AI Based PDM in Manufacturing Industry 4.0: A Bibliographic Review

AI Based PDM in Manufacturing Industry 4.0: A Bibliographic Review

Abstract-Predictive maintenance (PdM) techniques backed by data analytics and artificial intelligence (AI) have become increasingly popular in today's dynamic manufacturing environment as a game-changing way to improve equipment longevity, operational effectiveness, and competitiveness. In order to clarify the revolutionary effects of artificial intelligence (AI), data analytics, and predictive maintenance on maintenance procedures, this study explores the complex interactions between these three technologies in the industrial sector. This study synthesizes existing knowledge, finds gaps, and extracts insights critical to comprehending the changing predictive maintenance landscape through an exhaustive examination of the literature from 2014 to 2024. The effectiveness of several AI algorithms, such as logistic regression, support vector regression, random forests, neural networks, and linear regression, is assessed in relation to predictive manufacturing. The research delves into various machine learning algorithms to see which one is most appropriate for addressing predictive maintenance problems in manufacturing environments. Additionally, the study looks at optimization techniques to boost the accuracy and efficacy of AI-

driven maintenance forecasts, utilizing data analytics insights for better maintenance scheduling. Real-time insights and predictive capabilities are provided by the integration of Big Data, IoT, and cyber-physical systems, which transforms maintenance operations in the context of Industry 4.0. Experience-based, model-based, physics-based, data-driven, and hybrid methods to PdM implementation are examined, taking into account their distinct needs and capacities. Additionally, the study looks into how Industry 4.0 technologieslike robotics, cloud computing, augmented reality, and IIoTcan help with predictive maintenance tasks.

The research's insights enhance our understanding of predictive maintenance in the context of Industry 4.0 and provide practitioners, scholars, and industry stakeholders with important direction as they navigate the intricate terrain of maintenance optimization and digital transformation.

Keywords: Predictive Maintenance, Artificial Intelligence, Manufacturing

1. INTRODUCTION

   The maintenance strategy paradigm has seen tremendous evolution in the dynamic manufacturing context, moving from reactive tactics to proactive methodologies. Utilizing data analytics and artificial intelligence (AI), predictive maintenance is a game-changing idea that maximizes equipment durability, boosts operational effectiveness, and minimizes downtime. The present study aims to investigate the complex interactions of predictive maintenance, artificial intelligence, and data analytics within the manufacturing industry. In the current economic environment, which is characterized by intense globalization and markets that are becoming more demanding, industries are under pressure to increase the productivity and efficiency of their production lines in order to boost their competitiveness and please consumers. Industry 4.0 was created by connectivity, data, new gadgets, inventory reduction, customization, and controlled production. As of right now, this trend appears inexorable [27]. This suggests that in order to integrate all of the new technologies and therefore boost production, automation techniques must be used [2]. Through the introduction of cognitive automation and the subsequent

   implementation of the concept of intelligent production, which results in intelligent products and services, the technologies of the Internet of Things (IoT) and Big Data, as well as the integration of artificial intelligence (AI) methods and cyber-physical systems (CPS), play a significant role in this context. Businesses are able to tackle the demands of a

   much more dynamic environment thanks to this creative strategy. Through the use of predictive maintenance, businesses can extend the life of their equipment, prevent unscheduled downtime, and reduce energy usage and expenses [3]. With an estimate for 2024 to 2030, Figure 1 depicts the global trend toward the use of this kind of smart maintenance. Modern maintenance techniques generally differ based on the various learning models applied and the various issues that the machinery or equipment faces [4].

   Figure 1. Size of Maintenance 4.0 market worldwide in 2020 and 2021 with forecast for future years indexed by * (from 2024 to 2030) [5,6,7]

   1. Background :

      Reactive maintenance methods were the norm in the past, which frequently led to unscheduled downtime and rising expenses. Predictive maintenance, which uses cutting-edge technologies to detect equipment problems before they happen, completely changed this environment. This move toward predictive techniques stems from the realization that data may be a potent prediction tool if used properly.

