© JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

IoT-enabled Predictive Maintenance for Mechanical

Systems: Innovations in Real-time Monitoring and
Operational Excellence
AADIT SHARMA1, BOLAJI IYANU ADEKUNLE2, JEFFREY CHIDERA OGEAWUCHI3,
ABRAHAM AYODEJI ABAYOMI4, OMONIYI ONIFADE5
1
Astro Global, India
2
Data Analyst, Federal Ministry of Mines and Steel Development, Nigeria
3
Brickwall Global Investment Company, Lagos State, Nigeria
4
Hagital Consulting, Lagos State, Nigeria
5
Kenaitze indian tribe, Kenai, AK, USA

Abstract- The advent of the Internet of Things (IoT) I. INTRODUCTION

has significantly transformed maintenance strategies
for mechanical systems, transitioning from reactive 1.1. Evolution of Maintenance Strategies in
and preventive approaches to intelligent, predictive Mechanical Systems
maintenance frameworks. This paper explores the
Maintenance practices in mechanical systems have
integration of IoT technologies—specifically sensor
evolved significantly over the past few decades,
networks, edge computing, and cloud
transitioning from reactive models to more proactive
infrastructure—into mechanical system monitoring
and intelligent strategies [1]. Reactive maintenance,
to enable real-time diagnostics and failure
commonly known as the "run-to-failure" approach,
prediction. It outlines the evolution of maintenance
dominated early industrial operations [2]. This model
strategies and highlights how embedded sensing and
involved addressing equipment breakdowns only after
continuous data collection are foundational to
they occurred, which often led to prolonged
predictive analytics. Through detailed examination
downtimes, increased repair costs, and safety risks.
of system architecture, communication protocols,
While suitable for low-cost or non-critical equipment,
and machine learning methodologies, the paper
this strategy was inadequate for high-demand
illustrates how predictive models and digital twins
environments where reliability is essential [3].
enhance fault detection, equipment longevity, and
resource allocation. Case studies demonstrate Preventive maintenance emerged to address the
quantifiable operational benefits, including reduced limitations of reactive practices. This approach
unplanned downtime and cost savings. The strategic involves scheduled inspections and replacements
and organizational implications are analyzed, based on average life cycles or usage intervals,
emphasizing workforce transformation, regardless of the actual condition of components [4].
implementation barriers, and cybersecurity While preventive maintenance helped reduce sudden
considerations. Ultimately, this study presents a failures, it frequently resulted in unnecessary
comprehensive framework for implementing IoT- maintenance tasks and associated costs. Additionally,
enabled predictive maintenance and suggests future it lacked the flexibility to respond to unexpected
research directions centered on AI convergence, operational variations and complex wear patterns in
system interoperability, and sustainability in modern industrial systems [5].
industrial operations.
The advent of predictive maintenance marked a
Indexed Terms- Predictive Maintenance, Internet of paradigm shift in asset management. By leveraging
Things (IoT), Mechanical Systems, Real-time condition monitoring tools and data analytics,
Monitoring, Digital Twins, Operational Efficiency predictive strategies aim to forecast equipment failures
before they occur [6]. This method not only minimizes

