MASTER THESIS
6/27/2016
Cost-conscious RCM decision-making models
A Design Science Approach
MSc Technology & Operations ManagementUniversity of Groningen, Faculty of Economics and Business
Public version
Number of words: 10 759Author: D.N. Meeuwsen
Student number: S2314347
Address: Brugstraat 25a
9712AB Groningen
E-mail: d.meeuwsen@student.rug.nl
Supervisors: Dr.ir. W.H.M. Alsem Dr. J. Veldman Jan Flonk
Institution: University of Groningen University of Groningen Enexis
Address: Nettelbosje 2 Nettelbosje 2 Winschoterdiep 50
9747 AE Groningen 9747 AE Groningen 9723 AB Groningen
E-mail: w.h.m.alsem@rug.nl j.veldman@rug.nl jan.flonk@enexis.nl
Acknowledgement
First of all I would like to thank the asset department of Enexis in Groningen and all those interviewed who were willing to share their knowledge with me. With their help I have been able to collect the information required to write this thesis. In particular I would like to thank the gas grid manager Jan Flonk, the senior asset manager Evert van der Zee and the project manager Albert Pondes for involving me in this process and for giving me the opportunities, support and feedback I needed.
Furthermore, I wish to sincerely thank my first supervisor dr.ir. Wilfred Alsem and my second supervisor dr. J. Veldman from the university of Groningen (RUG) for their guidance and feedback but also for their efforts in challenging me to get the best out of this research.
Overall I found writing this thesis an interesting and instructive process and I am pleased with the results, hopefully after reading this thesis we can share the same opinion.
Abstract
The importance of appropriate techniques in RCM decision-making is underscored in literature and hence insufficient attention has been paid to the formulation of a
methodological framework for selecting such appropriate techniques in maintenance decision-making. In addition, RCM focuses mostly on reliability and costs are therefore rather
underexposed and often not explicitly mentioned. The purpose of this thesis is to overcome these shortcomings by developing an RCM decision-making model which also takes costs into account as a decision criterion. Research, both into the available literature as well as into the design science has been conducted at a grid company in order to develop RCM decision-making models which guide practitioners by selecting the most suitable and cost-effective maintenance tasks for a particular failure mode. It turns out that a comparison should be made between the costs of performing the technically feasible maintenance tasks which are able to detect a latent failure or prevent a future failure. Hereby the costs of proactive-, redesign-, restoration- and failure-finding tasks should be considered. If it is also possible not to perform any maintenance at all, a comparison should be made between the costs of repairing/replacing the component that causes the failure including the costs of the production losses. The RCM decision-making model developed for the grid company enables its practitioners to make good and systematic decisions regarding the maintenance and costs of gas stations. By using (an adapted version of) the developed RCM decision-making model, companies are able to select the most suitable and cost-efficient maintenance tasks for a particular failure mode in a consistent and structured way.
Table of Contents
List of abbreviations ... 1
List of figures and tables ... 2
List of figures ... 2
List of tables ... 2
1. Introduction ... 3
2. Theoretical background ... 5
2.1 Asset centric industry ... 5
2.2 Maintenance function ... 5
2.3 Maintenance decisions ... 5
2.4 Situating risk assessment in maintenance decision-making ... 7
2.5 Reliability Centred Maintenance (RCM) ... 7
2.6 The RCM decision-making process ... 8
2.7 Costs in the RCM decision-making process ... 12
3. Methodology ... 14
3.1 Company setting Enexis ... 14
3.2 The Design Science Research process model ... 15
3.3 Validation ... 16
3.4 Methodology steps ... 17
4. Results ... 19
4.1 Extended RCM decision-making model ... 19
4.2 The extended RCM decision-making model applied to Enexis ... 25
5. Discussion of the results ... 28
5.1 Discussion results extended RCM decision-making model ... 28
5.2 Discussion results adapted RCM decision-making model for Enexis ... 29
6. Conclusion and recommendations ... 30
7. Limitations and further research directions ... 33
References ... 34
Appendix A: Definitions maintenance policies and tasks ... 38
Appendix B: RCM tasks per RCM process step ... 39
Appendix C: Decision logic trees ... 41
Appendix E: Gas distribution system ... 46
Appendix F: Risk-matrix of Enexis ... 47
Appendix G: Stakeholder analysis– Enexis ... 48
Appendix H: Description design science research process steps ... 49
Appendix I: Gas station assessment model- Enexis ... 50
1
List of abbreviations
ALARP As Low As Reasonably Practicable BCM Business centred maintenance CBM Condition based maintenance CM Corrective maintenance
FMEA Failure modes and effects analysis IAEA International Atomic Energy Agency LCC Life cycle costing
MSG Maintenance Steering Group
NASA National Aeronautics and Space Administration
PM Preventive maintenance
RBIM Risk based inspection and maintenance RCM Reliability centred maintenance
RPN Risk priority number (risk associated with the failure mode) TBM Time based maintenance
TECOP Technical risks, Economic risks, Commercial risks, Organizational risks and
Political risks
2
List of figures and tables
List of figures
Figure 1 The RCM process (Arno et al., 2010; the department of the Army, 2006). Figure 2 General principle of a decision logic tree (IAEA, 2007).
Figure 3 The RCM decision-making process according to the literature (Arno et al.,
2010; the department of the Army, 2006; IAEA, 2007)
Figure 4 Design Science Research Process Model (McCarty, 1980) Figure 5 The extended RCM decision-making process
Figure 6a The extended RCM decision-making model - part A: Hidden failures
Figure 6b The extended RCM decision-making model - part B: Safety and Environmental consequences (hazardous situation)
Figure 6c The extended RCM decision-making model - part C: Direct and adverse effect on operational capability
Figure 6d The extended RCM decision-making model - part D: No direct and adverse effect on operational capability (but instead on the economic capability) Figure 7 The RCM decision-making model for Enexis
List of tables
Table 1 An overview of maintenance concepts (Waeyenbergh and Pintelon, 2002). Table 2 Decision options in the RCM decision-making process (IAEA, 2007; Moubray,
1997; MSG-s, 1993)
Table 3 Costs in the RCM decision-making process (IAEA, 2007; Moubray, 1997; MSG-s, 1993)
3
1. Introduction
As businesses become more complex, resources become scarcer and margins become razor thin, it is even more critical to release the most possible value from assets by incorporating the right maintenance strategy. A maintenance strategy which can ensure the availability,
reliability and efficiency of plant equipment is by far the most effective way of reducing the risk of serious component failure and unscheduled downtime (Jabar, 2008). Because studies have shown that nowadays major maintenance costs can still be saved, many companies have implemented initiatives in order to optimize their plants’ maintenance function (Jabar, 2008). Reliability centred maintenance (RCM) is a proven technology to optimize maintenance processes and another logical step in improving the overall performance and reliability while reducing the total life cycle cost of an asset (Arno et al., 2015). Optimization of the
maintenance process by the use of RCM is promising due to the fact that physical assets form the basic infrastructure of all businesses and their effective management is essential to overall success (Woodward, 1997). Although RCM involves identifying actions that are both cost-effective and reduce the probability of a failure, costs are still rather underexposed and often not explicitly mentioned. One of the main reasons for this is that RCM focuses mostly on reliability and less on costs (Waeyenbergh and Pintelon, 2002). Reducing costs is of great importance to many companies because it increases a business’ profitability. In this way value for money can be optimized (Hamandani and Khorshidi, 2013).
