Available online at www.sciencedirect.

com

ScienceDirect
Procedia Manufacturing 11 (2017) 1135 – 1144

27th International Conference on Flexible Automation and Intelligent Manufacturing, FAIM2017,

27-30 June 2017, Modena, Italy

A conceptual framework of knowledge conciliation to decision

making support in RCM deployment
Flávio Piechnickia,b, Eduardo Louresa, Eduardo Santosa*
a
Pontifical Catholic University of Parana, Curitiba, 80215-901, Brazil
b
Federal Institute of Parana, Telêmaco Borba, 84269-090, Brazil

Abstract

This paper proposes a conceptual framework that conciliates tacit and explicit information from the maintenance function,
generating a new knowledge base used in analyzing and improving decisions in deploying a customized RCM (Reliability
Centered Maintenance) model. The transformation of raw information into formal knowledge must generate personalized records
in a single database, being available for the RCM deployment phases. By identifying trends and applying Process Mining
techniques, hidden patterns and relationships can be uncovered. MCDM/A (Multi Criteria Decision Making/Analysis) methods
support the decisions in the stages of RCM implementations. Improving maintenance strategies is an important approach in
increasing system reliability and reducing costs.

©©2017
2017TheTheAuthors. Published
Authors. by Elsevier
Published B.V.B.V.
by Elsevier This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee
Peer-review under responsibility of the scientific committee of theof theInternational
27th 27th International Conference
Conference on Flexible
on Flexible Automation
Automation and and
Intelligent Manufacturing.
Intelligent Manufacturing

Keywords: Maintenance Management; Reliability Centered Maintenance; Knowledge Management; Multi Criteria Decision Making/Analysis;
Maintenance Strategies.

* Corresponding author. Tel.: +55-42-99969-0056; fax: +55-42-3221-3000.

E-mail address: flavio.piechnicki@ifpr.edu.br

2351-9789 © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 27th International Conference on Flexible Automation and Intelligent Manufacturing
doi:10.1016/j.promfg.2017.07.235
1136 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144

1. Introduction

The current industrial systems must ensure equipment uptime and reliability, in order to achieve the predefined
goals and parameters. In many manufacturing environments, the condition of the equipment or process has
significant impact on the quantity and quality of units produced [1]. In this context, the maintenance function aims
to increase the equipment lifetime, or at least the mean time between failures, with possible costly repair.
Furthermore, the maintenance policy is expected to be effective and reduce the frequency of job interruptions and
their unwanted effects for the process as a whole.
Researchers around the world are striving to develop and apply new techniques and methodologies for the
improvement of maintenance processes. The goal is to use them for optimal maintenance policies, impacting
positively on MPIs (Maintenance Performance Indicators), increasing quality and productivity and reducing costs.
Methods like Decision Making [2], Statistics [3], Failure analysis [4], Fuzzy logic [5], Modeling [6], among
others are, most of the time, employed, in isolation, addressing the use of tacit or explicit information, with
qualitative or quantitative approaches. Some efforts are made to achieve collaborative maintenance information
integration, exploring frameworks from the standpoint of development for maintenance functions [7,8].
Industrial systems have different levels of maturity, in organizational, structural, technological and human
resources aspects, requiring customized database models that can provide maximum information availability and
generating appropriate knowledge bases. Companies must evolve from a classical data management approach to
using information and knowledge as critical business assets. Quality practices and data management are still
essential, but these practices must be improved to meet the demands of the environment and of the business, with an
appropriate cost-benefit ratios.
The RCM methodology emerges in this scenario providing a framework capable of reducing maintenance
activities and their related costs as far as possible without affecting the overall performance of the plant, product
quality, safety or environmental integrity [9]. However, for a successful deployment of RCM, aspects of each
conceptual model and the respective specific variables of the industrial system selected should be addressed. This
perception becomes difficult given the diversity of technologies involved in the MMIS (Maintenance Management
Information Systems), EAM (Enterprise Asset Management), the presence of Industry 4.0 aspects, the level of
automation systems, organizational characteristics and the maintenance structure as a whole.
Thus, the maintenance information databases must be customized and applied according to the plant resources
available. To support this, this research proposes a custom framework for the provision of tacit and explicit data
applicable in RCM deployments. Figure 1 presents the general research scope where KDD (Knowledge Data
Discovery) is the hub of the elements involved.
CMMS (Computerized Maintenance Management Systems) manage the life cycle of maintenance work-orders in
accordance with the maintenance policy selected (containing knowledge about maintenance actions, past and
planned, degradation models, boundary thresholds) [10]. Measurement metrics are not adapted to real needs, which
have a strong human factor; nor is there a roadmap of the amount of data to be collected, their processing or how
they are applied in decision making [11].
Thus, the goal of this study is to create a conceptual framework to address tacit and explicit data in a single
database, in order to support the decision dimensions of RCM deployments. This base should be explored in search
of consistent patterns, detecting systematic relationships between variables. By identifying trends in maintenance
actions, hidden patterns and relationships can be discovered that will support maintenance decision-making, making
companies more productive, increasing quality and reducing costs. In this context, Process Mining techniques can be
applied for validation.
The research herein is structured as follow: section 2 introduces a review of the paper’s main terms and
discussing general concepts including the RCM methodology, information and knowledge approaches, data sources,
process mining and decision making. Section 3 exposes the proposed RCM deployment, mapping data and
information types, under tacit and explicit aspects. The section introduces RCM decision inputs and possibilities of
MCDMs tools and Process Mining approaches. Finally, section 4 presents the conceptual framework to support
RCM deployment model, from the standpoint of data fusion, process mining and decision making techniques.
 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144 1137