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 270

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

unplanned downtime but also optimizes resource cost-efficiency that were previously unattainable with
utilization by targeting maintenance actions precisely legacy systems [13].
when they are needed. Despite its advantages,
traditional predictive maintenance techniques often 1.3. Research Objectives and Scope
rely on isolated sensors and periodic data analysis,
The central objective of this paper is to examine the
limiting their responsiveness and scalability in
integration of IoT in predictive maintenance for
dynamic environments. This gap set the stage for a
mechanical systems, with a focus on real-time
more integrated and intelligent solution enabled by
monitoring technologies and their role in achieving
emerging technologies [7].
operational excellence. Specifically, the study aims to
1.2. Emergence of IoT in Industrial Applications evaluate how sensor networks, data communication
frameworks, and intelligent analytics contribute to
The integration of smart technologies into industrial minimizing equipment failures, reducing maintenance
systems, particularly through the Internet of Things, costs, and improving asset longevity. It further seeks
has redefined maintenance and monitoring capabilities to assess the implications of these innovations on
[8]. In the context of mechanical systems, this maintenance strategies, workforce transformation, and
development represents a crucial component of the enterprise-wide decision-making.
broader Industry 4.0 movement, which emphasizes
automation, data exchange, and cyber-physical To achieve these goals, the paper explores both the
integration. By embedding sensors, connectivity technological and operational dimensions of IoT-
modules, and computational intelligence into enabled predictive maintenance. From a technical
equipment, IoT enables continuous and context-aware perspective, it investigates sensor types, data
monitoring of critical mechanical components [9]. architectures, communication protocols, and
analytical models that enable proactive maintenance
One of the most transformative benefits of this strategies. From an operational standpoint, it considers
technology is its ability to collect high-frequency, how industries implement these systems, the
high-resolution data in real time. IoT-enabled sensors challenges they encounter in terms of integration and
can track temperature, vibration, pressure, acoustic interoperability, and the measurable outcomes realized
signals, and other operational parameters with in performance and cost savings.
exceptional granularity [10]. This continuous data
stream allows systems to detect deviations from The scope of this paper encompasses various
optimal conditions almost instantaneously, paving the mechanical systems across industrial domains,
way for accurate fault diagnosis and timely including manufacturing lines, HVAC systems,
interventions. Additionally, real-time alerts can be industrial pumps, and turbines. While the focus is
generated and transmitted through edge devices or predominantly on heavy-duty applications, the
cloud platforms, allowing remote maintenance teams analysis also considers scalable use cases applicable to
to take informed decisions without being physically small and medium enterprises. Geographically, the
present on-site [11]. study draws on global case examples, though
particular emphasis is placed on industrial regions
Moreover, this real-time intelligence is not confined to adopting Industry 4.0 frameworks. This
individual machines but can be aggregated across comprehensive approach ensures that the insights
entire production lines or facilities. IoT facilitates derived are both practical and broadly relevant to
interconnectedness across different mechanical assets, stakeholders across sectors seeking to modernize their
creating a digital ecosystem where each unit maintenance strategies.
contributes to a collective awareness of system health
[12]. This integration supports advanced analytics,
trend analysis, and machine learning applications that
further enhance predictive capabilities. As such, IoT
not only augments predictive maintenance but also
introduces operational transparency, scalability, and

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 271

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

II. CORE TECHNOLOGIES AND SYSTEM 2.2. Data Transmission and Cloud Integration
ARCHITECTURE
Once data is collected and processed at the edge, it
2.1. IoT Sensors and Edge Devices must be transmitted efficiently and securely to central
platforms for further analysis and long-term storage.
The foundation of IoT-enabled predictive maintenance This is enabled by a variety of communication
lies in the deployment of specialized sensors that protocols, each suited to different industrial
continuously monitor the operational state of requirements. Message Queuing Telemetry Transport
mechanical systems [2]. These sensors measure key (MQTT) is widely used due to its lightweight design
parameters such as vibration, temperature, acoustic and reliability over low-bandwidth connections. It
signals, and pressure. Vibration sensors, for instance, allows for real-time message exchange between
are particularly effective in identifying early signs of sensors, edge devices, and cloud servers. For large-
imbalance, misalignment, or bearing failures in scale outdoor deployments or energy-constrained
rotating machinery [14]. Temperature sensors detect environments, Low Power Wide Area Network
overheating that could signal excessive friction or technologies such as LoRaWAN offer long-range,
electrical faults, while acoustic sensors capture high- low-power communication capabilities [20].
frequency sound patterns that often precede
mechanical degradation. Pressure sensors, commonly In high-speed industrial environments, cellular
used in hydraulic and pneumatic systems, monitor technologies like 5G provide low-latency, high-
deviations that may indicate leaks or clogging [15]. bandwidth transmission essential for mission-critical
applications. 5G networks also support a higher
Edge devices act as intermediaries between the sensors density of connected devices, making them ideal for
and higher-level data processing systems. These facilities with extensive sensor arrays. These protocols
devices, which include industrial gateways and form the backbone of the data transmission layer,
embedded microcontrollers, collect sensor data and enabling seamless and scalable communication
conduct preliminary filtering, aggregation, and between field equipment and remote analytics
anomaly detection locally [16]. By processing data infrastructure [21].
near the source, edge computing reduces latency and
minimizes the need for constant cloud connectivity. Cloud integration is the next crucial layer in the
This is especially important in environments where system. Cloud platforms offer scalable storage,
immediate response times are critical to operational compute power, and analytical capabilities that far
safety and system uptime [17]. exceed what is possible on-site. They support the
ingestion of high-volume, high-velocity data streams
Furthermore, edge devices contribute significantly to and provide interfaces for data scientists and
system efficiency and resilience. In the event of maintenance engineers to access historical records and
network disruptions, they can continue functioning perform in-depth analytics [22]. Moreover, cloud
autonomously and store data temporarily for later environments facilitate remote diagnostics,
transmission [18]. They also enable data compression collaborative decision-making, and cross-site
and encryption, supporting secure and bandwidth- monitoring, which are vital for global enterprises.
efficient communication. With built-in intelligence, Security measures such as end-to-end encryption,
some edge devices even run lightweight machine multi-factor authentication, and role-based access
learning models for on-site analytics. This localized controls ensure the confidentiality and integrity of
processing helps ensure that only meaningful and pre- sensitive operational data [23].
processed data is transmitted to centralized platforms,
optimizing the performance of the entire predictive 2.3. System Architecture for Predictive Maintenance
maintenance ecosystem [19].
The architecture of an IoT-enabled predictive
maintenance system typically consists of five
interconnected layers: data collection, transmission,
storage, analytics, and visualization. At the base is the