The underexposure of costs becomes particularly evident when possible maintenance tasks (if any), that could detect a latent failure or prevent a future failure, need to be determined. It turns out that costs only play a role after this process, when the economic impact of the
maintenance tasks need to be estimated. So when the maintenance tasks are being determined, it is not clear which maintenance tasks are most suitable from a cost-effective point of view. Thereby, according to Braaksma et al. (2013) this RCM decision-making process relies for a great part on expert judgement and experience. The importance of appropriate techniques in maintenance decision-making is underscored in literature and insufficient attention has been paid to the formulation of a methodological framework for selecting such appropriate
techniques in maintenance decision-making (Pintelon and Van Puyvelde, 2013; Chemweno et al., 2015). The purpose of this thesis is to overcome these shortcomings by developing an RCM decision-making model which takes costs into account as a decision criterion. As a result, this RCM decision-making model guides practitioners by selecting the most suitable maintenance tasks for a particular failure mode. In addition, decisions about maintenance could be made in a consistent and structured way whereby expert judgment and experience play a less prominent role. Invariably, this enhances the maintenance decision-making process of RCM with regards to selecting appropriate maintenance actions.
4 stations whereby costs are also taken into account as a decision criterion. Because Enexis strives to reduce the yearly costs by 4% in order to remain attractive and offer fair and good prices, costs are important. The following research question has been developed from the above mentioned theoretical and managerial issues:
‘What does an RCM decision-making model look like that explicitly considers costs as a decision-making criterion?’
The essence of this research is to provide practitioners with an extended version of an RCM decision-making model which help them to select the most suitable maintenance task(s) for a particular failure mode in a structured way. Existing RCM decision-making processes are combined and extended in order to integrate costs as a decision-making criterion. Theoretical relevance is hence to map and identify the costs involved in the RCM decision-making process and to integrate costs as a decision-making criterion in an RCM decision-making model. Contribution from a managerial point of view is to provide knowledge about how the RCM decision-making process works in practice. The extended RCM decision-making model is therefore adapted to the specific characteristics of Enexis. In other words, an RCM
decision-making model has been developed for Enexis so its stakeholders are able to make good and structured decisions about maintenance of gas stations.
5
2. Theoretical background
This theoretical background starts with information about the asset centric industry. Hereafter maintenance in general and several maintenance policies and concepts are explained. The next paragraph is devoted to risk assessment because of its crucial and strategic role in
maintenance with regards to mitigating equipment failures. Because risk assessment is embedded in the RCM concept, the next two paragraphs are devoted to RCM and how its decision-making process works in order to select the most suitable maintenance policy for a particular failure. The last paragraph relates costs to the RCM decision-making process.
2.1 Asset centric industry
The industry in general is under increasing pressure to reduce costs, to maximize returns, to meet tougher performance and production targets and to comply with regulatory requirements on assets (Ouertani et al. 2008). On these grounds, asset centric companies like Enexis are looking for opportunities to manage and reduce stakeholder pressure as well as the cost of maintaining their assets; to improve the performance and extend the life of those assets; to speed up information analysis and decision-making processes and to gain competitive advantage throughout the asset life cycle (Chandima et al., 2012). Equipment maintenance and system reliability are important factors that affect the organization’s ability to provide quality and timely services to customers and to be ahead of competition. Maintenance function is hence vital for sustainable performance of any manufacturing plant in the asset centric industry (Muchiri et al, 2011).
2.2 Maintenance function
In an industrial term, the definition of maintenance is often described as an activity carried out on any equipment or system to ensure its reliability in performing its functions. Maintenance philosophies and activities have evolved from simply fixing an equipment malfunction to a more complex and integrated approach (Jabar, 2008). So over the past years, maintenance has become more important in the industry and the role of maintenance has grown into a much more prominent purpose in the plant operation. The same holds for Enexis, where reliability and safety are especially highly valued. For example, three different kinds of inspections are performed by mechanics on gas stations whereby, depending on the type of inspection, different measuring equipment is used. As for every company, maintenance function is also for Enexis vital for sustainable performance.
2.3 Maintenance decisions
6 decisions as strategic in nature. The five main objectives of maintenance are ensuring the plant functionality (availability, reliability, product quality etc.); ensuring the plant achieves its design life; ensuring plant and environmental safety; ensuring cost effectiveness in maintenance and effective use of resources (Muchiri et al., 2011). Once the maintenance objectives are outlined, maintenance strategy formulation (Pinjala, 2008) is necessary to help decide which type of maintenance needs to be done, when to do it, and how often it can be done. According to Pintelon and Van Puyvelde (2006), maintenance decision-making can be broadly explained in terms of maintenance actions (basic elementary work), maintenance policies and maintenance concepts:
Maintenance policies are the rules or set of rules describing the triggering mechanism for the different maintenance actions (Muchiri et al., 2011). Examples of these policies are reactive maintenance, condition based maintenance (CBM), and proactive maintenance (MSG-3, 1993; Arno et al., 2010). A definition of the maintenance policies can be found in appendix A.
Maintenance concepts entail the general decision structure for both maintenance actions and policies and therefore apply to a strategic level (Gits, 1984; Gits, 1992). Examples are
reliability centred maintenance (RCM), total productive maintenance (TPM), life cycle costing (LCC) and business centred maintenance (BCM), see table 1 for the main advantages and disadvantages of these maintenance concepts.