2. Background

This section presents important research topics, introducing relevant concepts around the proposed RCM
deployment framework.

2.1. Reliability Centered Maintenance

Contemporary approaches involving reliability concepts have been widely used in corporate maintenance, with
the aim of increasing equipment uptime. The intention is to maintain the plant’s equipment park at the highest
possible level of availability (uptime). For this, reliability management is used to eliminate the effects of system
failures, caused by equipment problems and often by human factors. In its most usual definition, reliability is the
probability that a component will not fail in performing its function within the threshold boundaries specified by the
production system [12-16]. Indicators used to measure the equipment reliability can be summarized applying the
observed failure rate and the mean time between failures (MTBF). The failure rate is calculated as the total number
of failures divided by the total operating time and the MTBF is the inverse of the failure rate [17].
In this context, the Reliability Centered Maintenance (RCM) emerged in the airline industry in the 1960s as a
methodology capable of determining the most effective maintenance strategy, reducing maintenance activities and
related costs without affecting plant performance, product quality, safety or environmental integrity [9]. Since then,
several standards applicable in different segments have been created and adapted over time. In a contemporary
approach and in industrial settings, the minimum criteria that RCM processes must meet are basically defined by
three technical standards: (i) IEC 60300-3-11 [18]; (ii) SAE-JA1011 [19]; (iii) SAE-JA1012 [17,20]. The approach
is comprised of concepts in Preventive Maintenance (PM), Predictive Maintenance (PdM), Real-time Monitoring
(RTM), Run-to-Failure (RTF) and proactive maintenance techniques, with the objective of reducing the probability
that a machine or component will fail during its life cycle with a minimum of maintenance interventions.

2.2. RCM approaches

Asset reliability is determined in the design phase. Deficiencies can be detected in the design, installation,
commissioning and operation phases, due to human errors in decision making, generating consequences. Skills and
behavioral factors can affect the performance of the equipment and of the production process [21]. The reliability of
assets depends on the reliability of people, production processes and the equipment used in it. Qualitative
information is embedded throughout maintenance management. Quantitative information is generated, for the most
part, by systems and processes. A recent research [22] performs a SWOT (Strengths, Weaknesses, Opportunities and
Threats) analysis of RCM frameworks between 1978 and 2014. Performed by professionals or consultants, most of
the times, RCM models have applied methods of reliability analysis with qualitative nature, with but a few based on
quantitative analysis. The main RCM activities are organized as elements that are identified, presented and analyzed.
In the results, difficulties are observed in developing RCM as a function of the quantity and quality of maintenance
information. Qualitative tools such as FMEA (Failure Modes and Effects Analysis), FTA (Fault Tree Analysis),
FFA (Functional Failure Analysis), decision diagrams, among others, are widely used in RCM implementations.
These studies have the human factor in common, hindering the generation of a solid and reliable base of asset
maintenance knowledge to be used in decision making. Quantitative approaches are mentioned, like the analysis of
relationship between system reliability and maintenance efforts, logical and structured reliability analysis, failure
rate modeling, economic analysis of maintenance tasks and the use of standardized components for reliability
analysis.
The decisions of the maintenance policies to be adopted should be made, as far as possible, with an appropriate
model. According to [23] the objective function should be cost per time unit, over a time period. If there are no exact
analytical models, simulation models can be used. It is important to emphasize that technology is an important factor
to support the existence and quality of the information available. Modern systems based on Industry 4.0
requirements can operate with the support of a greater informational density and tools to support decision making.
The data collected with the aid of these systems is saved in clouds. Products that integrate with cloud computing in
1138 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144