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 272

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

data collection layer, where various sensors and edge component failure. These models are trained on time-
devices capture real-time machine parameters. This series data from sensors, enabling them to forecast
raw data is often pre-processed locally to ensure failure events based on leading indicators like
relevance and reduce transmission loads. The increased vibration or temperature anomalies [25].
transmission layer involves communication protocols
that securely relay this data to centralized or In cases where labeled failure data is scarce or
distributed platforms. unavailable, unsupervised learning techniques are
employed to identify anomalous behavior without
Once the data reaches the storage layer, it is organized prior categorization. Clustering algorithms like k-
within scalable databases, often hosted on cloud means or DBSCAN can group sensor patterns and flag
services or hybrid cloud-edge architectures. This layer outliers that may suggest emerging faults.
ensures that both real-time and historical data are Autoencoders and principal component analysis are
readily available for processing. In the analytics layer, also useful in reducing data dimensionality and
advanced algorithms detect anomalies, predict detecting deviations in complex systems. These
failures, and generate maintenance recommendations. models are particularly effective in early-stage fault
These analytics may incorporate time-series modeling, detection and preventive alert generation [26].
pattern recognition, or machine learning techniques
tailored to the specific behavior of mechanical Streaming data analytics further enhances the
systems. responsiveness of predictive models. By continuously
ingesting and analyzing live sensor data, machine
The final layer, visualization, provides intuitive learning systems can update their predictions in near
dashboards and alerting mechanisms for end-users. real time. This capability supports dynamic
Operators can interact with real-time metrics, trend maintenance scheduling and immediate fault isolation
graphs, and maintenance schedules via web-based [27]. Adaptive learning models, which retrain as new
interfaces or mobile applications. Importantly, this data becomes available, improve accuracy over time
architectural model must also account for and adjust to evolving equipment conditions. The
cybersecurity and scalability. Security features such as integration of supervised and unsupervised learning in
data encryption, device authentication, and intrusion predictive maintenance allows for a comprehensive
detection are integrated across layers to mitigate cyber and robust approach to equipment health monitoring,
threats. Scalability is achieved through modular minimizing unplanned outages and maximizing asset
system design and cloud-native services, allowing availability [28].
organizations to expand their monitoring capabilities
as operational demands grow. This layered 3.2. Digital Twins and Simulation Models
architecture ensures a robust, responsive, and future-
Digital twins represent a significant innovation in
proof approach to predictive maintenance.
predictive maintenance, offering a virtual replica of
III. PREDICTIVE ANALYTICS AND physical assets that mirrors their real-time operational
MAINTENANCE OPTIMIZATION behavior. These models integrate live sensor data with
engineering models and historical records to simulate
3.1. Machine Learning for Failure Prediction system dynamics under various conditions. By
continuously synchronizing with physical systems,
Machine learning plays a pivotal role in predictive digital twins provide a contextualized and evolving
maintenance by enabling systems to learn from view of component performance, degradation trends,
historical patterns and detect subtle deviations in and potential failure points [29].
mechanical behavior. Supervised learning algorithms
are commonly applied when labeled data—indicating The use of digital twins allows for advanced
normal and failed conditions—is available [24]. diagnostics and “what-if” simulations, enabling
Algorithms such as decision trees, support vector operators to explore the impact of specific operating
machines, and logistic regression are used to classify scenarios or intervention strategies before
system states and estimate the probability of implementing them on physical assets. For example,

