Proceedings of the 2nd International Facilities Engineering & Management Conference,

Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

PREDICTIVE MAINTENANCE FOR ENERGY SYSTEMS IN BUILT

ENVIRONMENTS USING DEEP LEARNING MODELS
Najeem O. Adelakun1, Samuel A. Omolola2
1
Senior Engineer/Researcher, Federal College of Education Iwo, Osun State, Nigeria.
2
Department of Electrpnics Engineering, Federal Polytechnic Ilaro, Ogun State, Nigeria.
[email protected], [email protected]

ABSTRACT
Energy systems in built environments often face inefficiencies due to reactive maintenance, leading to
unexpected failures, higher operational costs, and energy waste. Traditional methods lack the predictive
capabilities needed to optimize energy use, necessitating advanced solutions. Predictive maintenance,
powered by deep learning, offers a data-driven approach to improve energy system performance and extend
equipment lifespan. This study develops a deep learning model to predict maintenance needs, reducing
energy consumption and operational costs. Real-time sensor data from HVAC systems, lighting, and power
units in commercial buildings is analyzed using neural networks to identify patterns and forecast failures.
The model's accuracy is validated through simulations and case studies, demonstrating its effectiveness in
minimizing downtime and enhancing efficiency. The findings highlight significant cost savings and
improved energy usage, emphasizing the value of predictive maintenance in optimizing energy systems.
Facility managers should adopt these models to enhance efficiency and reduce costs, with further research
on model refinement encouraged.
Keywords: Predictive maintenance, Deep learning, Energy systems, Built environments, Facility
management

INTRODUCTION
Energy systems in built environments play a crucial role in ensuring the efficiency, reliability, and
sustainability of modern infrastructure. These systems encompass heating, ventilation, and air
conditioning (HVAC) units, lighting, power distribution networks, and other critical assets that
contribute to occupant comfort and operational effectiveness (Okoye & Adelakun, 2019;
Papadakis & Katsaprakakis, 2023; Rahmanparast et al., 2024). However, maintaining the optimal
performance of these energy systems presents significant challenges due to the increasing
complexity of infrastructure, aging equipment, and growing energy demands (Adelakun &
Olanipekun, 2020; Ahmad et al., 2022; Hassan et at., 2023). Traditional maintenance strategies,
such as reactive and preventive maintenance, have limitations in addressing these challenges, often
resulting in unexpected equipment failures, increased operational costs, and energy inefficiencies
(Achouch et al., 2023; Adelakun, 2024; Selcuk, 2017). As a result, predictive maintenance,
powered by deep learning models, has emerged as a transformative approach to enhancing the
management of energy systems in built environments. Maintenance strategies for energy systems
have evolved over time. Historically, reactive maintenance, which involves repairing equipment
only after a failure occurs, has been the predominant approach in many facilities (Ayanlowo &
Adelakun, 2023). While this method may seem cost-effective in the short term, it leads to
substantial downtime, high repair costs, and potential safety hazards (Feng & Li, 2022; Zhao et
al., 2022). Preventive maintenance, which involves scheduled servicing based on estimated
equipment lifespans, aims to mitigate these risks. However, it is not always efficient, as it does not

https://doi.org/10.5281/zenodo.14849013 52 - 60 52
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