the field can provide data that enable predictive maintenance and provide information about optimization
possibilities in production.

2.3. Data, information and knowledge for decision making dimensions

Tacit and explicit data have real importance in the industrial scenario, being a bridge for correct decision making
in maintenance. Some researches refer to this collaborative approach with a view that when coming from reliable
sources, data and information are important to the whole process. Data fusion can be defined as the process of
combining data and knowledge from different sources in order to maximize the use of the information available to
increase reliability while minimizing the amount of data retained.
To improve the quality of quantitative information, data mining techniques have been used to treat the raw
information and improve the consistency of the information available. Process Mining can be defined as the
discovery of knowledge by analyzing task execution logs [24] and is used to check the difference between what was
going to happen and what actually did. Process Mining has been applied in many areas, such as business processes,
decision-making, software engineering, etc. [25]. [26] proposed the use of Process Mining for organizing and
maintaining knowledge in a generic process consisting in the creation of a knowledge collection and maintenance
system. The quality of knowledge being generated is important for collaboration between levels of the organization.
In this context, mining processes can be applied to review maintenance records, discovering the process and
assisting in creating a model of optimal maintenance policies.
The use of mining techniques in data whose source is maintenance planning is proposed by [27]. A methodology
using a mining approach to extract accurate fault rate data from WOs (Work Orders) and DD (Downtime Data) is
presented by [28]. The purpose of these initiatives is to explore the available information, evaluating possible
consistent patterns, such as association rules or temporal sequences, and detecting systematic relationships between
variables. With this, the knowledge base to be generated becomes more robust and reliable, increasing the credibility
of the decision making.
Maintenance managers have access to large volumes of data and have a complicated task that is to turn that data
into information that supports maintenance actions [29]. Important criteria include the need for multidisciplinary
team building, decision support tools, reliability data, and assistive technologies [13, 30]. In this context,
publications involving CBM - Condition-Based Maintenance [31], risk assessment [32], TPM - Total Productive
Maintenance [33] are examples of applied research to improve decision-making in maintenance.
MCDM methods are commonly used in maintenance decision making. These techniques allow information on
maintenance goals and objectives to be converted into evaluation criteria and brought into a framework that
incorporates stakeholder views. A literature review on researches with applications of MCDMs in the selection of
maintenance strategies is found in [34]. For the evaluation of the criteria adopted in MCDMs commonly are found
in the literature studies using the Multi-Criteria Decision Analysis (MCDA) method that consists in the analysis of
multiple criteria in decision-making approaches. It is an important intervention to address multiple and conflicting
goals where decision makers value them differently. Applications of MCDA methods allow a structured and
consistent evaluation, integrating quantitative and qualitative criteria [35].

3. RCM deployment model proposed

Around the world, RCM is considered an imperative methodology in the field of industrial maintenance that can
be functional in improving equipment uptime and reliability and reducing operational and maintenance costs [22].
To take advantage of it, the RCM classical approach seeks to answer seven questions presented sequentially in
system or process analyses: (i) What functions should be preserved? (ii) What are the functional failures? (iii) What
are the Failure Modes? (iv) What are the effects of failure? (v) What are the consequences of failure? (vi) What are
the applicable and effective tasks? (vii) What are the other alternatives? To answer these questions systematically,
the RCM deployment process in maintaining devices or systems can be summarized in steps, shown in Figure 1
[36].
 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144 1139

Fig. 1. RCM deployment steps and outputs [36].

