosion
Jet Fire Intensity radii
Failed Succeeded Yes No
High 99 1 99 1
Medium 80 20 80 20
Low 50 50 50 50
No 0 100 0 100
When the active barrier fails to be activated, a tanker explosion is assumed to be triggered. The
severity of a tanker explosion is determined by the number of affected compartments. In
Australia, a typical fuel tanker consists of six compartments, which store different types of fuel
products. The more compartments are influenced, the more severe is the explosion. However,
in this study, the severity of a tanker explosion remains major because the minimum size of a
single compartment contains 5700 L of petrol. Figure 3.8 shows an explosion with an
overpressure of 0.689 bar when only one compartment of gasoline release is ignited, which can
be seen to lead to a major explosion affecting the whole target site. Therefore, for tanker fire
53
|
UWA
|
quantification, only statements of ‘yes’ or ‘no’ are determined; ‘yes’ refers to major severity
of explosion.
Figure 3.8 Circled area of a tanker explosion range of 0.689 bar
Quantification of Evacuation Time and Evacuation
Evacuation is assumed to be impossible when explosion is triggered because explosion
normally occurs within a very short period of time in which people are not able to escape.
Therefore, for release-induced explosion, evacuation is considered failed since people are
assumed to be unaware of the release until it is ignited and explodes. However, for fire-induced
explosion, evacuation before the fire reaches the tanker is possible. The time of evacuation is
assumed to be influenced by the fire severity. Consequently, successful evacuation depends on
the time of evacuation. Their relationships are illustrated in Table 3.13 and 3.14.
Table 3.13 Interrelationship between jet fire and evacuation time
Evacuation Time
Fire
Sufficient Short Little
Major 0 20 80
Medium 10 60 30
Minor 70 20 10
Table 3.14 Interrelationship between Evacuation time and evacuation
Evacuation
Evacuation Time
Succeed Shelter inside shop Fail
Sufficient 80 10 10
Short 40 30 30
Little 10 30 60
54
|
UWA
|
Quantification of Time of Day and Number of People
Quantifying the number of people depends on the time of day as the number of people at petrol
stations varies throughout the day. The time of day is classified into five categories. A total of
22 cases are recorded in Table 3.2 with detailed accident times. Table 3.15 shows the
probabilities of times when explosion occurred during a day.
Table 3.15 Hours of explosion at a service station
Hours Cases Probability
8:00–10:00 a.m. 1 4.6%
10:00 a.m.–4:00 p.m. 11 50%
4:00–7:00 p.m. 5 22.7%
7:00–10:00 p.m. 2 9.1%
10:00 p.m.– 8:00 a.m. 3 13.6%
The peak hours during the day are 8:00–10:00 a.m. and 4:00–7:00 p.m., while the quietest
hours are at night from 10:00 p.m. to 8:00 a.m. The target site has eight dispensers and
customers can use both sides of each dispenser to refuel their cars, which means 16 customers
can refuel in the same time. Throughout the observations of petrol stations with similar sizes,
two car waiting in the queue is quite normal at the peak time. Therefore, the customer number
is then decided to be 16 refueling and 16 queueing. The number of staff members is reduced to
two people during a normal period and one person during night time. The fuel tanker driver is
always counted as one person. The total number of people in the service station is considered
high, medium or low as shown in Table 3.16.
Table 3.16 Number of people in the service station at different times of day
Hours Server Tanker Driver Customers Queue Total State
8:00–10:00 a.m. 3 1 16 16 36 High
10:00 a.m.–4:00 p.m. 2 1 12 0 15 Medium
4:00–7:00 p.m. 3 1 16 20 40 High
7:00–10:00 p.m. 2 1 12 0 15 Medium
10:00 p.m.–8:00 a.m. 1 1 4 0 6 Low
Quantification of Building Damage
The store area of the target station is a single-storey structure. According to CCPS (2010),
overpressure of over 35 and 17 kPa will cause severe and moderate building damages.
Therefore, the severity of building damages is determined based on the explosion overpressures
55
|
UWA
|
at the centre of the retail store as listed in Table 3.17. Table 3.17 also shows the classification
of building damage.
Table 3.17 Quantification of building damage
Building Damage
Overpressures (kPa)
Spill Severity (L) Major Medium Minor
at Store Centre
(>35 kPa) (17–35 kPa) (<17 kPa)
40 20.48
200 33.9
22 16.19
315 38.4
70 24.82
80 25.9
50 22.18
50 22.18
50 22.18
300 37.9
750 48.29
20 15.55
5000 72.33
400 40.95
8400 76.54
Quantification of Human Loss
As mentioned, fire-induced human loss is not considered in this study. Therefore, human loss
is calculated based on two aspects, which are people inside a dangerous explosion range and
people who shelter inside the retail store. People who are inside the affected range and are not
able to escape quickly will sustain serious injury or fatality directly from the explosion, while
the risk of people who evacuate to the store depends on the building damage. Therefore, the
severity of human loss is determined by the number of people present in the service station, the
evacuation situation and the explosion severity and is classified as either major, medium, minor
or none.