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 273

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

simulating increased load or altered temperature pressure, and fan motor currents. Fault detection
conditions can reveal stress points and predict the algorithms enabled timely interventions before system
likelihood of future faults. These insights empower degradation led to performance issues or energy
more accurate decision-making, reduce the risk of inefficiencies [35]. This not only preserved occupant
incorrect interventions, and extend asset lifecycles comfort but also resulted in measurable energy savings
through proactive maintenance planning [30]. and reduced service costs. These cases underscore the
tangible benefits of predictive maintenance in diverse
In addition to enhancing fault prediction accuracy, mechanical environments—demonstrating how
digital twins facilitate better communication between timely, data-driven decisions can significantly
maintenance teams, engineers, and decision-makers. improve asset reliability, lifecycle value, and
Graphical interfaces linked to the twin provide operational excellence [36].
intuitive visualizations of system health and
performance indicators [31]. Integration with IV. OPERATIONAL AND STRATEGIC
enterprise systems like ERP and CMMS ensures that IMPACTS
predictive insights translate into actionable work
orders [32]. The adoption of digital twins in 4.1. Operational Efficiency and Cost Reduction
mechanical system maintenance is accelerating,
The integration of Internet of Things technologies in
particularly in industries with complex and high-value
predictive maintenance significantly enhances
assets such as aerospace, energy, and manufacturing.
operational efficiency by minimizing unplanned
Their ability to combine real-time monitoring with
equipment downtime. Real-time monitoring enables
simulation-based foresight makes them indispensable
maintenance teams to detect deviations in component
tools for modern maintenance optimization [33].
behavior early, allowing interventions before a failure
3.3. Case Examples of Predictive Optimization occurs [37]. This proactive approach extends asset
lifespans and reduces the frequency of costly
Numerous real-world implementations illustrate the emergency repairs. For industries reliant on
effectiveness of predictive maintenance in enhancing continuous operations, such as manufacturing,
operational efficiency and cost control. In the aviation, and power generation, improved equipment
manufacturing sector, a leading automotive company uptime directly translates to increased productivity
deployed IoT-based predictive analytics on its robotic and profitability [38].
assembly lines. By using vibration and temperature
sensors on servo motors, combined with machine Beyond uptime improvements, predictive
learning models, the firm detected early signs of wear maintenance optimizes resource allocation by shifting
and avoided critical failures. This approach reduced maintenance schedules from fixed intervals to
unplanned downtime by 30% and extended the useful condition-based interventions. This eliminates
life of high-cost components by 20%. unnecessary part replacements and labor costs
associated with routine but often redundant
In the energy industry, a wind farm operator utilized inspections [39]. Maintenance crews can focus their
digital twins and predictive algorithms to monitor efforts on assets that genuinely require attention,
turbine blade stress and gearbox temperature in real enhancing workforce productivity. Predictive systems
time. By simulating environmental and operational also help reduce spare parts inventory requirements, as
scenarios, the system predicted maintenance needs parts are ordered based on predicted needs rather than
weeks in advance. As a result, maintenance activities forecasted consumption, improving inventory
were strategically scheduled during low-wind periods, turnover rates and reducing warehousing costs [37].
minimizing revenue losses and improving safety
conditions for field technicians [34]. Moreover, the data-driven nature of predictive
maintenance facilitates continuous process
Similarly, in the HVAC industry, building improvement. Historical and real-time data collected
management systems equipped with predictive through connected sensors can be analyzed to identify
analytics monitored compressor cycles, refrigerant inefficiencies in system operations. These insights