Each phase of the RCM deployment model is composed of inputs, controls, mechanisms and outputs, comprising
an information flow. The information generated after the RCM implementation (Feedback) is made available to the
model for continuous improvement.

3.1. Mapping the knowledge and scope of analysis

A mix of information with qualitative and quantitative characteristics is available in all phases of RCM analysis.
At each stage there is a predominance the generation of tacit or explicit knowledge, or mixed approaches. Thus,
based on the RCM implementation model presented in Figure 1, a comparison mapping of the knowledge type used
in the deployment phases as well as scope of analysis (Table I) was developed.

Table I. RCM Deployment: steps, knowledge type and scope of analysis.

Knowledge
RCM Step Scope of analysis
Tacit Explicit
RCM Plan X X Analysis of individual and collective tacit knowledge and skill sets, to establish the RCM team and
program management, analysis and support. Tacit knowledge is applied in choosing the level of
analysis, based on the knowledge of the RCM team experts. Explicit knowledge is used in selecting the
system to be analyzed, investigating the MSIs (Maintenance Significant Items), like KPIs (Key
Performance Indicators). Quantitative data is important in this phase in identifying the critical systems
based on process parameters and maintenance and process indicators.
Hardware X Logical division of an item into subsystems and progressively smaller elements (components) with
partitioning decreasing complexity levels. It is a step based on tacit knowledge from team experts.
FMECA X X Tacit knowledge is used in identifying and analyzing functions, failures, failure modes and criticality,
being a predominant approach in this step. Explicit knowledge is employed in FMEA or FMECA
extensions in analyzing process parameters, criticality and risk assessment.
1140 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144

Significant X Consists in analyzing failure effects on safety, environment, operations, or economics requirements.
function Tacit knowledge is applied in qualitative decision tools, like logic diagrams.
selection
RCM task X X Tacit knowledge is an important information source to determine which of several options is best
evaluation suited to prevent the occurrence of a given failure mode or, if unable to prevent it, reduce the
consequence of the failure to a level that is acceptable for the program. Maintenance task efficiency
evaluation commonly uses explicit knowledge from maintenance indicators applying them in decision
making approaches, in order to choose the best maintenance policies and techniques.
RCM task X Tacit approaches are applied in this phase using the answers from specialists in decisional diagrams,
selection proposed in RCM standards and publications, used in defining adequate maintenance tasks and their
frequency, based on qualitative analysis by specialists.
Implementation X The RCM Implementation Manager uses tacit knowledge to establish a method of review and approval
that ensures RCM methodology and to verify continuously if the RCM program is properly and
effectively applied, maintaining an audit trail of RCM recommendations and implemented actions.

3.2. RCM deployment model

A concise RCM deployment model was developed based on this mapping, with emphasis on knowledge
conciliation. The main objective is to facilitate the deployment process, focusing on those phases where qualitative
and quantitative techniques can be applied, improving the maintenance knowledge inventory available for
maintenance decision making (strategies). The model can be summarized in three phases: (i) Resource and system
selection; (ii) Failure, effect and criticality analyses and (iii) task assessment and feedback. Figure 2 presents the
model proposed. The three phases of the model are explained next.

Fig. 2. RCM deployment model proposed.

3.2.1. Phase 1: resource and system selection

The main objective of this phase is to appropriately select the system to be analyzed, based on MSIs
(Maintenance Significant Items) chosen by experts (based on process characteristics). Process data, flowcharts,
procedures and technical content could be used in supporting this phase. MCDM techniques are proposed using
criteria from maintenance indicators to select the most important system to be analyzed.

3.2.2. Phase 2: failure, effects and criticality analysis

This phase is basically the application of FMECA (Failure Modes, Effects and Criticality Analysis) that consists
in an extension of FMEA, with qualitative and quantitative approaches. Tacit knowledge from maintenance experts
 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144 1141

are used in listing system functions, functional failures, failure modes and causes and effects analysis. Explicit
knowledge like failure rates are used in quantitative analysis. The criticality of the system consists in the RPN (Risk
Priority Number), whose results are listed in a ranking based on failure frequency, severity and detectability. These
parameters can be extracted in a customized way based on the level of maintenance maturity (asset data complexity
and accuracy). The classical approaches to RPN calculation use tacit knowledge, applying the answers of experts
based on a scale of criticality. The collection of information should be recorded on a spreadsheet that will ensure the
documentation of the failure modes associated with each functional failure, its causes and effects, assisting in the
analysis of maintenance actions. The application of MCDM techniques is used to help in structuring and applying
decisions in connection with RPN parameter definitions (occurrence, severity, detection) and decisional diagrams
(including maintenance policy definition). To support data mining and decisions processes, event and task
identification (ID) systems must be established for each failure mode, in order to improve the maintenance plan.