RCM BCM TPM LCC
Advantages
Traceability Accuracy Increased productivity Improvement of the designer-user interface
Cost savings Business-centred approach
Increased quality Life cycle costs is of central importance
Rationalisation Integrating auditing possibilities
Cost reduction Correct adaptation brings considerable benefits in most cases
Plant improvement Increased moral, safety and environmental care
Feedback of information on design
Education Involves the operators Full integration Involves operators and
maintainers
Disadvantages
Complexity Complexity Not really a maintenance concept
Rather theoretical management philosophies
Extensive need of data Extensive need of data
No decision rules for basic maintenance policies
Difficult implementation, life cycle cost analysis is complex Focus on ‘reliability’ Cost and profit are not
taken into account
Less structured No concept
improvement mechanism available
Table 1: An overview of maintenance concepts (Waeyenbergh and Pintelon, 2002).
7 regarding the other maintenance concepts in table 1, costs are positively mentioned. Starting with TPM, where a cost reduction should occur after implementing it. However, a downside of TPM is that TPM is incomplete as a maintenance concept because it does not provide clear rules in deciding which basic maintenance policies will be used (Waeyenbergh and Pintelon, 2001). The next maintenance concept is LCC. Here, life cycle costs are of central importance but its life cycle cost analysis is very complex and less structured. For example in the case of Enexis, all the life cycle costs of gas stations are rather difficult to determine. In addition, this concept is difficult to implement because of constraints on both cash and time, uncertainty of forecasting demand and product life, etc. (Waeyenbergh and Pintelon, 2001). That brings us to RCM, where not ‘costs’ but ‘reliability’ is of central importance (Waeyenbergh and Pintelon, 2002). Although RCM is rather complex and in need of extensive data, costs are saved and it is a structured approach.
As can be concluded from table 1, each maintenance concept has both advantages and disadvantages. RCM is however most in line with Enexis’ maintenance policy due to the previously named arguments and in particular due to the fact that Enexis focuses mostly on reliability (and not on business objectives, costs or productivity) in order to ensure safety and preserve functionality in the most economic manner.
2.4 Situating risk assessment in maintenance decision-making
For operable assets, the operation and maintenance phase is quite critical, often constituting as much as 70% of the assets total cost of ownership (Koronios et al., 2007). Risk management is viewed as rather crucial with regards to mitigating equipment failures. For operable assets, this entails formulating effective maintenance strategies (Chemweno et al., 2015). As such, risk management forms an important aspect in asset management. Risk is a combination of consequence of failure and probability of failure. As part of the risk management process, risks have been identified and evaluated across the 5 TECOP categories; Technical risks, Economic risks, Commercial risks, Organizational risks and Political risks. Perceived risks in the maintenance decision-making domain are largely technological given that maintenance decisions focus on the equipment’s operational phase (Pintelon and Van Puyvelde, 2013). However the other risks (ECOP) could be as relevant and important as technological risks and should therefore not be underexposed by the asset management. So given the strategic
importance of maintenance programs towards sustaining the organizational competitiveness, the role of risk assessment cannot be ignored (Pintelon and Van Puyvelde, 2013). Hence when selecting the right maintenance strategy, risk assessment performs a crucial role by providing an important decision support structure that aids the selection process. Decision support frameworks mentioned in the literature where risk assessment is embedded include the Reliability Centred Maintenance (Moubray, 1997) and Risk Based Inspection and Maintenance (RBIM) (Khan and Haddara, 2003).
2.5 Reliability Centred Maintenance (RCM)
8 static equipment and not really suitable for complex machinery like gas stations (DNV, 2009). In addition, the main failures are loss of containment or structural damage and there are usually only a few failure modes recognized. Failure modes are the ways, or modes, in which an asset can fail (Braaksma et al., 2013). RCM on the other hand is suitable when there are many failure modes and consequences. This also holds for Enexis’ gas stations, where many failure modes are recognized. Lastly, according to Conachey et al. (2008) the input of RCM provides a process feed for optimization of maintenance tasks for maximum uptime. This is incorporated in the RCM decision-making model for Enexis. Based on the above mentioned arguments, RCM will hence be further explained.
RCM is an ongoing process that gathers data on performance and uses this data to improve planning for future maintenance (Sandham, 2013). The concept of RCM has been adopted across several government and industry operations as a strategy for performing maintenance. RCM programs can be implemented and conducted in several ways and use different kinds of information. For example, when focussing on asset information of components that have the highest ‘criticality’ could improve returns on investment (Braaksma et al., 2013). The decision as to how the RCM program is implemented should be made by the end user based on: (1) consequences of failure, (2) probability of failure, (3) historical data (4) risk tolerance (criticality) and (5) resource availability (Pride, 2010). However according to Braaksma et al. (2013) expert judgement is also an important aspect in this decision process.
RCM had its origins in the airline industry and provides a logical structured framework for determining the optimum mix of applicable and effective maintenance activities required to sustain the desired level of operational reliability of systems and equipment. This maintenance method strives to find a balance between system reliability, performance and maintenance costs by taking into consideration many parameters such as the effects of redundancy, spares costs, maintenance labour costs, equipment ageing and repair times (Bertling, 2002). The two primary objectives of RCM are ensuring safety through preventive maintenance actions and preserving functionality in the most economic manner by increasing reliability and
availability (department of the Army, 2006). In other words, RCM seeks to minimize
maintenance and improve reliability throughout the life-cycle (Pride, 2010). Although RCM involves identifying actions that are both cost-effective and reduce the probability of a failure, costs are still rather underexposed and often not explicitly mentioned. One of the main
reasons for this is that RCM focuses mostly on reliability and not on costs (Waeyenbergh and Pintelon, 2002). This becomes particularly evident when possible maintenance actions (if any) that could detect a latent failure or prevent a future failure need to be determined because costs are not mentioned in this process.
2.6 The RCM decision-making process
9 represents the process where decisions need to be made in order to select the most suitable maintenance tasks per failure mode.
Figure 1: The RCM process (Arno et al., 2010; the department of the Army, 2006).
