RISK-BASED MAINTENANCE & INSPECTION DECISIONS

John Woodhouse, Project Manager Eureka MACRO programme

Managing Director, The Woodhouse Partnership Ltd

Abstract

MACRO is a multi-industry, dti-supported joint venture (European ‘EUREKA’ project

EU1488), that is developing guidance and tools for better cost/risk evaluation and Asset
Management decision-making. It is bringing together the technical aspects of reliability
and risk-based analysis with the economic factors, human issues and data uncertainty.
This paper focuses on some of the resulting ‘best practice’ methods, analytical tools and
the results that are being achieved.

The boundaries of the project have been chosen to meet the perceived priorities and a
practical development timeframe. Nevertheless the range of analysis areas to be addressed
is wide and includes:

• Asset Life Cycle Costing: the hands-on tools for project evaluation, equipment
replacement, life extension, change control & modifications.
• Maintenance Optimisation: evaluating preventive maintenance strategies,
optimum intervals, legal/safety compliance, and environmental constraints.
• Inspection & Condition Monitoring: setting inspection, monitoring & test
intervals, optimal condition reaction points, and cost/benefit of monitoring
methods.
• Work Grouping & Shutdown Strategy: evaluating optimal task groupings,
shutdown intervals and opportunities.
• Materials & Resources: cost/risk optimisation of slow-moving spares,
consumables, supplier comparisons, materials pooling and alliances.

This paper concentrates on just two of these areas: Maintenance Optimisation and
Inspection/Condition Monitoring strategies as introduced overleaf:
1. INTRODUCTION

The international MACRO project has been busily collating the experiences and best
practices of collaborating organisations in many industries. In addition to the development
of innovative technical methods, MACRO is generating procedural guidance and training
programmes to implement risk-based management techniques. High among these are the
procedure guidance notes for reviewing or setting maintenance & inspection strategy. Of
course this subject has received a lot of exposure over the last few years - mostly focussing
on particular ‘initiatives’ such as RCM, TPM or other acronym-packaged frameworks.
The following paper is a summary of the MACRO project observations and
recommendations regarding best practice and the use of various tools (such as Function &
Criticality Analysis, FMEA, RCM and optimisation methods).

Four major areas of work have been at the core of MACRO developments during the last 6
months. These take the form of Working Parties, discussing best practices, the
specifications for any analytical tools that might be required and practical experiences of
implementation. The four active Working Parties at this stage of the project are:

a) Navigator: developing a master routemap for selecting appropriate analysis tools for
different types of problem and decision.
b) Maintenance Strategy: combining the experiences of different tools in different
circumstances to develop a robust self-help guide. This also includes the innovative
developments in quantitative risk-based techniques.
c) Inspection Strategy: again, introduction of quantitative techniques to set condition
monitoring or test/inspection strategy. This team is providing the hard tools for Risk-
based Inspection.
d) Lifespan: the Life Cycle Costing area developed to the extent of linking specific
decision requirements to the necessary tools and techniques to justify the optimum
solution.

2. MAINTENANCE STRATEGY PROCESS

2.1. ‘Horsesfor courses’

The first conclusion reached by the Maintenance Strategy Working party was the need for
a mixture of methods to determine what work is worth doing and when. No single formula
yet on offer was found to be suited to different industries, or even to different processes,
plant types or departments within the same company. The depth of analysis effort, and the
value-for-money of such analysis, is clearly dependent upon the importance of arriving at
the correct strategy. Criticality filtering of the systems, equipment and failure modes is
vital to avoid ‘analysis paralysis’ and loss of direction. Visible return for the effort is also
essential to maintain enthusiasm and management support for any systematic initiative.