3.2.3. Phase 3: task evaluation and feedback

This phase addresses many of the issues considered prior to implementing an RCM program, within a context of
knowledge and decision. Through applying adequate techniques, the tasks selected are improved and updated on the
maintenance plan. As the level of maintenance maturity increases, previously selected tasks can be changed to the
most adequate type (for example, a task in a failure mode defined as Run-to-failure can be changed to TBM (Time-
Based Maintenance) or CBM (Condition-Based Maintenance)). The parameters adopted, such as calendar,
frequency, period and time of interventions, can be reviewed.
The strategies adopted in maintenance plans must be continuously analyzed. The data extracted is now in a single
knowledge database and can be mined applying Process Mining techniques, revealing inconsistent patterns and
verifying systematic relationships in the maintenance task information set. Hidden patterns and relationships can be
discovered and used in support to MCDA techniques. Results tracking have to be formalized with a readable format
and feedback on the RCM deployment.

4. General RCM framework proposed

A general conceptual framework is proposed in order to execute the deployment process (Figure 3). It is
composed of three parts: (i) a knowledge (tacit and explicit) conciliation system; (ii) a KDD (knowledge data
discovered) as a knowledge source to be applied in RCM decisions approaches with MCDM support and (iii)
maintenance tasks analyses with MCDA and Process Mining techniques, to track the evolution of knowledge
abstraction, to aid in the selection of the best maintenance policies and provide feedback on the results.

Fig. 3. Conceptual RCM deployment proposed.

1142 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144

4.1. Conciliation system – discovering knowledge

The knowledge conciliation system is the initial collection and treatment of tacit and explicit data. Tacit data
consists in qualitative information based on experience from experts. The sources of this type of data can include
relations, observations, reports, norms, values, operating standards and maintenance procedures. The data can be
collected using forms, questionnaires, interviews, etc. Explicit data can be extracted from maintenance indicators,
archives, information systems like CMMS and EAM, data logs, process data (sensors, controllers and actuators),
smart devices and data sources related to Industry 4.0. There are several tools and methods for explicit data
collection, but they are not always available in all processes. In addition, the level of maturity of the company must
be analyzed from organizational, structural and technological aspects. Process Mining is an important approach to
analyze this data source and validate it, transforming the information into a reliable source of knowledge. After
collecting the tacit and explicit data available, a raw information data base is generated. A customized information
model is created generating new knowledge that is stored in a KDD.

4.2. Decisions - KDD to RCM deployment

The new knowledge base consists in all the primary data originated in the conciliation system. It consists in tacit
(experience-based) and explicit (processes and systems) information. According with the maintenance structure
maturity the knowledge generated in this phase can include: (i) creation of a guide to assist the specialists involved
(planning); (ii) accurate information about the hardware of the systems (hierarchical structure and analysis of the
components) and their respective functions; (iii) information about failure modes, effects and criticality; (iv) in
service data and MDE (Manual Data Entry); (v) system operation and maintenance indicators.
The proposed RCM deployment model uses the new knowledge as inputs for techniques employed in the first
two phases of the analysis (Figure 3) in a customized way, according to the characteristics of the system selected. In
these phases, MCDM is considered a complex maintenance decision-making tool involving both quantitative and
qualitative factors. It can be applied in selecting the system (under the MSI aspects specified as decision criteria), to
support in the FMECA implementation flow (including parameters adopted in RPN calculation), to evaluate the
consequences of failures, to drive answers for decision diagrams, among others. The final objective is to choose the
optimal probable options in maintenance actions and frequencies to be adopted for each failure mode.