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 274

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

enable process adjustments that further reduce energy employees can thrive in a data-enriched industrial
consumption, improve throughput, and support ecosystem [32].
sustainability goals. Collectively, these outcomes
position predictive maintenance not only as a tool for 4.3. Strategic Deployment Challenges
asset health but also as a strategic lever for operational
Despite the clear benefits, deploying IoT-enabled
excellence and cost containment across diverse
predictive maintenance faces several strategic
industrial sectors [40].
challenges. The most immediate barrier is the
4.2. Workforce Transformation and Skill substantial upfront capital investment required for
Requirements sensor installation, edge devices, communication
infrastructure, and cloud integration. Many
The shift from manual to data-driven maintenance organizations, particularly small- to medium-sized
approaches is fundamentally transforming the enterprises, struggle with justifying these investments
industrial workforce. Traditional maintenance relied without immediate and measurable return on
heavily on physical inspections, intuitive judgments, investment. This financial hurdle is often compounded
and reactive troubleshooting [41]. With IoT-enabled by the need to retrofit legacy systems, which may not
systems generating vast volumes of data, the role of be readily compatible with modern IoT technologies.
maintenance personnel is evolving to include data
analysis, digital system management, and algorithm Data integration is another significant challenge,
interpretation. This paradigm shift necessitates especially in environments where multiple equipment
upskilling workers in domains such as data literacy, brands and control systems coexist. Consolidating
diagnostics interpretation, and basic machine learning heterogeneous data formats into a unified analytics
comprehension [42]. platform requires robust middleware solutions and
standardized communication protocols. Additionally,
Organizations are increasingly investing in data ownership and access rights concerns arise when
interdisciplinary training programs to build hybrid integrating third-party service providers, particularly
competencies among their technical staff. in shared industrial facilities. These complexities can
Maintenance engineers are being trained in sensor slow deployment and compromise the integrity of
technologies, cloud platforms, and data visualization analytics results if not properly managed from the
tools, while data scientists are introduced to the outset [45].
mechanical behavior of industrial assets. This
convergence of IT and operational technology is also Cybersecurity is a critical and growing concern in IoT-
encouraging greater cross-functional collaboration, as based maintenance ecosystems. As more endpoints are
maintenance teams now work alongside data analysts, connected to networks, the attack surface expands,
cybersecurity experts, and systems engineers to increasing vulnerability to cyber threats that could
maintain and improve predictive maintenance disrupt operations or compromise sensitive data.
platforms [43]. Ensuring end-to-end encryption, secure authentication
protocols, and continuous monitoring becomes
Furthermore, the redefinition of roles is leading to new essential [46]. Moreover, regulatory compliance and
career pathways in industrial environments. Roles adherence to industry-specific cybersecurity standards
such as maintenance data analyst, digital twin must be prioritized. Organizations must balance
operator, and reliability engineer are gaining innovation with risk management, ensuring that
prominence [44]. At the organizational level, predictive maintenance deployments are both effective
companies are restructuring their teams to support a and secure [47].
digital-first maintenance strategy, which includes
hiring new talent with backgrounds in mechatronics, CONCLUSION
embedded systems, and analytics. The transformation
IoT-enabled predictive maintenance represents a
reinforces the need for continuous learning and
transformative shift in the management of mechanical
adaptive workforce development strategies, ensuring
systems. By integrating real-time data acquisition,