Input is needed in order to be able to make decisions regarding the selection of the most suitable maintenance tasks per failure mode. The data that will be used as input for the maintenance decision-making process is based on the earlier RCM process step ‘apply RCM decision logic’. Here the risk priority number (RPN) for each failure mode and effect is determined based on data about the failure’s severity, probability of occurrence and
detectability. Due to the fact that it is unlikely that all the resources will be available to tackle all improvement opportunities, there is a requirement to prioritize and to take into account several trade-offs. The maintenance tasks with the highest priority numbers, that can detect a latent failure or prevent a future failure, need to be determined first. This is done at the next RCM process step ‘identify maintenance tasks’ where the maintenance tasks for a particular failure mode are identified and screened by the use of a decision logic tree.
10 of equipment (Pride, 2010). See appendix A for a list and definitions of all the possible
maintenance tasks that can detect a latent failure or prevent a future failure. Decision logic trees consist of a series of Yes-No questions. The answers to these questions lead to a specific path through the tree. The questions are structured to meet the objectives of the RCM
analysis: ensure the safe (non-hazardous) and economical operation and support of a product while maximizing the availability of that product. ‘As low as reasonably practicable’
(ALARP) arguments are used to justify the acceptance or rejection of maintenance tasks that lead to risk changes between these limits (French et al., 2005). This objective is met by selecting failure-finding or preventive maintenance (PM) tasks when appropriate, some combination of PM and failure-finding tasks and by corrective maintenance (CM) or redesign when PM and failure-finding tasks are neither applicable nor effective (Department of the Army, 2006; Arno et al., 2010; MSG-3, 1993). Note that not all of the maintenance tasks have been (or even can be) performed. The selected maintenance task(s) are influenced by the failure causes that the analyst determines to be dominant and worthy of prevention. This is in turn dependent on the maintenance philosophy of the company and the cost-effective
application of various maintenance tasks (IAEA, 2007).
Figure 2: General principle of a decision logic tree (IAEA, 2007).
11 After the screening and identification of maintenance tasks, the economic impacts of
performing these tasks need to be determined. This is done in the last process step ‘make recommendations and package final maintenance program or approach’. Here the cost of performing the maintenance in terms of both direct and indirect (loss of production) costs to the potential savings of preventing a failure are compared in order to eventually develop a maintenance tasking schedule for the analysed equipment. So regarding the RCM decision-making process, costs actually only become apparent when the economic impact of the maintenance tasks needs to be determined.
In order to find the decision criteria in the RCM decision-making process, a comparison is made between the decision-making processes (for selecting the most suitable maintenance tasks per failure mode) of IAEA (2007), Moubray (1997) and MSG-3 (1993). It turns out that the RCM decisions making process consists of five main decision categories, each with its own decision options. In table 2 the five decision categories and its decision options are provided.
Decision category Decision options
Kind of failure Evident or hidden
Possible damage of the failure Hazardous (safety or environmental) or non-hazardous Possible operational consequence Adverse and direct effect on operational capability or no
effect on the operational capability (economic effect) Technical feasibility/applicable Feasible or not feasible
Worth doing/effective Cost-effective or not cost-effective
Table 2: Decision options in the RCM decision-making process (IAEA, 2007; Moubray, 1997; MSG-s, 1993)
12 Figure 3: The RCM decision-making process according to the literature (Arno et al., 2010; The department of the Army, 2006; IAEA, 2007; Moubray, 1997; MSG-s, 1993)
2.7 Costs in the RCM decision-making process
In each case, a task is only suitable if it is technically feasible and worth doing. If this is however not the case, the task as a whole is rejected (Moubray, 1997). If a task is selected, a description of the task and the frequency with which it must be done are recorded, along with an estimation of the economic impact of performing the maintenance task. The latter contains a comparison of performing the maintenance task in terms of both direct and indirect (loss of production) costs to the potential savings of preventing a failure. It turns out that costs only play a role at the last RCM process step where the economic impact of the maintenance task needs to be determined. This means that only the costs of the maintenance tasks which are suitable for a given scenario need to be compared in order to see which maintenance task is most economical. In table 3 an overview and description of the possible costs that could be eligible in the RCM decision-making process (for selecting the most suitable maintenance tasks per scenario) is provided.
Costs Description
Proactive maintenance costs
- Lubrication and/or servicing task costs
- Scheduled on-condition task costs
- Scheduled discard task costs
All the costs that appear when a preventive, condition-based or predictive maintenance strategy is performed to stabilize the reliability of equipment or machines.
Failure-finding costs
- Operational and/or visual check costs
- Inspection and/or functional check costs
All the costs that appear when a failure-finding maintenance strategy is performed to reveal hidden failures that have already occurred.
Redesign/replacement costs
- Redesign task costs
- Replacement task costs
All the costs that appear when a redesign is performed. These entail the costs of engineering effort, tooling modifications, and changes in production labour, materials, and overhead. Plus all the costs that appear when a replacement is
What:
A list of the RPN’s ranking from most risky to least risky.
Method:
Per failure mode and effect: severity * probability of occurrence * detectability.
What:
Screen and determine maintenance task(s) per failure mode.
Method:
Decision logic tree.
Input Process
What:
Determine the economic impact of performing the maintenance task(s). Method: Estimation based on experience. Process Maintenance tasking schedule for the analyzed equipment.
Output
Options:
- Proactive maintenance tasks
- Redesign/replacement maintenance tasks - Failure-finding maintenance tasks - Restoration maintenance tasks
Options:
- Kind of failure
13 necessary.
Repair costs
- Restoration task costs
All the costs that appear when a repair maintenance strategy is chosen. Restoration tasks keep the asset operating at its present condition or to bring an asset back to an earlier condition.
Table 3: Costs in the RCM decision-making process (IAEA, 2007; Moubray, 1997; MSG-s, 1993). There is also an option not to perform any maintenance at all, this is called a ‘run-to-failure’ maintenance strategy and results only in corrective maintenance tasks. Therefore, the costs of performing a run-to-failure strategy eventually result in restoration tasks costs.