4.3. Analysis - Maintenance strategies

Decisions define important actions in the maintenance scenario. The strategies adopted in the last phase of the
RCM deployment proposed (stemming from the FMECA analysis) can undergo changes in process or components
over time. This includes changes in process characteristics, parameters, type and level of technology employed,
available resources, optimizations, etc. Analytics for MCDA approaches can support the update of maintenance
plans using, among other items, information from KDD about Work Orders (scheduled, completed and past due),
the estimated time versus actual time spent in maintenance activities and machine downtime versus uptime. A
connection with Process Mining techniques is observed in order to improve data consistency and update the
information when necessary. The use of storage fields is considered important to back up the information and
provide data security.

Conclusion

The adaptation and customization of tools or methodologies applied in the maintenance function has been shown
to be an important initiative. Frameworks with contemporary approaches of more flexible and simplified models
have gained space in related research. The framework presented brings important highlights in RCM development,
with a succinct and dynamic model, enabling the deployment of tools widely disseminated by researchers around the
world.
The research emphasizes an important relation between tacit and explicit data, proposing an integration and
generation of a single knowledge base for later use in the adaptation stages of RCM development. As compared to
 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144 1143

the traditional models, it is simplified through the reduction of the implementation stages, enhancing the knowledge
and the possible techniques applicable in data treatment (Process Mining) and decision making/analysis (MCDM/A)
for the maintenance function.
The level of maintenance maturity from structural, organizational and technological standpoints must be observed
and mapped in the planning phase of RCM deployment. The decision making and analysis models and the Process
Mining techniques depend on the application of this requirement.
With the use of the proposed framework, the appropriate maintenance tasks can be identified and inserted into a
viable maintenance plan. The tacit and explicit database allow the development of the model to be continually
improved. By applying the right techniques, we have the possibility of choosing the appropriate maintenance
activities and optimize parameters based on the type of maintenance, the frequency of tasks and the time intervals for
a more efficient and effective delivery. Well-defined maintenance strategies significantly improve results in meeting
business goals by increasing facility reliability, eliminating unnecessary activities and reducing costs.