The overall flowchart that has emerged from the MACRO team is one of multi-level
analysis. Dependent upon process or functional criticality, differing levels of analysis
effort should be applied. At the top end, perhaps 5-10% of the most vital corporate
functions, quantitative risk and performance analysis is warranted. For the next 40-60% of
‘core business’ activities, template and rule-based methods (such as RCM or RBI) are
more appropriate. At the lower levels of process criticality, not even the simple tabulated
questions and FMEA work are worthwhile - a cruder but quicker ‘filter’ is required.
Figure 1 shows the overall flowchart up to the point of individually identified and justified
maintenance tasks. The process of consolidation and optimisation of an overall
© The Woodhouse Partnership Ltd
2
maintenance programme is the subject of a specific MACRO working party, which is
developing dynamic optimisation methods for work clustering and shutdown strategies
(“APT-SCHEDULE”).

1. Functional Analysis

2. Criticality Assessment
a) System level
Typically 40%

b) Asset/Failure mode level

Typically 10% Typically 50%

Quantified Manufacturer’s
Rule-based Analysis Recommendations
3. Deterioration e.g. RCM, RBI
Modelling or Current Practice
Performance

Equipment Failure Mode Consequence M’tce Method

“Reverse RCM”
filter
Risk

tasks > reasons

Intervals worth
cost/benefit & intervals
calculating
Optimum
reaction points Design
Designchanges
changes
Preventive
Preventivetasks
tasks
Detective
Detectivetasks
tasks
Corrective
Correctivetasks
tasks
Construct optimal
4. groupings & schedule

Opportunities
Cost & & Constraints
Risk
Impact

Task combinations
Copyright 1994 The Woodhouse Partnership Ltd

Figure 1. Strategy methods depend upon Operational Criticality

© The Woodhouse Partnership Ltd

3
2.2. Functional focus
In order to determine which physical assets are worth maintaining to what degree (and
which are worth analysing to what level), a shift of emphasis is vital. Maintenance
strategy has historically been directed at types of equipment. The same recommendations
on maintenance work and intervals are issued, whatever the operational role or importance
of that equipment. The service interval and task list for a Vauxhall Cavalier is the same,
whether the car is just one available from the company vehicle pool, or is a doctor’s sole
means of transport! Clearly the consequences of breakdown can be very different – so the
importance of reliability differs and the level of maintenance should be adjusted
accordingly.

Despite the common horror stories of inappropriate application, one of the big advantages
of RCM logic is that it considers equipment function and loss of function as important
criteria. However, in its original form, RCM only attaches these characteristics as an
attribute to the equipment. The whole asset list is reviewed and, for each piece of
equipment, the functional failure, operational or other consequences and the failure modes
are all employed to determine the appropriate maintenance strategy. Unfortunately, by this
route, the analysis of each equipment’s characteristics has to be almost complete before it
becomes clear whether it was worthwhile examining in the first place. The team doing the
review has to apply nearly the whole procedure to find out which items were worth
reviewing at all! To allow an earlier filter and prioritising of such analysis effort,
therefore, a clear understanding is needed of which systems do what, and what happens if
they do not. This process mapping or ‘functional breakdown’ can shift the focus
dramatically. Not only does it provide a means of prioritising the maintenance strategy
studies, it also achieves a wider operational awareness (it can be a revelation to
maintainers and operators alike) and invariably stimulates ideas for design or procedural
improvement.

The methods for mapping equipment functions are similar to those employed in ‘business
process re-engineering’. However, the terminology and process (or systems) viewpoint
may not be familiar to the operators and maintainers who should be involved in developing
the map, so guidance and facilitation is usually needed. A summary of a simple level of
Input-Process-Output diagramming was published in the Maintenance Journal in June
1997. More detailed methods (such as the ICOM format, which separates out the Inputs,
Constraints, Outputs and Mechanisms) are available but the method is less important than
the fact of considering operational requirements and failure likelihood/consequences first,
and the necessary equipment (and its maintenance requirements) second. The underlying
objective is to direct the costly analysis effort at the most important functions or core
business of the organisation.