consider real-time operational data and may lead to unnecessary maintenance or premature
component replacements (Gupta et al., 2023; Kumar et al., 2024). Predictive maintenance offers a
data-driven alternative by leveraging machine learning and deep learning models to analyze
historical and real-time sensor data, detect anomalies, and forecast potential failures before they
occur (Adelakun 2023; Chen et al., 2023; Li et al., 2023; Omolola & Adelakun, 2023; Wu et al.,
2024). This approach enables facility managers to schedule maintenance proactively, optimize
resource allocation, and extend the lifespan of energy systems, ultimately reducing energy waste
and operational costs.
Deep learning, a subset of artificial intelligence, has demonstrated remarkable success in predictive
maintenance due to its ability to analyze vast amounts of time-series data and recognize complex
patterns that traditional statistical methods cannot detect (Adelakun, 2024; De Luca et al., 2023;
Ran et al., 2019). Models such as recurrent neural networks (RNNs), long short-term memory
(LSTM) networks, and convolutional neural networks (CNNs) are commonly used to process
sensor data from HVAC systems, lighting, and power units (Adelakun & Baale, 2024). These
models can identify deviations from normal operating conditions, detect early signs of equipment
degradation, and predict future failures with high accuracy (Adelakun, 2024). For instance, LSTM
networks are particularly effective in predictive maintenance because they can capture long-term
dependencies in sequential data, making them suitable for time-series analysis (Gao et al., 2023).
When applied to energy systems, these models can analyze temperature fluctuations, vibration
patterns, and power consumption trends to determine when a component is likely to fail. This level
of predictive insight enables facility managers to implement targeted maintenance interventions,
reducing unplanned downtime and enhancing energy efficiency.
Predictive maintenance using deep learning models has been successfully implemented across
various energy-intensive sectors, including commercial buildings, industrial facilities, and smart
cities (Billanes et al., 2025; Nguyen et al., 2021). In commercial buildings, HVAC systems
consume a substantial portion of total energy usage. Implementing predictive maintenance in these
systems can lead to significant energy savings by identifying faults such as refrigerant leaks,
clogged filters, or sensor malfunctions before they cause inefficiencies (Mulumba, 2014).
Similarly, lighting systems benefit from predictive maintenance by detecting voltage fluctuations,
ballast failures, and sensor degradation in smart lighting networks. Power distribution systems also
leverage predictive analytics to monitor transformer health, cable insulation integrity, and circuit
breaker performance, thereby preventing catastrophic failures and ensuring grid stability
(Abraham et al., 2024; Kuponiyi et al., 2023). Despite its advantages, predictive maintenance faces
several challenges that must be addressed for widespread adoption. One of the primary challenges
is data quality and availability. Effective predictive models require large volumes of high-quality
sensor data, but many older buildings lack the necessary IoT infrastructure to collect such data
efficiently (Patel et al., 2022). Additionally, data preprocessing, including noise reduction, feature
selection, and anomaly detection, is crucial to improving model accuracy and reliability. Another
challenge is model interpretability and trust. Deep learning models, particularly deep neural
networks, are often considered "black-box" systems due to their complex architectures. Facility
managers may struggle to understand how these models generate predictions, leading to reluctance
in adopting AI-driven maintenance strategies (Egwim et al., 2023). Therefore, explainable AI

https://doi.org/10.5281/zenodo.14849013 52 - 60 53
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

(XAI) techniques must be integrated to enhance transparency and provide actionable insights
for decision-makers. Furthermore, the computational complexity of deep learning models requires
significant processing power and storage capacity. While cloud-based platforms offer scalable
solutions, concerns regarding data security, privacy, and latency must be carefully managed (Wang
et al., 2023). Edge computing and federated learning are emerging as potential solutions to mitigate
these challenges by enabling on-site data processing while maintaining security and efficiency.
Predictive maintenance has been increasingly adopted due to advancements in artificial
intelligence (AI) and sensor technology. Various studies highlight the effectiveness of machine
learning and deep learning techniques in predicting failures and optimizing maintenance
schedules. Traditional statistical models, such as regression analysis and time-series forecasting,
have been used for predictive maintenance but often struggle with complex and non-linear data
patterns. Deep learning models, including recurrent neural networks (RNNs), convolutional neural
networks (CNNs), and long short-term memory (LSTM) networks, have demonstrated superior
performance in handling large-scale and dynamic datasets. These models can process time-series
data from energy systems, identifying patterns and predicting failures with high accuracy. Despite
their potential, challenges such as data preprocessing, feature selection, and model interpretability
remain areas of active research. This study builds on existing literature to explore the application
of deep learning models for predictive maintenance in energy systems.
METHODOLOGY
This research involved the collection of real-time sensor data from various energy-consuming
systems within a commercial office building. The dataset captures energy usage information from
multiple systems, including HVAC units, lighting fixtures, office equipment, and ventilation
systems. These systems are integral to energy management in built environments, influencing both
efficiency and sustainability. Hourly power consumption data was recorded for several key
components. These include air conditioning units, electrical sockets, lighting installations,
photocopiers, printers, ceiling fans, standing fans, and a pumping machine. Each component was
monitored to assess its contribution to overall energy usage and identify potential areas for
optimization. The analysis of this data allows for a better understanding of energy distribution
across different systems, enabling more effective energy-saving strategies. In addition to power
consumption measurements, failure prediction and maintenance requirements were also recorded.
These were based on the operational status of each system and historical failure patterns. By
tracking performance trends, the study aimed to enhance maintenance planning, reduce downtime,
and improve system reliability. The findings from this research contribute to the development of
data-driven energy management strategies, which are crucial for improving efficiency, reducing
operational costs, and promoting sustainability in commercial buildings.
RESULTS AND DISCUSSION
The deep learning model successfully provided predictions for system failures and maintenance
requirements based on historical data and real-time sensor inputs. The results are structured into
three key categories: Failure Prediction, Maintenance Prediction, and Visualization.