Note that the costs, which should be taken into account in the RCM decision-making process, depend on the situation. For example, if a maintenance task is not technically feasible, the costs of performing this task should not be compared to technically feasible maintenance tasks because it does not make sense to compare an option that is not eligible for a particular
situation. On the other hand, if both proactive tasks and failure-finding tasks are suitable for a particular situation and time period, then the option should be chosen that is most economical. So, when the maintenance tasks for a failure mode need to be determined, costs are
underexposed and often not even explicitly mentioned. This is noteworthy because here a great part of the decision which maintenance to perform for a particular failure mode is determined. In other words, only after screening the maintenance options to see which are suitable for a particular situation are costs being compared. This happens in the last RCM process step where the economic impact of the eligible maintenance tasks is estimated. It turns out that then a comparison is made between the costs of (1) proactive tasks, (2) redesign tasks, (3) restoration tasks, and (4) failure-finding tasks. If it is however also possible to perform no maintenance at all, then a comparison should be made between the costs of
14
3. Methodology
3.1 Company setting Enexis
The main goal of this research is to analyse the decision-making process of RCM and which costs need to be considered in this process. An extended RCM decision-making model is developed, the most suitable maintenance actions among several scenarios are selected and costs are explicitly mentioned. This model is then adapted to the company Enexis in order to provide their stakeholders with an RCM decision-making model which enables them to make good and systematic decisions regarding maintenance of gas stations.
This research will employ a study at the Dutch energy and electricity distribution company ´Enexis´. Enexis is the second largest grid company in the Netherlands with a total of 2.7 million customers and owns approximately 244.000 gas stations. The revenue in 2015 was 1.35 billion euros and about 4,000 employees are working for Enexis. The six strategic pillars of Enexis are (1) reliability, (2) safety, (3) legislation, (4) affordability, (5) customer
satisfaction, and (6) sustainability.
15 rarely happens) or by inspections. How often such inspections need to be done depends on (1) long and short term plans, (2) budget, (3) experience, and (4) failure probabilities (which in turn are based on Enexis’ risk-matrix, see appendix F).
The asset management of Enexis declared that it is hard to determine what the optimal
maintenance policy of the gas stations is and doubts whether the double equipped gas stations runs at the expense of too high costs. On the other hand, the management also doubts whether non-redundant gas stations and networks are technically feasible, safe enough and cheaper than double equipped gas stations. Besides taking redundancy into account, other
characteristics also need to be considered as well e.g. if a failure is evident or not or if it is cheaper to replace components or instead renew a whole gas station. See appendix G for a stakeholder analysis and who is authorized to take which decisions about gas stations and/or gas station components.
Last but not least, routines are so common at Enexis that there is no systematic and structured decision-making model available telling practitioners the needed maintenance actions for a particular situation. Asset management wonders what a RCM decision-making model, which guides practitioners by choosing the most suitable maintenance task for a particular situation and also considers costs, looks like.
3.2 The Design Science Research process model
Before starting to conduct research, it is crucial to know what kind of research is most appropriate. Research can be very generally defined as an activity that contributes to the understanding of a phenomenon (Vaishnavi and Kuechler, 2015). Both literature research and design science research are most appropriate because of developing improvement solutions to problems based on existing literature. Thereby, design science is an explorative way of doing research (Holmstrom et al., 2009). The aim of design science is to serve as the archival venue of science-based design knowledge across multiple disciplines. Design science research has dual goals of artefact development and knowledge production and is both iterative and incremental of nature (Baskerville et al., 2015). So, a design problem has a problem-solving paradigm and is a construction of innovative solutions to practical problems. Design science exists of two types (Hietschold et al., 2014). Type one leads to a theory or theoretical results. Type two leads to validated design principles. In this case we are dealing with design science type two because an individual artefact is designed in order to improve some context.
Artefacts are not given, rather they need to be designed to improve or influence a given problem context in order to contribute to stakeholder goals. So the aim of an artefact is that it interacts with the problem context in order to improve the context. This is exactly the purpose of the extended RCM decision-making model developed and then adapted to Enexis specific characteristics.
In figure 4, a model of the general design science research process is described. Even though the different phases in a design process and a design science research process are similar, the activities carried out within these phases are considerably different. There are many design
16 model of McCarty (1980) is used because of its completeness; at every process step the
corresponding knowledge flows and outputs are given as well.
A description of the five design science research process steps can be found in appendix H. This research focuses however on the first three steps of the design science research process whereby in step three an extended and an adapted RCM decision-making model is developed as artefacts. The fourth and fifth step is beyond the scope of this research, thus not discussed and elaborated further in much detail. The developed artefacts in step 3 are models which consist of several steps. A model is a set of propositions or statements expressing
relationships among constructs. They are proposals for how things are or should be (Vaishnavi and Kuechler, 2015).
3.3 Validation
Although the evaluation phase is beyond the scope of this research, the model will
nevertheless be partly validated by asking stakeholder(s) if the RCM decision-making model is sufficient accuracy and built for the purpose at hand. By validating the model, is it more surely that it will actually be used when the implementation phase takes place.
Validation is an important aspect in design science research (Van Strien, 1997; Wieringa,
2009).There are two key concepts in validation: the ideas of sufficient accuracy and models
that are built for a specific purpose (Carson, 1986). So, the purpose or objectives of a model must be known before it can be validated. Validity exists of internal and external validity. External validity shows to what degree the solution design can be generalized beyond the immediate case and internal validity shows in which degree the solution design indeed satisfy stakeholders’ goal and criteria (Voss et al., 2002).
17 developed model. An expert could have other or additional ideas regarding the model than the stakeholders have and could therefore be valuable. At the end, validity is a binary decision with a conclusion of ‘yes’ or ‘no’ (Robinson, 2008). In the discussion section is determined if the model is valid and thus sufficiently accurate for the purpose at hand.
3.4 Methodology steps
The methodology is explained by means of thirteen methodology steps in order provide a guideline how to generate a substantiate answer on the research question. In table 4 a
distinction is made between theoretical and managerial steps in order to be able to develop an extended RCM decision-making model and then adapt this model to the specific
characteristics of Enexis.
Step Theoretical (literature) Managerial (Enexis)
1 Determine the type of research and why a
study at Enexis is relevant for this research.
Design science steps ‘awareness of problem and suggestion’.
2 In-depth analysis of how the RCM process looks like and how it works, including what all the steps in the RCM process entail.
3 Determine where in the RCM process
decisions need to be made in order to select the most suitable maintenance tasks per failure mode. Determine on what criteria those decisions are based.
4 Determine what RCM decision-making models already exist.
5 Determine which costs are considered in the RCM decision-making process.
6 Determine the current maintenance
management policy used at Enexis. Design
science steps ‘awareness of problem and suggestion’.
