Pipeline Technology Conference 2016

Asset Integrity Management

and the Dilemma of Maintenance Planning

Dr.-Ing. Felix Weise

Dipl.-Ing. Florian Dinauer
M.Sc. Peter Wiecha
Dipl.-Ing. Christoph Schmidt
DNV GL Oil & Gas Germany

1 Asset Integrity Management

The planning and optimization of maintenance programs is an important part of an

Asset Integrity Management and can be a complex and difficult task for managers of
pipeline network operations as well as for their maintenance personnel due to a
possible lack of component data, information and decision variables. In addition
pipeline network operators in Germany need to fulfil legal requirements (e.g. German
"Fernrohrleitungsverordnung", see [4] and [5]) which prescribe the frequency of
maintenance activities and thus limit the freedom of the maintenance planning
process. Due to the fact that pipeline operators are part of the energy distribution
network a highly reliable operation with a minimum number of unplanned shutdowns
and unavailability’s are mandatory and requested by the customers. Interruptions of
the distribution process and unplanned stops can lead to high financial losses. The
maintenance departments are forced to maximize the availability of the operation,
often acting with a limited maintenance budget and resources. Considering the
before mentioned limitations the planning, evaluation and optimization of preventive
maintenance activities is a complex and difficult task that can be only fulfilled with a
systematic approach and a quantitative decision basis.
For this special purpose DNV GL developed a methodology called Risk Based
Inspection (RBI) that uses quantitative risk reduction measures for the definition of
cost efficient inspection plans. A more intensive inspection effort is awarded to plant
components that have been identified within the RBI analysis as risk drivers
compared to those equipment that have only limited impact on the overall risk of a
pipeline distribution network. In addition, the confrontation with the own risk as well
as the associated quantification of the benefits of specific inspection measures
creates a robust internal information database. In the further process this can be
used as a basis for decision making with regard to the specification of safety
objectives related to the organization as well as the verification of the necessary
inspection and maintenance budget.
Beside inspections, regular maintenance activities and the storage of spare parts are
additional risk-reduction measures intended for practical usage. The inspection is the
most widely used measure related to Risk Based Inspection (see [1] and [2]), but a
concept for decision making in the context of the storage of spare parts has also
already been established which is based on risk reduction (see [3]). In contrast to
inspections and spare part holding, for dedicated maintenance activities (e.g.
 Pipeline Technology Conference 2016

lubrication of gears) a detailed embedding into the concept of risk-based planning

has not been carried out so far.
Therefore this article primarily presents a risk-based maintenance model which can
be utilized for planning of preventive and periodic maintenance measures (e.g.
lubricating, greasing or adjusting system components) whereas inspections are more
related to checking and measuring the actual condition of a component.
The definition of the system is the first step during the execution of a RBI project (see
Figure 1). In this step the covered plant is divided into functional packages and
maintainable items. Subsequently the possible failure modes and the likelihood for a
failure mode and consequence are investigated. The failure rate is almost constant
during a normal operation mode and can be calculated based on the average lifetime
(MTTF “Mean Time To Failure”) of a component. In addition to the likelihood the
potential extent of the damage is taken into account for the risk calculation. The
damage generated by a component failure with regard to the down time generally
consists of a fixed part, such as the costs to start and stop the plant for repair and a
variable part, which quantifies the loss of production for the duration of the down
time.
Based on the calculated risk value the planning and optimization of the available
maintenance measures is carried out iteratively and their influence on the component
risk is determined. The depth of the analysis is developed with the user and ranges
from a first screening of the plant components and a qualitative risk classification, an
individual check of measures and finally to a complete quantitative optimization of all
maintenance activities taking the complete system into account.
Additionally, maintenance measures that have been carried out already can be taken
into consideration for the risk assessment and will be evaluated with regard to their
potential to reduce the risk by holding the maintenance budget constant. The
assessment of the current maintenance and inspection program can also be utilized
for the development of corporate key performance indicators (KPI) in the area of
maintenance and thus initiate a continuous improvement process. The different steps
of the RBI are schematically shown in Figure 1.