https://doi.org/10.5281/zenodo.14849013 52 - 60 54
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

The model classified system failure predictions into two categories: 0 (No Failure) and 1
(Failure Detected). It identified potential failures by analyzing historical operational data and
sensor readings. For instance, at specific timestamps (e.g., 2024-10-22 03:00:00), the model
predicted a potential system failure, indicated by a 1 (Yes) in the dataset. Such insights enable
proactive maintenance interventions, reducing unexpected downtimes and optimizing equipment
lifespan. The Maintenance Required column in the dataset flagged components that required
servicing based on detected anomalies in energy consumption patterns. The model identified
increased energy usage linked to potential wear and malfunctioning, triggering maintenance alerts.
From the dataset, four instances were recorded where maintenance was required (1 = Yes), while
the remaining instances indicated no immediate maintenance needs (0 = No). This approach helps
facility managers prioritize preventive maintenance tasks efficiently. To enhance interpretation,
bar charts were generated for both failure and maintenance predictions:

Fig 1: Word Cloud of Energy-Consuming Components

Fig 1 presents a word cloud visualization of the key energy-consuming components analyzed in
this study. The word cloud highlights the most frequently referenced components, with larger text
sizes indicating higher occurrences in the dataset. Prominent terms such as Air Conditioning,
Lighting, Printers, Photocopiers, Fans, and Pumping Machine reflect their significant contribution
to overall energy consumption. This visualization provides a quick, intuitive representation of the
dataset, emphasizing the major systems influencing power usage and facility maintenance
requirements.

https://doi.org/10.5281/zenodo.14849013 52 - 60 55
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

Fig 2: Failure Prediction at Different Timestamps

Fig 2 presents the failure prediction results over time, indicating system conditions at hourly
intervals on October 22, 2024. The dataset classifies system status as 1 (Failure Predicted) or 0
(No Failure) based on historical trends and real-time sensor data. The results show that most
timestamps recorded 0, suggesting normal system operation. However, failures were predicted at
3:00 AM and 4:00 AM, marked by 1 in the dataset. These occurrences suggest possible anomalies
or performance issues during those hours, highlighting critical periods for proactive maintenance
to prevent unexpected breakdowns and optimize system reliability.

Fig 3: Maintenance Prediction at Different Timestamps

https://doi.org/10.5281/zenodo.14849013 52 - 60 56
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