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 275

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

edge processing, and advanced analytics, industries for broader system-wide implementation while
are moving beyond traditional reactive and time-based minimizing financial and operational risk.
maintenance strategies. This evolution has delivered
substantial technological advantages, including Equally important is stakeholder training and
improved fault detection accuracy, real-time organizational readiness. Transitioning to predictive
diagnostics, and seamless integration of remote maintenance requires workforce upskilling in areas
monitoring capabilities across critical infrastructure. such as sensor data interpretation, digital system
The deployment of multi-sensor arrays and cloud- integration, and cybersecurity. Companies should
enabled platforms facilitates continuous condition establish cross-functional teams involving
monitoring, ultimately enhancing system visibility and maintenance engineers, data scientists, and IT
responsiveness. professionals to foster collaboration. Employee
engagement and change management strategies are
Operationally, predictive maintenance has proven its necessary to gain buy-in and ensure alignment across
capacity to reduce unplanned downtime, optimize departments.
maintenance scheduling, and extend equipment life
cycles. It supports informed decision-making by From a technical architecture standpoint, adopting a
providing timely insights into component health, hybrid model that leverages both edge and cloud
enabling proactive interventions. Additionally, computing is advisable. Edge devices support real-
organizations benefit economically through cost time analytics and immediate decision-making at the
avoidance associated with emergency repairs, spare site level, while cloud infrastructure facilitates long-
part overstocking, and inefficient resource use. These term storage, pattern recognition, and dashboard
advantages collectively contribute to greater visualization. Ensuring interoperability between
reliability, safety, and operational continuity in high- legacy equipment and modern IoT solutions through
demand environments such as manufacturing, energy, middleware and open standards further enhances
and logistics. implementation success. Additionally, organizations
should embed robust cybersecurity protocols from the
The paper has also highlighted the ecosystem of outset to safeguard data integrity and system
innovations—ranging from digital twins and edge functionality.
computing to data-driven simulation models—that are
accelerating predictive maintenance adoption. These REFERENCES
technologies not only improve performance but also
[1] E. C. Fitch, Proactive maintenance for
offer scalable and secure frameworks suitable for
mechanical systems (no. 5). Elsevier, 2013.
deployment across complex, multi-asset
environments. Overall, IoT-based predictive [2] C. J. Turner, C. Emmanouilidis, T. Tomiyama,
maintenance emerges not just as a technological A. Tiwari, and R. Roy, "Intelligent decision
enhancement but as a strategic enabler of operational support for maintenance: an overview and future
excellence and competitive advantage in the era of trends," International Journal of Computer
Industry 4.0. Integrated Manufacturing, vol. 32, no. 10, pp.
936-959, 2019.
For industries exploring IoT-enabled predictive [3] L. Metso, D. Baglee, and S. Marttonen-Arola,
maintenance, a structured and phased approach is "Maintenance as a combination of intelligent it
essential to ensure successful deployment. A systems and strategies: a literature review,"
recommended starting point is the implementation of Management and production engineering
small-scale pilot projects on selected high-value review, vol. 9, no. 1, pp. 51-64, 2018.
assets. These pilots enable the validation of sensor
[4] E. I. Basri, I. H. Abdul Razak, H. Ab-Samat, and
configurations, edge-device compatibility, and
S. Kamaruddin, "Preventive maintenance (PM)
analytical model accuracy within a controlled
planning: a review," Journal of quality in
environment. Pilot outcomes provide critical feedback
maintenance engineering, vol. 23, no. 2, pp. 114-
143, 2017.