7 Determine how the general decision structure
at Enexis looks like and who is involved in that process. Also determine which
stakeholders and decision makers specifically have to deal with the gas stations by
conducting a stakeholder analysis. Design
science steps ‘awareness of problem and suggestion’.
8 Determine what decision-making criteria
Enexis uses regarding the maintenance of gas stations. Design science steps ‘awareness of
problem and suggestion’.
9 Determine which costs need to be taken into
account regarding the decision-making process of gas stations. Design science steps
18 10 Make a comparison between what is stated in the literature about the RCM decision-making
processes and the costs to be considered. Draw a conclusion about the differences/similarities in order to be able to develop a systematic and structured RCM decision-making model whereby costs are explicitly mentioned among several scenarios. This extended decision-making model will help to decide which maintenance actions practitioners need to make.
Design science step: ‘development’.
11 Adapt the model that is developed in step 10 to Enexis specific characteristics (e.g. redundant gas stations or no redundant gas stations). In this step an RCM decision-making model for Enexis’ gas stations is developed whereby the most suitable maintenance task for a particular situation is selected in a structured way. Design science step: ‘development’.
12 Discuss and validate the developed RCM decision-making models. Design science step
‘evaluation’.
13 Provide recommendations, a conclusion and further research directions. Table 4: Methodology steps
Literature and design science research will both be conducted in order to generate an integral conclusion. Differences and similarities are compared in order to give an answer on the research question from a theoretical and managerial point of view. The theoretical
methodology steps (steps 2-5) are containing a literature research and hence can be found in the theoretical background section. Those steps are based on reading and comparing
numerous articles and other relevant information. The managerial methodology steps on the other hand are all based on semi-structured interviews with stakeholders (see appendix G for a stakeholder analysis) and relevant data about Enexis and its gas stations. Relevant data could for example be data about the replacement and repair costs of gas stations’ components. The step 1 and 6-9 are all part of the design science steps ‘awareness of problem’ and
‘suggestion’ because the output of the ‘awareness of problem’ step is a proposal and is intimately connected to the suggestion phase. The content of those steps can be found in paragraph 3.1 where the case selection and setting is described. Steps 10-12 on the other hand are based on both theoretical and managerial input and therefore form the basis of the
19
4. Results
4.1 Extended RCM decision-making model
In this paragraph an extended RCM decision-making model is developed which also takes costs into account as a decision criterion. Like explained in the theoretical background
section, it turns out that costs are not explicitly mentioned when the maintenance tasks (which could detect the latent failure mode or prevent the failure mode from happening) are screened and identified. Only hereafter the economic impact will be estimated. So the RCM decision-making process actually consists of two process steps: (1) the screening and determination of the maintenance tasks per failure mode and (2) the economic impact of performing the particular maintenance tasks. However, in the extended RCM decision-making model, the screening and identification of the maintenance task and the economic impact of these are simultaneously determined. In other words, costs are explicitly mentioned and already integrated when the maintenance tasks are being screened and identified. So in the extended RCM decision-making model, costs are integrated, explicitly mentioned and already
compared to each other to see what the most economic maintenance tasks are. Hence, the determined maintenance task for a specific failure mode is not only technically feasible but also most cost-effective. The output of the extended RCM decision-making model is therefore not just the kind of maintenance tasks (that could detect a latent failure or prevent a future failure), but already the most suitable maintenance task and the costs of performing it. This is further explained and illustrated in figure 5.
Figure 5: The extended RCM decision-making process
What:
A list of the RPN’s ranking from most risky to least risky.
Method:
Per failure mode and effect: severity * probability of occurrence * detectability.
What:
Screen and determine maintenance task(s) per failure mode and determine the economic impact of the determined maintenance task(s) simultaneously
Method:
Extended RCM decision-making model. See figure 6.
Input Process
Maintenance tasking schedule for the analyzed equipment.
Output
Options:
- Proactive maintenance task(s) costs
* Scheduled on-condition task(s) costs
* Scheduled lubrication and/or servicing task(s) costs * Scheduled discard task(s) costs
- Redesign maintenance task costs
* Scheduled replacement or redesign task(s) costs - Failure-finding maintenance task(s) costs
* Operational and/or visual check task(s) costs * Inspection and/or functional check task(s) costs - Restoration maintenance task costs
* Restauration tasks(s) costs
20 As can be concluded from Figure 5, the input for the extended RCM decision-making model is the same as the input for the RCM decision making process according to the literature: a list of the risk priority numbers (RPN’s), ranking from most to least risky. The RPN’s are again based on the severity, probability of occurrence and detectability per failure mode. After having determined the RPN’s, the failure modes with the highest risk numbers should be the input for the extended RCM decision-making model. The extended RCM decision-making model is given in figure 6 and exists of four parts. Like in the RCM decision models of Moubray (1997) and MSG-3 (1993), the extended RCM decision-making model is also divided into four parts where a distinction is made between (1) kind of failure, (2) possible damage of the failure, (3) possible operational consequence, and (4) possible economic consequence. Those four distinctions are illustrated by the letters A-D. Each letter has other characteristics whereby different maintenance policies should be performed. For example, failure-finding tasks are only eligible by a hidden failure and component(s) that results in a failure having direct and hazardous effects could only be solved by either a combination of PM tasks or a redesign/replacement.
In order to make the extended RCM decision-making model as complete and detailed as possible, all the maintenance tasks that could detect a latent failure or prevent a future failure should be identified. This is done by looking to all the maintenance tasks already existing in the current RCM decision-making processes. Those are compared and combined to eventually integrate all the relevant maintenance tasks into one decision-making model. For example when dealing with a hidden failure (see part A of figure 6), the costs of preventive
maintenance tasks which are technically feasible are first compared to each other. The most cost-effective preventive maintenance task is then compared to the most cost-effective failure-finding task that is technically feasible etc. In the end, the output of the extended RCM decision-making model is/are maintenance task(s) that are most suitable for a particular failure mode. With the term ‘most suitable’ is meant that the determined maintenance tasks are both technically feasible and most cost-efficient. Hence the costs of the possible
21
Will the loss of function caused by this failure mode on its own become evident under normal circumstances?
Scheduled on-condition task costs
Is the proactive task to detect whether the failure is occurring or about to occur technically feasible and cheaper than a scheduled
lubrication and/or servicing tasks, a scheduled discard task or one of the failure-finding tasks?