Figure 1: How to manage a RBI project

 Pipeline Technology Conference 2016

2 Risk Based Maintenance Planning

Maintenance should be conducted regularly to avoid degradation and wear as

defined in DIN 31051 (see [3]); i.e. such activities extend the life time of a component
[4]. In contrast inspection activities “only” improve the knowledge about the damage
state and condition of a component. Maintenance measures actually modify the
physical state of a component. Improving the state of the component leads generally
to an extension of the average life (= Mean Time To Failure MTTF).
The life-prolonging effect of a maintenance operation is affected by many factors.
The quality of a specific maintenance measure is of essential importance for
example. It can be described as efficiency α of the maintenance package.
Effectiveness describes the quality of maintenance measure and thus their impact on
the extending of the component life. In general the theory distinguish between the
two extreme conditions "as good as new" (α = 1) and "as bad as old" (α = 0). When a
maintenance operation leads to a state, which is in accordance with the state shortly
after commissioning the component is “As good as new”. If the considered
component reaches only the state "as bad as old" after performing a maintenance
operation the component state isn`t changed by the maintenance operation at all.
Another influence factor which should be taken into account during modeling
maintenance measures is the maximum improvability (L) of the considered
component. The improvability describes the maximum possible percentage of
extending the component life. It can be assumed that also a fully effective
maintenance activity increases the component life only by a certain percentage.
Normally a component specific upper limit is assumed. The life extending impact of a
measure is formally shown in the equation below:

MTTF’ = (1 + α • L) • MTTF

MTTF' is the average life after a maintenance procedure.

The effect of maintenance measure on the component reliability is shown in Figure 2.
Two reliability functions, representing almost an exponential failure behavior, are
shown. The time tW represents a characteristic point of maintenance time.
 Pipeline Technology Conference 2016

Figure 2: Maintenance measures improve the state of a component

The first curve (black) represents the original function of reliability before performing
maintenance. At the time of commissioning, the component reliability is 100% (R =
1). The reliability decreases with increasing operating time, i.e. increasing the
likelihood that the component at a later date has already failed. Again to increase the
reliability of the component at a given time, the maintenance is carried out according
to a given maintenance interval at the time tW. Assuming a fully efficient maintenance
(α = L = 1) the reliability of the component will be restored to its original value. This
means that the component is after performing a fully efficient service in the state,
which it had at the time of commissioning (i.e. the reliability is 100% again). From this
point, the component reliability decreases again, until another maintenance measure
is carried out at a later date.
Now the important question is: What is the optimal time interval tW* of a dedicated
maintenance activity? The curve shown in Figure 3 represents an idealized curve of
the risk reduction potential (on the y-axis) with increasing maintenance intervals (on
x-axis). This simplified graph can be helpful to answer the above mentioned question.
 Pipeline Technology Conference 2016

Simple reflections allow expecting the following behavior:

If the component is maintained right at the time of commissioning/start of the
machine or plant, also a fully efficient maintenance activity may cause no risk
reduction, because the component has experienced no wear and tear at this time.
Maintenance at this point of time would have no influence on the state and condition
of the component and will bring therefore probably no risk reduction (it would even
worsen the condition of the component under specific circumstances). With an
increased operating and running time of the machine, the reliability of the component
decreases according to Figure 2. This increases the likelihood that the component, at
the time of the maintenance, has already failed. Also a fully effective maintenance
can restore the component state only with diminishing probability when the time
increases, namely only if the component has not failed yet. In this case a preventive
maintenance measure has a lower risk reducing effect compared to a maintenance
operation performed at an earlier date.
This means that the maintenance effect, described as the potential for risk reduction,
rises continuously from the date of commissioning of the component and then at
some point decreases until finally no risk reduction effect is observable. The curve
indicates clearly a maximum whose determination is essential for the planning of the
optimal point of maintenance time.

Figure 3: Schematic representation of the maintenance-related risk reduction

when choosing different maintenance points
 Pipeline Technology Conference 2016

3 Applying of a maintenance improvement project

 

As practical illustration of the maintenance model described in chapter 2 will be

summarized in this chapter based on three selected components that will be
analyzed and the maintenance measures will be optimized with regard to the best
maintenance interval.
A gearbox, a chain and a bearing of a technical facility/machine, whose specific
mean time to failure (MTTF) values has been taken from an expert database (e.g.
OREDA or organization internal databases), will be analyzed. The available
maintenance measures for each component are given in Table 1.