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 276

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

[5] E. Gustavsson, M. Patriksson, A.-B. Strömberg, [15] M. Compare, P. Baraldi, and E. Zio, "Challenges
A. Wojciechowski, and M. Önnheim, to IoT-enabled predictive maintenance for
"Preventive maintenance scheduling of multi- industry 4.0," IEEE Internet of things journal,
component systems with interval costs," vol. 7, no. 5, pp. 4585-4597, 2019.
Computers & Industrial Engineering, vol. 76, pp. [16] J. Cecílio and P. Furtado, "Wireless sensors in
390-400, 2014. heterogeneous networked systems," Wireless
[6] T. Zhu, Y. Ran, X. Zhou, and Y. Wen, "A survey Sensors in Heterogeneous Networked Systems,
of predictive maintenance: Systems, purposes vol. 2, pp. 39-59, 2014.
and approaches," arXiv preprint [17] A. M. Rahmani et al., "Exploiting smart e-Health
arXiv:1912.07383, 2019. gateways at the edge of healthcare Internet-of-
[7] H. Ab-Samat and S. Kamaruddin, "Opportunistic Things: A fog computing approach," Future
maintenance (OM) as a new advancement in Generation Computer Systems, vol. 78, pp. 641-
maintenance approaches: A review," Journal of 658, 2018.
Quality in Maintenance Engineering, vol. 20, no. [18] D. Wu, X. Huang, X. Xie, X. Nie, L. Bao, and Z.
2, pp. 98-121, 2014. Qin, "LEDGE: Leveraging edge computing for
[8] O. Vermesan and P. Friess, Internet of things resilient access management of mobile IoT,"
applications-from research and innovation to IEEE Transactions on Mobile Computing, vol.
market deployment. Taylor & Francis, 2014. 20, no. 3, pp. 1110-1125, 2019.
[9] M. Liu, J. Ma, L. Lin, M. Ge, Q. Wang, and C. [19] S. N. Shirazi, A. Gouglidis, A. Farshad, and D.
Liu, "Intelligent assembly system for mechanical Hutchison, "The extended cloud: Review and
products and key technology based on internet of analysis of mobile edge computing and fog from
things," Journal of Intelligent Manufacturing, a security and resilience perspective," IEEE
vol. 28, pp. 271-299, 2017. Journal on Selected Areas in Communications,
[10] G. Lampropoulos, K. Siakas, and T. vol. 35, no. 11, pp. 2586-2595, 2017.
Anastasiadis, "Internet of things in the context of [20] U. Raza, P. Kulkarni, and M. Sooriyabandara,
industry 4.0: An overview," International "Low power wide area networks: An overview,"
Journal of Entrepreneurial Knowledge, vol. 7, ieee communications surveys & tutorials, vol.
no. 1, 2019. 19, no. 2, pp. 855-873, 2017.
[11] V. Tsiatsis, S. Karnouskos, J. Holler, D. Boyle, [21] N. A. Mohammed, A. M. Mansoor, and R. B.
and C. Mulligan, Internet of Things: technologies Ahmad, "Mission-critical machine-type
and applications for a new age of intelligence. communication: An overview and perspectives
Academic Press, 2018. towards 5G," IEEE Access, vol. 7, pp. 127198-
[12] G. Lampropoulos, K. Siakas, and T. 127216, 2019.
Anastasiadis, "Internet of things (IoT) in [22] S. Ouf and M. Nasr, "Business intelligence in the
industry: Contemporary application domains, cloud," in 2011 IEEE 3rd International
innovative technologies and intelligent Conference on Communication Software and
manufacturing," people, vol. 6, no. 7, pp. 109- Networks, 2011: IEEE, pp. 650-655.
118, 2018. [23] V. Josyula, M. Orr, and G. Page, Cloud
[13] C. Gomez, S. Chessa, A. Fleury, G. Roussos, and computing: Automating the virtualized data
D. Preuveneers, "Internet of Things for enabling center. Cisco Press, 2011.
smart environments: A technology-centric [24] M. Sivakumar, M. Maranco, and N. Krishnaraj,
perspective," Journal of Ambient Intelligence "Data Analytics and Artificial Intelligence for
and Smart Environments, vol. 11, no. 1, pp. 23- Predictive Maintenance in Manufacturing," in
43, 2019. Data Analytics and Artificial Intelligence for
[14] L. Franceschini and A. Midali, "Industrial IoT: a Predictive Maintenance in Smart
cost-benefit analysis of predictive maintenance Manufacturing: CRC Press, pp. 29-55.
service," 2019.