3.2.5 Case study – Calculation and Discussion
After the quantification of basic nodes and interrelationships between nodes, human loss can
be estimated for a specific site based on the equation introduced in Section 3.2.3. The final
results are shown in Figure 3.9.
56
|
UWA
|
Figure 3.9 shows that when a release occurs, the probability of an explosion is about 3.1%, and
the possibility of further human loss is about 2.7%, which means that when an explosion
occurs, the probability of human loss is very high. Meanwhile, if human loss occurs, major
human loss takes a large percentage (2.3%), while possibilities of medium and minor losses
are only 0.3% and 0.1%, respectively. This result indicates that explosion accidents inside
service stations may result in significant consequences.
Figure 3.9 Final results of the BN calculation
Meanwhile, sensitivity studies on basic nodes and tanker explosion are also conducted in this
study. The proposed BN has three basic nodes: ignition source, release scenario and time of
day. To conduct a sensitivity study, each state of every basic node is assumed to occur with a
probability of 100%, while the probabilities of other states are determined to be 0%.
First, Table 3.18 indicates that the ignition sources SF, UH and F may cause human loss with
a probability of only about 0.6%. This happens because SF, UH and F are normally located
outside the catchment where only a major spill may reach. On the contrary, S, SD and L lead
to the rise of the probabilities of human loss by about 4.6% because these three ignition sources
57
|
UWA
|
are able to ignite a spill anywhere at the service station. In the meantime, S takes a very large
percentage (21.3%) of total probability among ignition sources, while SD and L only take 5.3%
and 1.4%, respectively. Therefore, smoking is determined to be the most dangerous ignition
source at petrol stations and should be totally banned.
Table 3.18 Sensitivity study on ignition sources
Human Loss
Ignition Source Abbreviation
Major Medium Minor No
Smoking S 4.56 0.74 0.35 94.35
Arcing A 2.52 0.29 0.02 97.17
Hot ember or ash HE 2.52 0.29 0.02 97.17
Spark or flame SF 0.58 0.05 0 99.37
Unclassified heat UH 0.58 0.05 0 99.37
Static discharge SD 4.56 0.74 0.35 94.35
Friction F 0.58 0.05 0 99.37
Lightning L 4.56 0.74 0.35 94.35
Second, the results in Table 3.19 show that the release scenario of hose rupture causes the most
catastrophic consequences by causing a 4.43% major human loss. As to the other four release
scenarios, the probabilities of human loss are slightly lower when overfilling or coupling failure
dominates, while the probabilities decrease significantly if only misconnection or vapour
recovery occurs. Table 3.19 also lists the frequency of each release scenario from AcuTech
Consulting Group (2014) and indicates that overfilling is the most frequent scenario. Therefore,
as the most frequently occurring and the most destructive release scenarios, overfilling and
hose rupture should be paid more attention when the tanker driver conducts the refuelling job.
Table 3.19 Sensitivity study on release scenarios
Human Loss
Release Scenario Frequency per Year
Major Medium Minor No
Overfill 10–1 2.34 0.32 0.08 97.26
Misconnect 10–3 1.44 0.27 0.16 98.13
Hose rupture 10–5 4.43 0.43 0.01 95.13
Coupling failure 10–3 2.04 0.3 0.11 97.55
Vapour recovery 10–3 1.14 0.26 0.19 98.41
The basic nodes of time of day only affect the severities of human losses. From Table 3.20,
one can conclude that there is not much difference among times during daytime. However, the
probability of major human loss decreases by about 1% at night because of the relatively small
number of customers. Thus, the refuelling job can be moved to night time if possible.
58
|
UWA
|
D Leak Rate 3 Major; Moderate; Minor.
E Ignition 2 Yes; No.
F Jet Fire 4 Major; Medium; Minor; No.
G Safety Barrier 2 Yes; No.
H Tank Fire 2 Yes; No.
Evacuation
I 3 Sufficient; Short; Little.
Time
J Evacuation 2 Evacuated; Failed.
8 AM – 10 AM; 10 AM – 4 PM; 4 PM – 7 PM; 7 PM – 10 PM;
K Time of day 5
10 PM – 8 AM.
L No. of people 3 Large; Medium; Small.
M Human loss 4 Major; Medium; Minor; No.