Fig 3 illustrates the predicted maintenance requirements for various time intervals on
October 22, 2024. The dataset categorizes maintenance needs as 1 (Yes, Maintenance Required)
or 0 (No Maintenance Needed) based on system performance data. The results show that most
timestamps recorded 0, indicating normal operation. However, maintenance was predicted at 2:00
AM, 5:00 AM, and 7:00 AM, suggesting potential wear or system inefficiencies during these
periods. These predictions highlight critical windows for preventive maintenance, ensuring timely
interventions to mitigate possible failures, extend equipment lifespan, and enhance overall
operational efficiency. The results of the study indicate variations in power consumption and their
impact on system performance. The total power consumption fluctuates between 8.324 kW and
8.518 kW, with increased consumption observed during specific periods. Failure predictions
occurred at 3:00 AM and 4:00 AM, aligning with higher energy usage, particularly from the
photocopier and printer. Additionally, maintenance requirements were flagged at 2:00 AM, 5:00
AM, and 7:00 AM, suggesting preventive actions are necessary to avoid system failures. The
analysis reveals a correlation between increased power consumption and system failures,
emphasizing the need for continuous monitoring and timely maintenance. These findings highlight
the importance of optimizing power usage and implementing predictive maintenance strategies to
ensure operational efficiency and prevent unexpected equipment breakdowns. By addressing
maintenance needs proactively, reliability can be improved while minimizing downtime and
energy waste.

CONCLUSION
This study demonstrates the effectiveness of predictive maintenance using deep learning models
for energy system management. By analyzing power consumption patterns and failure predictions,
it becomes evident that proactive maintenance can enhance operational efficiency, minimize
downtime, and optimize energy usage. The findings reveal that power consumption fluctuations
correlate with system failures, emphasizing the importance of real-time monitoring and timely
interventions. Predictive maintenance models provide a data-driven approach to identifying
potential failures before they occur, reducing repair costs and improving overall system reliability.
Facility managers and engineers can leverage these models to enhance decision-making, ensuring
that maintenance activities are performed only, when necessary, rather than following rigid
schedules. This approach leads to cost savings, extends equipment lifespan, and enhances
sustainability efforts by reducing unnecessary energy consumption. Future research should focus
on refining the deep learning model by incorporating more advanced architectures such as
transformer-based models or reinforcement learning techniques to improve prediction accuracy.
Additionally, expanding the dataset to include diverse energy systems across multiple facilities
will enhance the model's generalizability. Exploring the integration of IoT sensors and real-time
data processing can further improve the predictive capabilities of the model. Finally, assessing the
economic impact of predictive maintenance adoption across different industries will provide
valuable insights for large-scale implementation.

https://doi.org/10.5281/zenodo.14849013 52 - 60 57
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

REFERENCES
Abraham, Doney, Øyvind Toftegaard, Binu Ben Jose DR, Alemayehu Gebremedhin, and Sule
Yildirim Yayilgan. "Consequence simulation of cyber attacks on key smart grid business
cases." Frontiers in Energy Research 12 (2024): 1395954.
https://doi.org/10.3389/fenrg.2024.1395954
Achouch, M., Dimitrova, M., Ziane, K., Sattarpanah Karganroudi, S., Dhouib, R., Ibrahim, H., &
Adda, M. (2022). On predictive maintenance in industry 4.0: Overview, models, and
challenges. Applied Sciences, 12(16), 8081. https://doi.org/10.3390/app12168081
Adelakun, N. O. (2024, May 14). Exploring the impact of artificial intelligence on information
retrieval systems. Information Matters, Vol. 4, Issue 5.
https://informationmatters.org/2024/05/exploring-the-impact-of-artificial-intelligence-on-
information-retrieval-systems http://dx.doi.org/10.2139/ssrn.4834942
Adelakun, N. O. (2024). Deep Learning Methods for Sentiment Analysis Using BERT and GRU
Models. Lambert Academic Publishing. London, United Kingdom. ISBN: 978-620-8-
01279-3
Adelakun, N. & Baale, A. (2024). Sentiment analysis of financial news using the bert model.
ITEGAM-Journal of Engineering and Technology for Industrial Applications, 10(48), 21-
27. https://doi.org/10.5935/jetia.v10i48.1029
Adelakun, N. O. (2024, January 3). Future trends in artificial intelligence for energy management.
Information Matters, Vol. 4, Issue 1. https://informationmatters.org/2024/01/future-trends-
in-artificial-intelligence-for-energy-management/ http://dx.doi.org/10.2139/ssrn.4703931
Adelakun, N. O. (2023, July 11). Navigating challenges and future trends in sentiment analysis for
investment decision making. Information Matters,
3(7). https://informationmatters.org/2023/07/navigating-challenges-and-future-trends-in-
sentiment-analysis-for-investment-decision-making
http://dx.doi.org/10.2139/ssrn.4513208
Adelakun, N. O. & Olanipekun, B. A. (2020) Outage Analysis and System Disturbances on 330
kV and 132 kV Transmission System in Nigeria, European Journal of Engineering
Research and Science, Vol. 5, Issue 1. Pp 100-
102, https://doi.org/10.24018/ejers.2020.5.1.1722.
Ahmad, T., Madonski, R., Zhang, D., Huang, C., & Mujeeb, A. (2022). Data-driven probabilistic
machine learning in sustainable smart energy/smart energy systems: Key developments,
challenges, and future research opportunities in the context of smart grid
paradigm. Renewable and Sustainable Energy Reviews, 160, 112128.
https://doi.org/10.1016/j.rser.2022.112128
Ayanlowo, O. F., & Adelakun, N. O. (2023). Transforming Facility Engineering Practices in
Nigeria: Insights from Abuja Metropolis. Proceedings of the 1st International Facilities
Engineering & Management Conference, Exhibition & AGM (IFEMCE 2023), The
Nigerian Institution of Facilities Engineering and Management. 26 -
33, https://dx.doi.org/10.2139/ssrn.4748787
Billanes, J. D., Ma, Z. G., & Jørgensen, B. N. (2025). Data-Driven Technologies for Energy
Optimization in Smart Buildings: A Scoping Review. Energies (19961073), 18(2).
https://doi.org/10.3390/en18020290
Chen, C., Fu, H., Zheng, Y., Tao, F., & Liu, Y. (2023). The advance of digital twin for predictive
maintenance: The role and function of machine learning. Journal of Manufacturing
Systems, 71, 581-594. https://doi.org/10.1016/j.jmsy.2023.10.010