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 277

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

[25] A. Diez-Olivan, J. Del Ser, D. Galar, and B. Building Simulation, 2018, vol. 11: Springer, pp.
Sierra, "Data fusion and machine learning for 953-975.
industrial prognosis: Trends and perspectives [36] S. O. Erikstad, "Merging physics, big data
towards Industry 4.0," Information Fusion, vol. analytics and simulation for the next-generation
50, pp. 92-111, 2019. digital twins," High-performance marine
[26] C. Joshua, L. Jhon, and I. Kola, "Using Anomaly vehicles, pp. 141-151, 2017.
Detection for Predictive Maintenance in Smart [37] A. Cachada et al., "Maintenance 4.0: Intelligent
Manufacturing," 2019. and predictive maintenance system architecture,"
[27] M. Mohammadi, A. Al-Fuqaha, S. Sorour, and in 2018 IEEE 23rd international conference on
M. Guizani, "Deep learning for IoT big data and emerging technologies and factory automation
streaming analytics: A survey," IEEE (ETFA), 2018, vol. 1: IEEE, pp. 139-146.
Communications Surveys & Tutorials, vol. 20, [38] E. Temer and H.-J. Pehl, "Moving toward smart
no. 4, pp. 2923-2960, 2018. monitoring and predictive maintenance of
[28] B. Qolomany et al., "Leveraging machine downhole tools using the industrial Internet of
learning and big data for smart buildings: A Things IIoT," in Abu Dhabi International
comprehensive survey," IEEE access, vol. 7, pp. Petroleum Exhibition and Conference, 2017:
90316-90356, 2019. SPE, p. D031S065R004.
[29] G. Kejela, R. M. Esteves, and C. Rong, [39] R. B. Shetty, "Predictive maintenance in the IoT
"Predictive analytics of sensor data using era," Prognostics and Health Management of
distributed machine learning techniques," in Electronics: Fundamentals, Machine Learning,
2014 IEEE 6th international conference on cloud and the Internet of Things, pp. 589-612, 2018.
computing technology and science, 2014: IEEE, [40] K. M. Ngu, N. Philip, and S. Sahlan, "Proactive
pp. 626-631. and predictive maintenance strategies and
[30] E. LaGrange, "Developing a digital twin: The application for instrumentation & control in oil
roadmap for oil and gas optimization," in SPE & gas industry," International Journal of
offshore Europe conference and exhibition, Integrated Engineering, vol. 11, no. 4, 2019.
2019: SPE, p. D011S003R001. [41] M. Analytics, "The age of analytics: competing
[31] B. R. Barricelli, E. Casiraghi, and D. Fogli, "A in a data-driven world," McKinsey Global
survey on digital twin: Definitions, Institute Research, 2016.
characteristics, applications, and design [42] L. Pullagura, B. Brahma, N. V. Kumari, L.
implications," IEEE access, vol. 7, pp. 167653- Ravikumar, S. K. G. Katta, and R. Chiwariro,
167671, 2019. "Industry 4.0 Design Principles, Technologies,
[32] A. Rasheed, O. San, and T. Kvamsdal, "Digital and Applications," in Computational Intelligence
twin: Values, challenges and enablers," arXiv in Industry 4.0 and 5.0 Applications: Auerbach
preprint arXiv:1910.01719, 2019. Publications, pp. 357-388.
[33] J. Lee, I. Cameron, and M. Hassall, "Improving [43] T. Hong, D. W. Gao, T. Laing, D. Kruchten, and
process safety: What roles for Digitalization and J. Calzada, "Training energy data scientists:
Industry 4.0?," Process safety and environmental universities and industry need to work together
protection, vol. 132, pp. 325-339, 2019. to bridge the talent gap," IEEE Power and
[34] E. Elmies, "Fault Detection of Offshore Wind Energy Magazine, vol. 16, no. 3, pp. 66-73, 2018.
Turbine Drivetrain, State-of-the-Art, [44] F. Tao, M. Zhang, and A. Y. C. Nee, Digital twin
Development Trend and Role of Digital Twin," driven smart manufacturing. Academic press,
NTNU, 2019. 2019.
[35] Y. Li and Z. O’Neill, "A critical review of fault [45] T. P. Raptis, A. Passarella, and M. Conti, "Data
modeling of HVAC systems in buildings," in management in industry 4.0: State of the art and

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 278

 © JUN 2019 | IRE Journals | Volume 2 Issue 12 | ISSN: 2456-8880

open challenges," IEEE Access, vol. 7, pp.

97052-97093, 2019.
[46] M. Abu-Elkheir, M. Hayajneh, and N. A. Ali,
"Data management for the internet of things:
Design primitives and solution," Sensors, vol.
13, no. 11, pp. 15582-15612, 2013.
[47] V. Jirkovský, M. Obitko, and V. Mařík,
"Understanding data heterogeneity in the context
of cyber-physical systems integration," IEEE
Transactions on Industrial Informatics, vol. 13,
no. 2, pp. 660-667, 2016.

IRE 1708643 ICONIC RESEARCH AND ENGINEERING JOURNALS 279