Table 1: Data of the Case Study

No Component Expert Judgement
Available Maintenance program
Ø MTTF [h]
1 Gear 11.000 - Maintenance after tw1 or tw*
2 Chain 17.000 - Three inspections of different effectiveness
3 Bearing 4.000 (Typ A,B and C in accordance API 580/581)

The maintenance effectiveness and the maximum improvement factors of the

component-specific maintenance measures are specified in table 2. In this example it
has been assumed that all components have a maximum improvement factor of
20%. Depending on the component specific maintenance measures the factor L can
vary from 0 to 100 %.
Furthermore there will be three different effective maintenance measures considered:
a highly effective (oil change, with 90%), an average effective (lubricate bearings with
60%) and a little effective maintenance measure (adjustment of the chain with 30%).

Table 2: Maintenance effectiveness and maximum improvement factors

No Component specific Maintenance Maximum
maintenance measure effectiveness (α) improvement (L)
[%] [%]
1 Change of gear oil 90 20
2 Adjustment of chain 30 20
3 Lubrication of bearing 60 20

The percentage of risk reduction with increasing maintenance interval lengths for the
three considered components is shown in Figure 4.
 Pipeline Technology Conference 2016

Changing the oil leads due to its high effectiveness compared to the other two
measures to the highest risk reduction of approximately 19%. The figure also shows
that the oil change should be performed after approximately 7,500 hours of operation
to obtain the greatest possible risk reduction.
The lubrication of the bearing would be optimal after approximately 2,500 hours of
operation and leads to a risk reduction of approximately 12.5% due to the lower
effectiveness.
The adjusting of the chain reaches the lowest risk reduction of 6.5% compared to the
two other measures due to the lower maintenance effectiveness and should be
carried out after approximately 10,000 hours of operation.
These results are naturally highly depending on determination of the input
parameters (such as α and L factor) for the available maintenance activities. This
determination should be performed in close coordination with the maintenance
experts of the plant.

Figure 4: The effect of a maintenance measure depends on the time of

conduction
 Pipeline Technology Conference 2016

4 Benchmarking of competitive maintenance and inspection activities

In the example shown in chapter 3, only three component-specific maintenance

measures were considered for the risk-based definition of the maintenance program.
But which measure the maintenance expert should go with, especially if he has to
decide whether to conduct only an inspection or additionally also maintenance
measures? Is it better that all considered components will be maintained and
inspected or only dedicated components and if so which components?
A cost-benefit analysis of the possible maintenance programs is helpful to answer
this questions where each component-specific action will be evaluated with regard to
their risk reduction potential (benefit) and compared with the activity-based costs.
For all specific measures a cost-benefit factor (see equation below) can be calculated
which is a measure of the effectiveness of the relevant maintenance program.

ROM  R MM
CBF 
K

ROM stands for the risk without maintenance and RMM for risk taking the appropriate
maintenance action into account. The difference between the two values represents
the maintenance-related risk reduction.
The ratio of the risk reduction (ROM - RMM) and the activity-based cost (K) finally
yields to the cost-benefit factor, which can be used as a performance indicator of the
respective maintenance/inspection activity or even combination.
The results of the cost benefit analysis carried out in the context of the example for 9
different component-specific maintenance programs are shown in Figure 5 (in
descending CBF order).
This graph can be used for efficient risk-based maintenance planning and
determination of priorities for the maintenance program. In this example the highest
cost benefit factor is achieved with an optimal maintenance of the chain from a risk
perspective view. The maintenance measures carried out after risk optimal intervals
for all three components compared to the other maintenance programs has the
highest CBF. In the maintenance schedule also combinations of various maintenance
measures can be taken into account, which will be evaluated with regard to their cost
benefit effectiveness.
In this example the analyzed maintenance programs were consisting of maintenance
and inspections. Other maintenance measures such as, for example, spare parts
storage, redundancy, replacement of old by new components and improvement
measures can be also integrated into the analysis.
 Pipeline Technology Conference 2016

In this example CBF prioritization of maintenance programs has shown that

maintenance measures in combination with a simultaneous conduction of
inspections, independent of the type, have always a lower CBF compared to a
maintenance measure carried out on its own.