   2. Objectives: (After a Singler line objectives should be point wise)

      This study's major goal is to give manufacturers a thorough grasp of predictive maintenance while highlighting the revolutionary advantages that come from combining data analytics and artificial intelligence (AI). This study attempts to clarify the revolutionary influence on maintenance processes by carefully examining industry applications, fundamental data analytics concepts, and the implementation of machine learning models. The objective is to offer practitioners, researchers, and industry stakeholders a

      comprehensive understanding of the changing predictive maintenance scene, including real-world case studies, moral issues, and a look at potential future developments.

      Examining the Body of Existing Literature: This study will perform a comprehensive evaluation of the body of knowledge on predictive maintenance in order to provide a strong foundation. We seek to synthesize the existing state of knowledge, identify knowledge gaps, and extract useful insights that contribute to the expanding discourse on predictive maintenance through the analysis of academic articles, industry reports, and case studies.

      Utilizing AI for Predictive Manufacturing: The research will investigate the many uses of artificial intelligence (AI) in predictive maintenance, examining the efficacy of neural networks, logistic regression, support vector regression, random forests, and linear regression. We want to illustrate the advantages and disadvantages of every AI algorithm by analyzing real-world examples and industry use cases, giving manufacturers a comprehensive grasp of how each one might be applied in various situations.

      Identifying the Algorithm that Best Fits: Evaluating and contrasting different machne learning algorithms to see which one best suits the unique requirements and difficulties of predictive maintenance in manufacturing is a crucial component of this research. By conducting a thorough examination of algorithmic performance indicators, our goal is to assist practitioners in choosing the best model for their particular operational settings.

      Optimizing the Outcome: Predictive maintenance outcomes are optimized as a key focus of the study. This entails investigating methods to improve the precision and efficacy of AI-driven forecasts, utilizing data analytics insights to improve maintenance plans, and tackling moral dilemmas related to the application of optimized maintenance practices.

      Identifier	Issue
      MQ	Available models, methods, or flowchart for predictive in the Industry 4.0
      SQ1	How to disintegrate research with Industry curriculum?
      SQ2	Common modes of predictive models
      SQ3	Generating a proposed predictive application or monitoring?
      SQ4	Validation of the Results and Implications
      SQ5	Challenges opportunities and others

      Table 1.1. Research questions.

      Figure 2. Components of Modeling the Framework Literature Review

      • Scaife. A (2024) has responded to the question of how Artificial Intelligence (AI) can be used in conjunction with predictive maintenance to lower operations and maintenance costs in facility operations. The paper has presented some research that shows that it can reduce facility operations costs but more research is required to determine the progressive implications of AI in PdM in the industry.

      • Gawde S et. al (2022) have examined that there is still tremendous space for growth in the field of multi-fault diagnostics in Industry 4.0. The study lists the fundamental research in the subject and provides a comparative analysis of several elements pertaining to the diagnosis of multiple faults in industrial rotating machinery. The main obstacles and research gaps are also identified in the report. It provides answers by using the latest developments in AI to multi-fault detection, providing a solid foundation for further study in this area.

      • Peng et al. (2020) has demonstrated advancements, yet difficulties persist. As of right now, the primary data source is real-time sensor data, with limited means for

        computation and fusion. Assessments of the health state of equipment produce qualitative data with little room for interpretation or modification. The integration of equipment and a high degree of health indicator dimension are essential to the achievement of intelligent manufacturing. A new industrial revolution has been sparked by the industry's use of physical information systems, necessitating assessment and forecasting based on numerous health indicators. Building operational models is challenging, as component connections are intricate. Thus, it's critical to research intelligent monitoring technologies for industrial machinery in order to achieve real-time, full- state, and whole-process

      • Lee et. al (2019) have studied and researched the development of PdM systems utilizing the AI algorithms to indicate the component's conditions (normal, warning, and failure) in the systems, is the RUL of the bearing and the gap wear of the cutting tool, respectively. To classify the state of the tools, the SVM and ANNs (RNN and CNN) approaches are used in conjunction with several feature extraction techniques.