References

[1] B. Kazaz, T.W. Sloan, The impact of process deterioration on production and maintenance policies. European Journal of Operational
Research. 227 (2013) 88–100.
[2] N. Subramanian, R. Ramanathan, A review of applications of analytic hierarchy process in operations management. International Journal of
Production Economics. 138 (2012) 215–241.
[3] R.J.K. Netto, E.A.P. Santos, E.F.R. Loures, R. Pierezan, Using Overall Equipment Effectiveness (OEE) to predict shutdown maintenance.
Joint International Conference ICIEOM-CIO-IIIEC, 2015.
[4] C. PACIAROTTI, G. MAZZUTO, D. D’ETTORRE, A revised FMEA application to the quality control management. International Journal of
Quality & Reliability Management. 31 (2014) 788–810.
[5] A.W. Labib, A decision analysis model for maintenance policy selection using a CMMS. Journal of Quality in Maintenance Engineering. 10
(2004) 191-202.
[6] R.K. Sharma, D. Kumar, P. Kumar, Modeling and analysing system failure behaviour using RCA, FMEA and NHPPP models. International
Journal of Quality & Reliability Management. 24 (2007) 525-546.
[7] A.C. Marquez, J.N.D. Gupta, Contemporary maintenance management: process, framework and supporting pillars. The International Journal
of Management Science. 34 (2005) 313-326.
[8] G. Waeyenbergh, L. Pintelon, Maintenance concept development: a case study. International journal of production economics. 89 (2004) 395-
405.
[9] V.S. Deshpande, J.P. Modak, Application of RCM to a medium scale industry. Reliability Engineering e System Safety. 77 (2002) 31-43.
[10] A. Voisin, E. Levrat, P. Cocheteux, B. Iung, Generic prognosis model for proactive maintenance decision support: application to pre-
industrial e-maintenance test bed. Journal of Intelligent Manufacturing. 21 (2010) 177-193.
[11] M. Galar, A. Fernandez, E. Barrenechea, H. Bustince, F. Herrera, A review on ensembles for the class imbalance problem: bagging-,
[12] C. Ebeling, Reliability and maintainability engineering, McGraw Hill, USA, 1997.
boosting-, and hybrid-based approaches. Man, and Cybernetics, Part C: Applications and Reviews. 99 (2011) 1-22.
[13] J. Moubray, Reliability-centered maintenance, second ed., Industrial Press, New York, 1997.
[14] K. Neubeck, Practical Reliability Analysis, Pearson Prentice Hall, 2004.
[15] P.D.T. O’Connor, Practical Reliability Engineering, fourth ed., Wiley, 2002.
[16] D.J. Smith, Reliability, maintainability and risk – practical methods for engineers, seventh ed., Butterworth-Heinemann, 2005.
[17] J.T. Selvik, T. Aven, A framework for reliability and risk centered maintenance. Reliability Engineering and System Safety. 96 (2011) 324–
331.
[18] IEC. IEC 60300-3-11/ NEK IEC 60300-3-11: Dependability management, Part 3–11: Application guide Reliability centred maintenance.
Norsk elektroteknisk komite (NEK), 1999.
[19] SAE. SAE-JA1011: Evaluation criteria for reliability-centered maintenance (RCM) processes, Society of Automotive Engineers, 1999.
[20] SAE. SAE-JA1012: A guide to the reliability-centered maintenance (RCM) standard, Society of Automotive Engineers, 2002.
[21] V. Narayan, Business performance and maintenanceHow are safety, quality, reliability, productivity and maintenance related. Journal of
Quality in Maintenance Engineering. 18 (2012) 183-195.
[22] G. Gupta, R. P. Mishra, A SWOT analysis of reliability centered maintenance framework. Journal of Quality in Maintenance Engineering.
22 (2016) 130–145.
[23] P. Sinha, Towards higher maintenance effectiveness Integrating maintenance management with reliability engineering. International Journal
of Quality & Reliability Management. 32 (2015) 754–762.
[24] W.M.P Van der Aalst, A.J.M.M. Weijters, L. Maruster, Workflow mining: Discovering process models from event logs. IEEE
Transactions on Knowledge and Data Engineering. 16 (2004) 1128–1142.
[25] L. Cabac, N. Denz, Net Components for the Integration of Process Mining into Agent-Oriented Software Engineering. In ToPNoC. 5100
(2008) 86–103.
[26] M. Li, L. Liu, L. Yin, Y. Zhu, A process mining based approach to knowledge maintenance. Information Systems Frontiers. 13 (2011) 371–
380.
[27] S. Bjorklund, Utilisation of data mining principles in maintenance planning. Proceedings of the 21st Int. DAAAM Symposium. 21 (2010)
753-754.
[28] K. Arif-Uz-Zaman, M.E. Cholette, L. Ma, A. Karim, Extracting failure time data from industrial maintenance records using text mining.
1144 Flávio Piechnicki et al. / Procedia Manufacturing 11 (2017) 1135 – 1144

Advanced Engineering Informatics. xxx (2016) xxx–xxx.

[29] E.G. Berggren, Efficient Track Maintenance: Methodology for Combined Analysis of Condition Data. Proceedings of the Institution of
Mechanical Engineers. 224 (2010) 353-360.
[30] F. I. Khan, M. M. Haddara, Risk-based maintenance (RBM): a quantitative approach formaintenance/inspection scheduling and planning.
Journal of Loss Prevention in the Process Industries. 16 (2003) 561–573.
[31] A. Bousdekis, B. Magoutas, D. Apostolou, G. Mentzas, A proactive decision making framework for condition-based maintenance. Industrial
Management & Data Systems. 115 (2015) 1225-1250.
[32] P. Chemweno, L. Pintelon, A.V. Horenbeek, P. Muchiri, Development of a risk assessment selection methodology for asset maintenance
decision making: An analytic network process (ANP) approach. Int. J. Production Economics. 170 (2015) 663–676.
[33] A.S. Piechnicki, A.V.H. Sola, F. Trojan, Decision-making towards achieving world-class total productive maintenance. International Journal
of Operations & Production Management. 35 (2015) 1594–1621.
[34] M. Shafiee, Maintenance strategy selection problem: an MCDM overview. Journal of Quality in Maintenance Engineering. 21 (2015) 378-
402.
[35] A. Ishizaka, V.E. Pereira, Portraying an employee performance management system based on multi-criteria decision analysis and visual
techniques. International Journal of Manpower. 37 (2016) 628-659.
[36] NAVAIR 00-25-403, Management Manual: guidelines for the naval aviation Reliability-Centered Maintenance Process. Naval Air Systems
Command. USA, 2005.