Go to part B*
No
Is a scheduled lubrication and/or servicing task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled discard task or one of the failure-finding tasks?
Scheduled lubrication and/or servicing task costs
Is a scheduled discard task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled lubrication and/or servicing task or one of the failure-finding tasks?
Scheduled discard task costs
Is an operational and/or visual check task to detect the failure technically feasible and cheaper than an inspection and/or functional check task?
Costs of operational and/or visual check task(s)
Figure 6a: The extended RCM decision-making model - part A: Hidden failures
Is an inspection and/or functional check task to detect the failure technically feasible and cheaper than a
combination of PM and failure-finding tasks, a replacement/redesign and a restauration task?
Costs of inspection and/or functional check task(s) No No No Yes Yes Yes Yes Yes Yes No Replacement/redesign costs Run-to-failure approach: eventually leads to restauration tasks costs
Could the failure have a direct and hazardous effect?
Ye
s
No
Is a combination of PM and failure-fining tasks technically feasible and cheaper than a replacement/redesign? No Ye s The costs of a combination of PM and failure-finding tasks Part A No
Is a restauration tasks cheaper than a combination of PM and failure-finding tasks and a replacement/redesign?
Ye
s
22
Does the functional failure or secondary damage resulting from this failure have a direct and hazardous effect?
Figure 6b: The extended RCM decision-making model - part B: Safety and Environmental consequences (hazardous situation)
Scheduled lubrication and/or servicing task costs
Is a scheduled discard task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled lubrication and/or servicing task, a combination of PM tasks and a redesign/replacement?
Scheduled discard task costs
Is a combination of proactive tasks technically feasible and cheaper than a replacement/redesign? Costs of a combination of PM tasks No No Yes Yes Yes Replacement/redesign costs No No Ye s *Part B Go to part C **
Yes Scheduled on-condition task costs
No
Is the proactive task to detect whether the failure is occurring or about to occur technically feasible and cheaper than a scheduled lubrication and/or servicing tasks a scheduled discard task a combination of PM tasks and a redesign/replacement? ?
23
Does the failure mode have a direct adverse effect on the operational capability? **Part
C
Figure 6c: The extended RCM decision-making model - part C: Direct and adverse effect on operational capability No Ye s Scheduled on-condition task costs
Is the proactive task to detect whether the failure is occurring or about to occur technically feasible and cheaper than a scheduled lubrication and/or servicing tasks, a scheduled discard task and cheaper than the cost of the operational consequences plus the cost of the repair over the same time period??
Is a scheduled lubrication and/or servicing task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled discard task and cheaper than the cost of the operational consequences plus the cost of the repair over the same time period??
Scheduled lubrication and/or servicing task costs
Is a scheduled discard task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled lubrication and/or servicing task and cheaper than the cost of the operational consequences plus the cost of the repair over the same time period?
Scheduled discard task costs
If no scheduled maintenance is performed, are the operational consequences then still acceptable? No No Yes Yes Yes Yes Replacement/redesign costs No Run-to-failure approach: eventually leads to restauration tasks costs
24
Part D***
Does the failure mode have a indirect adverse effect on the operational capability? (This indirect means that it instead has an effect on the economic capability)
No
Is the proactive task to detect whether the failure is occurring or about to occur technically feasible and cheaper than a scheduled lubrication and/or servicing tasks, a scheduled discard task and the cost of repair or replacement/redesign over the
same time period?
Yes
Scheduled on-condition task costs
No
Is a scheduled lubrication and/or servicing task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled discard task and the cost of repair or replacement/redesign over
the same time period?
Yes
Scheduled lubrication and/or servicing task costs
Is a scheduled discard task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled lubrication and/or servicing task and the cost of repair or replacement/redesign over the
same time period?
No
Yes
Scheduled discard task costs
Are the repair costs higher than the costs of a
replacement/redesign over the
same time period?
Yes
No
Run-to-failure approach: eventually leads to restauration tasks costs
Replacement/redesign costs
25 4.2 The extended RCM decision-making model applied to Enexis
In this section a systematic and structured decision-making model for Enexis is developed so its stakeholders can use a common language regarding which maintenance tasks to perform in what situation. Because this model is derived from the previous extended RCM decision-making model, costs are automatically included.
Like in the RCM decision-making processes according to the literature, the failure modes occurring at gas stations are also ranked from most to least risky. Enexis has a risk-matrix which determines the risk priority numbers of their failure modes and effects (see appendix F for Enexis’ risk-matrix). So the input of the extended and existing RCM decision-making processes corresponds to the input of the adapted RCM decision-making process for Enexis. Although the input is comparable, the decision-making processes itself continuous slightly different due to Enexis specific characteristics.
First of all, in the adapted RCM decision model for Enexis, a main distinction is made between a redundant system and a non-redundant system. Enexis describes a system as redundant if there is overcapacity. This is of importance because the eligible maintenance tasks depend on a systems’ redundancy. For both scenarios (the scenario where gas stations are redundant and the scenario where gas stations are not redundant) an own RCM decision-making model could be developed. However, it is decided to combine both scenarios into one model in order to provide a clear overview of all the possibilities.
Both scenarios start with the distinction whether a failure is evident or hidden. Hereafter both scenarios continue in a different way. When dealing with a redundant system, corrective and/or preventive maintenance could take place and when having a non-redundant system, preventive maintenance should be performed. At a non-redundant system, failures should be prevented because there is no overcapacity. So there is a problem if a failure occurs at a non-redundant system because then the gas supply will be intermitted most of the time. Preventive maintenance therefore should, like the name says, prevent the gas supply from intermitting by preventing the failure from occurring. When looking to the redundant system, there is no immediate need to repair or replace the component that has failed, unless it results in a direct and hazardous effect. Because of its redundancy, the gas supply is almost never intermitted if a failure occurs and hence most of the time no further complications occur. However, the gas supply could still be intermitted if one day the other gas station breaks down as well, therefore the repair or the replacement of the component that has failed should be scheduled
nevertheless.
The different maintenance tasks, for detecting a latent failure or preventing a future failure, for gas stations are again corresponding to the possible maintenance tasks earlier determined and used in the extended RCM decision-making model. The eligible maintenance tasks in the RCM decision-making model for Enexis are hence: (1) proactive tasks, (2)
26 For the gas stations which are still in operation, an inspection and maintenance plan over a longer period of time is established. The determination of the maximum repair costs as a percentage of the replacement costs for a particular additional time period is part of this long-term plan. When it turns out that a component can be both repaired or replaced, a guideline is provided which tells practitioners whether it is more cost-efficient to replace or repair the component. This guideline is determined by both material and gas station experts and data of 2650 gas stations (which is approximately 21.5% of the total number of gas stations).