2.3. Criticality analysis

The commonly promoted versions of RCM focus firstly on equipment, and then its
operational context, failure modes, consequences and maintenance requirements. The
MACRO network has clearly concluded that this is wasteful and often results in duplicate
consideration of identical or similar maintenance requirements, and low-value analysis of
marginal equipment or failure modes. A common feature of successful implementations is
the criticality-based priority or filtering of which systems, equipment and failure modes are
worth analysing in the first place. This may not be formalised ‘criticality system’ but, if

© The Woodhouse Partnership Ltd

4
the review of maintenance requirements is to be systematic and/or will involve a wide
range of personnel from different backgrounds, then a consistent and generally agreed
ranking method is necessary. A survey of such methods, with practical applications
guidance, has been prepared as one of the reference documents of the MACRO project –
and was published in summary form at the Advances in Reliability Technology
Symposium in April (Manchester Univ.) and in the Maintenance & Asset Management
Journal last month.

The key feature of successful ranking methods is the combination of failure consequences
(safety, economic, environmental or others) with the likelihood or frequency of failures.
Degree of detail varies, along with the guidance and ‘weighting’ methods for scoring the
different elements, however the combination (usually a multiplication of frequency and
consequence) aims to prioritise the small-and-frequent among the big-and-rare.

Whatever method is used, the resulting process criticalities can be ranked and charted to
show any natural groupings. An histogram (Pareto-style) usually reveals a few ultra
critical activities, with a core group of functions beneath them, and a large number of low
criticality support functions (i.e. three bands of criticality can be chosen). See figure 2 for
an example.

Typical Process Criticality results

250.00
CRITICALITY SCORE

200.00

150.00

100.00

50.00

0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

FUNCTION or SUB-SYSTEM The small and many

The critical few
The core of the business

3. Maintenance Strategy Selection

The MACRO project has developed structured methods to review maintenance
requirements at each level of criticality. These merge existing best practices with
innovative improvements and tools.

Figure 2. Criticality determines analysis method that is appropriate

RCM and TPM logic, for example, are incorporated in their correct places, with supportive
guidance on the evaluation of efficiency-oriented maintenance, lifespan-related tasks (such
as painting or lubrication) and the cost/risk basis for setting inspection or maintenance
© The Woodhouse Partnership Ltd
5
intervals. A series of articles and case studies are being prepared to cover each of these in
turn, starting with the radical advances in the area of quantified (cost/risk) evaluation of
maintenance options. These and related methods are also being collated into Management
Mini-guides for publication next year. The corresponding analysis tools are already
available and are on display in the exhibition attached to this conference.

3.1. Quantified risk-based maintenance

Once the process criticalities have been identified, the analysis of maintenance
requirements splits into different levels of detail. For the few really vital processes (5-10%
of all systems), the approximations and black/white assumptions of RCM or RBI (‘risk-
based inspection’) are rarely sufficient to determine the optimal combination of operating
and maintenance strategies. For example, RCM and RBI both require a clear and
consistent definition of ‘functional failure’ - yet this is often a grey area in real life. What
degree of deterioration do you choose to define as ‘unacceptable’? The level of risk that is
worth taking, and the quality or performance that is achieved, or the life expectancy of the
asset are often trade-able commodities that can be influenced by the amount and type of
maintenance. In critical areas, the additional head scratching and data collection required
to quantify such risk and performance trade-offs can be very worthwhile. MACRO field-
testing has shown scope for multi-million pound savings through such a quantitative
approach. Examples range from painting programmes and lubrication schedules to major
overhaul or shutdown strategies and inspection/test intervals. A representative case study
is attached at Appendix A to this paper.