https://doi.org/10.5281/zenodo.14849013 52 - 60 58
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

De Luca, R., Ferraro, A., Galli, A., Gallo, M., Moscato, V., & Sperli, G. (2023). A deep
attention based approach for predictive maintenance applications in IoT scenarios. Journal
of Manufacturing Technology Management, 34(4), 535-556.
https://doi.org/10.1108/JMTM-02-2022-0093
Egwim, C. N., Alaka, H., Demir, E., Balogun, H., Olu-Ajayi, R., Sulaimon, I., ... & Muideen, A.
A. (2023). Artificial Intelligence in the Construction Industry: A Systematic Review of the
Entire Construction Value Chain Lifecycle. Energies 2024, 17, 182.
Feng, M., & Li, Y. (2022). Predictive maintenance decision making based on reinforcement
learning in multistage production systems. IEEE Access, 10, 18910-18921.
https://doi.org/10.1109/ACCESS.2022.3151170
Gao, Z., Chen, J., Wang, G., Ren, S., Fang, L., Yinglan, A., & Wang, Q. (2023). A novel
multivariate time series prediction of crucial water quality parameters with Long Short-
Term Memory (LSTM) networks. Journal of Contaminant Hydrology, 259, 104262.
https://doi.org/10.1016/j.jconhyd.2023.104262
Gupta, V., Mitra, R., Koenig, F., Kumar, M., & Tiwari, M. K. (2023). Predictive maintenance of
baggage handling conveyors using IoT. Computers & Industrial Engineering, 177, 109033.
https://doi.org/10.1016/j.cie.2023.109033
Hassan, Q., Algburi, S., Sameen, A. Z., Salman, H. M., & Jaszczur, M. (2023). A review of hybrid
renewable energy systems: Solar and wind-powered solutions: Challenges, opportunities,
and policy implications. Results in Engineering, 101621.
https://doi.org/10.1016/j.rineng.2023.101621
Kumar, R., Sangwan, K. S., Herrmann, C., & Ghosh, R. (2024). Development of a cyber physical
production system framework for smart tool health management. Journal of Intelligent
Manufacturing, 35(7), 3037-3066. https://doi.org/10.1007/s10845-023-02192-3
Kuponiyi, D. S., Olajide, M. B., Eko, M. A., Odeyemi, C. S. & Adelakun, N. O. (2023). Design
Analysis of Microgrid Power System for Telecommunication Industries in
Nigeria, International Journal of Electrical Engineering and Computer Science, vol. 5, pp.
135-156, 2023, https://doi.org/10.37394/232027.2023.5.15
Li, Z., He, Q., & Li, J. (2024). A survey of deep learning-driven architecture for predictive
maintenance. Engineering Applications of Artificial Intelligence, 133, 108285.
https://doi.org/10.1016/j.engappai.2024.108285
Mulumba, T. (2014). Predictive Maintenance: Fault Detection and Diagnosis in HVAC
systems (Doctoral dissertation, Masdar Institute of Science and Technology).
Nguyen, H. P., Le, P. Q. H., Pham, V. V., Nguyen, X. P., Balasubramaniam, D., & Hoang, A. T.
(2021). Application of the Internet of Things in 3E (efficiency, economy, and environment)
factor-based energy management as smart and sustainable strategy. Energy Sources, Part
A: Recovery, Utilization, and Environmental Effects, 1-23.
https://doi.org/10.1080/15567036.2021.1954110
Okoye, C. U. & Adelakun, N. O. (2019), Design and Evaluation of Electrical Services for an
Energy Efficient Home, Iconic Research and Engineering Journals, Vol. 3, Issue 6, Pp 95
– 102, https://dx.doi.org/10.2139/ssrn.3579943
Omolola, S. A. & Adelakun, N. O. (2024). Design and Implementation of a Remote Control Based
Home Automation System. Andalasian International Journal of Applied Science,
Engineering and Technology, 4(1), 13-19. https://doi.org/10.25077/aijaset.v4i1.115
Papadakis, N., & Katsaprakakis, D. A. (2023). A review of energy efficiency interventions in
public buildings. Energies, 16(17), 6329. https://doi.org/10.3390/en16176329