3.3.2 Case study – Quantifications
The quantification and calculation methods of the fire BN are the same as those of the explosion
BN. Therefore, they will not be introduced again. Figure 3.12 shows an example GIS map of
another petrol station which is used to conduct a case study of fire risks. The selected site
includes 6 petrol dispensers and a retail shop. The tanker refuelling area is circled in Figure
3.12. Some of the quantifications of fire factors are the same as those of explosion factors and
only quantifications of nodes with differences will be explained.
Figure 3.12 Target petrol station and circled refuelling area
Quantification of Leak Severity
The size of the catchment of the target site is 5.8 m*3.4 m, and the height of the bund is 50
mm. The cesspit is designed to contain 200 litres of petrol, which account for about 20% of the
61
|
UWA
|
The severities of the fires are determined based on the sizes of specific sites. In this case study,
the size of the site is 64 m*32 m, and the leak point is located at the centre of the target station.
Therefore, Table 3.23 shows the jet fire severity classification. The jet fire is considered to be
major when it covers the whole site, while it is considered to be moderate when half of the site
is safe. If less than half of the site is affected by the fire, the fire is classified as minor.
According to this classification, 12 of 30 cases are major fires, 10 are moderate, and 8 are
minor.
Table 3.23 Classifying the severity of the tanker fire distance.
No. of cases
Jet fire severity Fire Distance (m)
Major 32 – 12
Moderate 16 – 32 10
Minor 0 – 16 8
Quantification of Safety Barrier and Tanker Fire
The proposed BN considers the effectiveness of a safety barrier in preventing or delaying the
escalation of the initial fire to the fuel tanker. When the barrier activation fails, a tanker fire is
triggered. The severity of a tanker fire is determined by the number of affected compartments.
According to the Department of Mines and Petroleum (2009b), a typical fuel tanker in Australia
has a total of six compartments and normally uses a 100Ø-mm flexible discharge hose to refuel
the underground storage tank. In this study, the severity of the tanker fire is major because a
hose rupture is assumed to occur for a jet fire-induced tanker fire, and the hose rupture of a
100Ø-mm hose will lead to a release rate of 16.67 litres/s, which is large enough to cause a
major fire. Therefore, for tanker fire quantification, only statements of “yes” or “no” are
defined, and “yes” means major fire.
Quantification of Time of Day
Historical data regarding fire incidents at service stations is used to determine the times during
which fire accidents occurred. The data is mainly obtained from yearly dangerous goods
incident reports (Department of Minerals and Energy, 1996-2000; Department of Consumer
and Employment Protection, 2001-2007; Department of Mines and Petroleum, 2008-2015), a
report on fires at U.S. service stations (Evarts, 2011) and a report on petrol station accidents in
France from 1958 to 2007 (ARIA Technologies, 2009). Table 3.24 summarises the
63
|
UWA
|
From Figure 3.14, it is evident that, when release occurs, a jet fire has a probability of
approximately 3% to occur and the probability of a major jet fire occurring is around 1.6%.
Meanwhile, regarding human loss, there is only an approximately 2% chance that fires will
endanger people at a petrol station. However, major human loss accounts for a very large
percentage (1.65%), while the possibilities of medium loss and minor loss are 0.35% and
0.05%, respectively. This means that fire accidents at service stations may not always lead to
human loss. However, if human loss occurs, the consequences will probably be significant.
Compared to the explosion accidents, it can be seen that the probabilities of fire and explosion
remain same. This occurs because the same data of release scenarios and ignition probability
is used for both case study. However, the consequence of explosion accident is much severe
than that of fire events. As can be seen from Figure 3.09 and 3.14, the total human loss reduced
by around 30% (from 3% to 2%) and major human loss decreased from 2.3% to 1.6%.
Therefore, it can be concluded that fire accidents would be less catastrophic than explosion
events, though fires can cause severe human losses.
3.4 SUMMARY
In this chapter, a more accurate quantitative risk analysis method for explosion and fire
accidents at petrol stations is proposed. A Bayesian network (BN) is implemented as a risk
analysis tool in this study to estimate the probabilities of initial leaks and consequent domino
effects, such as ignitions, explosions, fires and human losses.
The proposed BN-based method is different from the conventional ones. It aims at building
relationships between risk influencing factors and exploring the importance of basic risk factors
rather than offering an annual fatality risks (AFR) number. The decision of risk reduction and
mitigation can then be made based on the evaluation of risk factors. The more important a risk
factor is, the more detailed risk assessment and mitigation measures should be focused on.
Two case studies of explosion and fire risks are conducted to illustrate the applicability of the
proposed method. In the case study, the quantification of each node is described in detail, and
results show that petrol leaks may lead to human loss such as death or injury with a probability
of approximately 3% and 2% when explosion and fire occurs respectively.
Two BN models with explosion and fire risk influence factors was built to model and evaluate
the risks of explosion and fire accident and human safety. The case study proved that the BN
65
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.