4. Risk based Inspection

The second big area of quantitative analysis is the setting of inspection and condition
monitoring strategy. Much work is going on in this area in the United States, albeit mostly
concentrated on the monitoring of static equipment (vessels, pipes etc) on petrochemical
plants. The MACRO focus has been wider and more quantitative. A cost/risk balancing
tool has been developed to evaluate optimal strategy in the light of various uncertainties
(rate of deterioration, point of failure, quality of measurement etc.) and the early results of
its application are extremely encouraging. Applications already proven include the
condition monitoring of wooden poles (electricity distribution), corrosion monitoring of
pipes and storage vessels, function testing of safety protection and standby equipment, and
instrumentation. An example of such application is illustrated in Appendix B to this paper.

Regarding the value of risk-based study (instead of simpler, rule-based approaches), the
MACRO team has performed some systematic comparisons: a sample of 5 condition
monitoring strategies, arrived at by RCM/RBI guidance, were checked with the APT-
INSPECTION analysis. 3 of the 5 were found to be about right (+/- 20% of optimal
timing) but one decision was a factor of 8x in error (the job should have been done at 1/8th
of the interval, and the other was not worth doing at all. On such a 40% hit rate -
generating significant improvement - the payback for the additional ‘brain-strain’ (in
practice, about 2 hours per problem tackled) was almost instantaneous.

© The Woodhouse Partnership Ltd

6
5. Other areas of interest

5.1. Lifespan
The fields of asset replacement, life extension and life cycle costing are receiving
increasing attention also. A North Sea “Joint Industry Project” is embarked on developing
an ISO standard for Life Cycle Costing, and the pressure for ‘alliances’ between
contractors, vendors and operators can only increase. The MACRO work is focussing on
delivering the practical tools that can aid the evaluation of specific decisions (such as the
choosing repair/replacement point, evaluating a life extension proposal or comparing A vs.
B equipment options). APT-LIFESPAN is due for public release toward the end of this
year.

5.2. Navigator

A very big task faced the MACRO “Navigator” working party. It has been trying to map
all the possible decision triggers (events or circumstances in which asset management or
engineering decisions need to be made) onto the common groupings of the appropriate
investigation, solutions and evaluations that have to be made. After developing a
consistent mapping format and hierarchical terminology, the many decisions have been
found to map quite elegantly onto a few generic templates. As a result, the key questions
and investigation steps have been condensed into an easy-to-follow routemap: one for
equipment-related operational problems, one for project environment or systematic studies
(e.g. maintenance strategy or materials requirements) and one for mitigation and
contingency options. A series of wallcharts, interactive CD-ROM and Web-site guidance
is under construction.

© The Woodhouse Partnership Ltd

7
Appendix A: RISK-BASE MAINTENANCE CASE STUDY

Maintenance vs. Asset Replacement decision: Reliability projections are combined with
failure consequences and the costs of planned maintenance/replacement. The uncertainty in
most data is handled by range-estimation and sensitivity testing - showing that the decision
is robust even in the light of widely varying assumptions. It can be shown, therefore, that
no further data is needed.

Calculated Results, showing optimal replacement point (12 years), the conflict between
planned expenditure and risk exposures, and the cost/risk penalties for replacing too early
or too late:

Appendix B: RISK-BASED INSPECTION CASE STUDY

© The Woodhouse Partnership Ltd
8
A typical corrosion monitoring example - with uncertainty about the deterioration rate, the
onset of deterioration, the point of potential failure and the measurement quality/accuracy.
Inputs include the direct and indirect costs of monitoring and different levels of failure
consequence. All data can be range-estimated and tested for sensitivity. Other examples
range from visual inspections of substations, ultrasonic testing, vibration monitoring to
operational checks of protection equipment and functional tests of standby equipment.

Results: showing optimal inspection interval at 6-7 years (current policy was 3-yearly, so a
50% reduction in monitoring costs is available). If regulator or safety requirement limit
inspection intervals to, for example, 3-yearly, the following graph shows the premium paid
for such compliance - and can help to justify alternative options (such as design changes,
negotiations with the authorities etc).

© The Woodhouse Partnership Ltd

9