Although the maximum repair costs as a percentage of the replacement costs for a certain additional life is based on an estimation and exceptions occur, this method will be
incorporated in the RCM decision-making model for Enexis because the general principle behind the model (repair the component if repair costs < replace costs over the same period of
time; replace the component if replace costs < repair costs over the same period of time) holds
for Enexis. For example, if the additional life time of a component can be extended with 15 years, the repair costs for that component should be maximum 45% of the costs to replace that component. In the RCM decision-making model for Enexis, the maximum repair costs as a percentage of the replacement costs are given for 5, 10, 15, 20 and 25 additional years. The values of those addition years for gas station components are common and therefore used in the RCM decision-making model for Enexis. Due to the fact that the formula is given (repair costs < replace costs * (1- net present value (remaining life)), the percentage for other
additional years can be calculated as well.
What also is incorporated in the adapted RCM decision-making model for Enexis, is the option to renew a whole gas station. This should be done if it turns out that this option is more cost-effective than replacing/repairing gas stations’ components. Enexis already provided a model with some guidelines whether this could (or should) be the case, see appendix I. This should for example be the case when there are legal restrictions and if there are no spare parts for components available anymore.
27
* A repair or replacement task depends on the maximum component repair costs as percentage of the replacement costs for the same time period. In formula form with a discount rate of 4%: repair costs < replace costs * (1- net present value(remaining life)).
Additional 5 years Additional 10 years Additional 15 years Additional 20 years Additional 25 years
Repair cost < 18% of the replace cost
Repair cost < 33% of the replace cost
Repair cost < 45% of the replace cost
Repair cost < 55% of the replace cost
Repair cost < 63% of the replace cost
No red uce the fail ure rate to an acc epta ble leve l o Yes Yes No red uce the fail ure rate to an acc epta ble leve l o No redu ce the failu re rate No redu ce the failu re rate to an acce ptab le leve l o Yes Yes Yes Yes Yes
Figure 7: The RCM decision-making model for Enexis
Yes No Yes No Yes No Yes No
Redundant gas stations?
Preventive maintenance Preventive/Corrective
maintenance
Evident failure?
Is one of the technically feasible PM tasks cheaper than one of the technically feasible failure-finding task and a restauration task? Does the functional failure or
secondary damage resulting from this failure have a direct and hazardous effect?
Run-to-failure approach. (Eventually leads to restauration tasks costs). Immediate repair or replacement of the component(s). * Scheduled repair or replacement of the component(s). * Evident failure? Immediate repair or replacement. *
Is the proactive task to detect whether the failure is occurring or about to occur technically feasible and cheaper than a scheduled lubrication and/or servicing tasks, a scheduled discard task and a scheduled replacement/redesign? Scheduled
on-condition task costs.
Is a scheduled discard task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled lubrication and/or servicing tasks and a replacement/redesign task?
Scheduled discard task costs.
Scheduled replacement/redesign task costs. Is a scheduled lubrication and/or servicing task to reduce the failure rate to an acceptable level technically feasible and cheaper than a scheduled on-condition task, a scheduled discard task and a scheduled replacement/redesign? Scheduled lubrication and/or servicing task costs. Yes No red uce the fail ure rate to an acc epta ble leve l o
Perform the cheapest PM task that reduces the failure rate to an acceptable level. No redu ce the failu re rate to an acce ptab le leve l o Perform the cheapest failure-finding task.
Is one of the technically feasible failure-finding tasks cheaper than a restauration task?
Is renewing a whole gas station cheaper than repairing/replacing the component(s)? Is renewing a whole gas
28
5. Discussion of the results
In this section the results are discussed. First the results of the extended RCM decision-making model are discussed and thereafter the results of the adapted version for Enexis.
5.1 Discussion results extended RCM decision-making model
The extended RCM decision-making model is different from the already existing RCM decision-making models because more specific maintenance options are incorporated and costs are explicitly mentioned. In the already existing RCM decision-making processes, the screening and identification of the maintenance tasks is a different part than the estimation of the economic impact of the maintenance tasks. The only monetary aspects mentioned when the maintenance tasks are screened and identified is if the maintenance task is ‘worth doing’ or ‘effective’. However, it is not clear when this is the case and no explicit decision criterion is given. In the extended RCM decision-making model on the other hand, this is stated explicitly by already mentioning and comparing the costs of the different maintenance tasks. In the end, the maintenance task is chosen which is technically feasible and most cost-efficient for a specific failure mode. The second main difference is that several maintenance tasks are combined and represented into one model. In the already existing RCM decision-making processes, less maintenance tasks are mentioned. In other words, there are fewer maintenance tasks options and most of the time they are less detailed. For example in the RCM decision model of Moubray (1997) a scheduled lubrication and/or servicing task is not mentioned. In the RCM decision model of MSG-3, a scheduled on-condition task is not mentioned and in the RCM decision models of Moubray (1997) and IAEA (2007) there is no distinction between failure-finding tasks. Due to the fact that the extended RCM decision-making models’ external validity is high, one of its main advantages is that it can be generalized beyond the immediate case of Enexis. Hence more companies can implement such an RCM decision-making model.
A point of discussion regarding the extended RCM decision-making model is that the
structure of this model is quite similar to some of the already existing RCM decision-making models (especially to the models of Moubray (1997) and MSG-3 (1993)). Thereby, the models of Moubray (1997) and MSG-3 (1993) are for literary concepts rather old and therefore its current relevance could be discussed. On the other hand, the more recent RCM decision-making models (e.g. Afefy 2010 and NASA 2008) also were taken into account by the development of the extended RCM decision-making model. However, those were less applicable and omitted due to the fact that the decision-making model of Afefy (2010) does not make a distinction between hidden or functional failures and model of NASA (2008) does not make a clear distinction between the different PM tasks and first considers a redesign instead of several PM tasks. This last reason is especially important because the RCM decision-making process considers maintenance above redesign (Moubray, 1997).
Nevertheless a critical note regarding the current relevance of the decision-making models of Moubray (1997) and MSG-3 (1993) should be made.