https://doi.org/10.5281/zenodo.14849013 52 - 60 59
 Proceedings of the 2nd International Facilities Engineering & Management Conference,
Exhibition, AGM (IFEMCE 2024), The Nigerian Institution of Facilities Engineering and
Management, 1st – 2nd November, 2024.

Patel, V. A., Bhattacharya, P., Tanwar, S., Gupta, R., Sharma, G., Bokoro, P. N., & Sharma,
R. (2022). Adoption of federated learning for healthcare informatics: Emerging applications
and future directions. IEEE Access, 10, 90792-90826.
https://doi.org/10.1109/ACCESS.2022.3201876
Rahmanparast, A., Bacak, A., Camci, M., Karakoyun, Y., Acikgoz, O., & Dalkilic, A. S. (2024).
A comparison of heating and cooling systems having radiant and ventilation systems
regarding thermal comfort. Journal of Thermal Analysis and Calorimetry, 149(23), 13551-
13579. https://doi.org/10.1007/s10973-024-13771-8
Ran, Y., Zhou, X., Lin, P., Wen, Y., & Deng, R. (2019). A survey of predictive maintenance:
Systems, purposes and approaches. arXiv preprint arXiv:1912.07383, 1-36.
Selcuk, S. (2017). Predictive maintenance, its implementation and latest trends. Proceedings of the
Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, 231(9),
1670-1679. https://doi.org/10.1177/0954405415601640
Wang, T., Liang, Y., Shen, X., Zheng, X., Mahmood, A., & Sheng, Q. Z. (2023). Edge computing
and sensor-cloud: Overview, solutions, and directions. ACM Computing Surveys, 55(13s),
1-37. https://doi.org/10.1145/3582270
Wu, Y., Sicard, B., & Gadsden, S. A. (2024). Physics-informed machine learning: a
comprehensive review on applications in anomaly detection and condition
monitoring. Expert Systems with Applications, 124678.
https://doi.org/10.1016/j.eswa.2024.124678
Zhao, J., Gao, C., & Tang, T. (2022). A review of sustainable maintenance strategies for single
component and multicomponent equipment. Sustainability, 14(5), 2992.
https://doi.org/10.3390/su14052992

https://doi.org/10.5281/zenodo.14849013 52 - 60 60