      • Khalastchi And Kalech() have presented that robots are increasingly used in daily life, performing tasks that are too dangerous or dull for humans. Fault Detection and Diagnosis (FDD) has been studied aimed to provide insights on FDD approaches for robotic systems, focusing on their characteristics, advantages, and challenges by examining FDD from the perspective of different characteristics and successful approaches, and analyzing the advantages and disadvantages of each approach. They also discuss research opportunities for robotic systems' FDD, introducing readers to the field of FDD.

      • Lee et. al (2017) have investigated that a sensor- monitored semiconductor manufacturing plant's maintenance decision support system process is simulated, and a Maintenance Policies Management framework under Big Data Platform is built. To categorize probable failure patterns and determine the machine status of the malfunctioning component, artificial intelligence is utilized.

      • Thoben et. al (2017) have presented paradigm changes in the manufacturing industries on data-driven, information and communication technology integration across the supply network, heightened automation, and energy conservation, sustainability, agility, and quality enhancements. Applications of smart manufacturing include the use of video streams and operational data from airplane cabin surveillance systems for improved archiving and analysis services, human-robot interaction on the shop floor, and CPLS for intralogistics. More research questions will surface as these projects gather momentum, and quick advancements are anticipated soon. Researchers may work with allied fields and sectors as 14.0 and smart manufacturing continue to gain traction in order to see their findings applied in practical settings.

      • Li et. al (2017) that the last few decades have seen an extremely high rate of research and development into the identification and prognosis of faults in mechanical systems. But because machine centers are so complex, there is still much work to be done in this area to improve the precision and dependability of data mining-based defect diagnosis and prognosis. In the context of Industry 4.0, the study looks into defect detection and prognosis in machine centers using data mining techniques in order to develop a methodical methodology and gather information for predictive maintenance. We present a system framework that incorporates the fault analysis and treatment process for predictive maintenance in machine centers, based on the principles of Industry 4.0. The framework consists of five modules: data preprocessing, data mining, sensor selection, and data collecting.

      • Li et. al (2016) have studied a completely novel interactive clustering method that enables domain experts steer the cluster constructing process in bearing deficiencies diagnosis. The approach that permits domain knowledge to direct it, generalizes an otherwise unbiased clustering algorithm into a biased one using shrinkage. A desired level of granularity and a specific location can be chosen for in-depth analysis by the domain expert. The adopted approach performs better than the impartial fuzzy c-means algorithm.

      • Vog et. al (2016) have reviewed the challenges, needs, methods, and best practices for Prognostics and Health

        Management (PHM) in manufacturing systems. It highlights the need for PHM capabilities that overcome current challenges and meet future needs. The paper also highlights the benefits of open-system architectures, cost-benefit analyses, method verification and validation, and standards in PHM systems.

        • Kang et. al (2016) have enhanced the competitiveness of the industrial revolution, enabling precise and efficient engineering decision-making in real time by integrating a variety of ICT technologies with already- existing manufacturing technologies. The study assessed and examined a number the Smart Manufacturing-related trends were identified and the future was predicted by conducting various analyses on the application areas and technoogy development levels

        • Lee et. al (2015) have presented a 5C architecture for Cyber-Physical Systems in Industry 4.0 manufacturing systems. It provides a viable and practical guideline for the manufacturing industry to implement CPS for better product quality and system reliability with more intelligent and resilient manufacturing equipment.

        • Wang et. al (2014) have presented a method for classifying and fusing multiple member classifiers to create a robust model for use in health diagnostics. The methodology includes a weighted majority vote with dominant technique, several multi-attribute classifiers as member algorithms, and a k-fold cross validation model. The 2008 PHM challenge problem shows that the fusion strategy performs better in terms of diagnostic accuracy and resilience than stand-alone algorithms.