Figure 5: Component-related CBF comparison for 9 different maintenance

programs
 Pipeline Technology Conference 2016

5 Availability optimization based on a system concept

 

The previous chapter showed that risk based maintenance is a helpful tool to
evaluate the benefit of different maintenance measures. The final criterion is the cost-
benefit factor which allows the choice of an optimal maintenance measure for a
component. For plants with many different components the risk based approach can
be carried out also for maintenance purposes. A problem which can occur in such
cases is that an isolated view on a single component does not reflect the reality. If
there is a failure at one component this can lead to a shut-off of the complete
machine, plant section or even the facility. This means that based on the failure of
one single component no other component can fail or wear out while the complete
plant or plant section is down. This fact is not accommodated if each component is
reflected separately. One possibility to take this fact into account during risk based
inspection and maintenance planning is a system based view.
The system concept is based on the availability (A) of a plant and can be described
as followed,

where MTTF is the Mean Time to Failure and MTTR is the Mean Time to Repair. This
equation shows that the availability is highly related to the MTTR. If the MTTR is very
low the availability is close to 100 % (A ≈ 1). However when the MTTR of an
equipment is very high the availability goes significantly down.
The availability of a system can be described with the system availability. For a
system with n serial operated components the system availability can be
described:

∙ ⋯

Beside the serial operation also parallel operated components can be described.
The risk of the complete plant can be calculated by using the system availability. For
this example the marginal return (MR) is the only asset number taken into account.
This means if the plant is not available the consequence is only a loss of marginal
return. The system risk is:

For 100% availability of the system the availability ( 1), which means the risk for
the system is 0. A system availability of 0 leads to a risk and therefore to a loss of the
full marginal return. Now the cost benefit factor can be used again to check the
efficiency of different maintenance measures regarding the system availability and
therefore the system risk.
 Pipeline Technology Conference 2016

6 Conclusion

In this article, a risk-based approach has been introduced which allows to optimize
the intervals of different maintenance measures regarding their risk reduction. The
results of this optimization can be used on the one hand for the verification of existing
maintenance programs and on the other hand used by maintenance plans for the
first-time installation.
This approach provides a quantitative basis for maintenance decisions in the
framework of risk-based maintenance planning that can be used in an ideal way for
the tradeoff of business and technical interests and thus the planning and
communication within the orgainisation. It quantitatively represents the benefits of
maintenance and helps to demonstrate these benefits also economically. On the
basis of quantified benefits and reduce risk associated a decision on the optimal use
of ever smaller maintenance budgets are met and communicated to all involved
divisions can be taken.
The system concept approach allows using the method also for very complex plants
with many different components. Thus it is possible not only to decide, which
maintenance measure is the best one, regarding the cost-benefit factor. The system
concept approach also shows for which components in the plant a maintenance
measure has a high cost-benefit factor. This means the system concept approach
allows deciding which components are the most critical ones in a plant. Ultimately,
the decision about the acceptance of the residual risk, including maintenance
programs and associated risk reductions is in the responsibility of the internal risk
management.
RBI may make an important contribution to the establishment and targeted control of
maintenance resources as a "tool" of the maintenance department by these are
focused on the technical high risk areas of the company.
 Pipeline Technology Conference 2016

7 References

[1] “Risk – Based Inspection”; API Recommended Practice 580; First Edition
May 2002; API Publishing Services, 1220 L Street, N.W., Washington, D.C.
20005

[2] „Risk – Based Inspection Base Resource Document“; API Publication 581;
First Edition May 2000; API Publishing Services, 1220 L Street, N.W.,
Washington, D.C. 20005

[3] DIN 31051 – Grundlagen der Instandhaltung, 2012-09

[4] Verordnung über Rohrfernleitungsanlagen (Rohrfernleitungsverordnung),

RohrFLtgV, 2002

[5] TRFL, Technische Regel für Rohrfernleitungsanlagen nach §9 Absatz 5

RohrFLtgV, 2010