        • Kaiser and Gebraeel (2009) have presented a predictive maintenance strategy based on sensory-updated deterioration (also known as the SUDM policy). The proposed maintenance policy makes use of state-of- the-art degradation models that predict and update the residual life distribution (RLD) by combining component-specific real-time degradation signals obtained during operation with reliability and degradation characteristics of the component population.

      Figure 3 Research Trend in the topic over Past Ten Years (2011-2023) [8]

      Figure 4 Distribution based on type of research [10]

      Industry 4.0

      In the transition from Industry 3.0 to the era of Industry 4.0, we witnessed a profound transformation where computers and industrial equipment are seamlessly connected, communicating and making decisions autonomously. This paradigm shift is made possible by the integration of cyber- physical systems, the Internet of Things (IoT), and the Internet of Systems. The resulting smart factory represents a significant leap forward, with machines becoming increasingly intelligent through the continuous accumulation of data. This digital interconnectedness, facilitated by the Internet of Things, Big Data, computer intelligence, and decision-making systems, revolutionizes the industrial landscape [9]. Industry 4.0 empowers companies to respond swiftly to market changes, deliver personalized products, and enhance operational efficiency. The real-time insights provided by informed Big Data play a crucial role in optimizing the entire value chain [10]. Technologies like the Industrial Internet of Things (IIoT) and physical network systems are indispensable components for collecting, processing, and storing data [11]. To fully leverage these advancements, the Industry 4.0 model advocates for the use of multidisciplinary technologies, acknowledging that some of these innovations, such as the Internet of Things (IoT), blockchain technology, big data, and artificial intelligence (AI), have already demonstrated their potential in revolutionizing sectors like healthcare and manufacturing [12]. In healthcare, Industry 4.0 promises precise and personalized services, employing predictive data analysis for early disease detection. The integration of blockchain technology safeguards patient data collected through IoT- connected devices. In manufacturing, Industry 4.0 facilitates more responsive value chains, fostering greater integration between manufacturers and customers [13]. This transformative wave extends its impact across diverse economic sectors, raising concerns about its implications on employment and stimulating debates on policies to support digital transformation [13]. Industry 4.0, encompassing technologies such as the Internet of Things, Big Data, Artificial Intelligence, Cloud computing, Cybersecurity,

      Blockchain technology, Additive manufacturing, Autonomous robots, and Augmented reality, is poised to influence sectors like the wood industry globally [12]. Despite its potential benefits, the construction industry has been slow to embrace Industry 4.0, prompting a comprehensive exploration of its implications, including political, economic, social, technological, environmental, and legal considerations [14]. These encourage highlight the taking after stages to form up the method stream for planning PdM framework:

      • Information securing and pre-processing: Collecting, cleaning and fittingly putting away tall quality information;

      • Diagnostics and Corruption location: Distinguishing components that show signs of inconsistencies, not performing ideally, or that a disappointment is up and coming. It moreover includes understanding causes of disappointment finding its root causes, disappointment classification into sorts and modes are also made here;

      • Prognostics: Foreseeing long-term wellbeing of a framework in order to maintain a strategic distance from potential disappointments. Models and exactness of expectations are vital here; and

      • Support choices making: Choosing on best approaches for maintenance planning, will be made based on the result of the prescient result higher surrender and lessening misfortunes.

        Figure 5. Stages in Data Analysis and Modelling

2. PDM IN MANUFACTURING INDUSTRY 4.0

   In the context of Industry 4.0, Predictive Maintenance (PdM) stands as a pivotal strategy leveraging digitized sensor data and advanced analytics to continuously monitor the condition of machine components or processes, aiming to ascertain when and where maintenance might be necessary [27]. The application of PdM has been noted to effectively minimize maintenance costs while maximizing the operational lifespan of devices [27]. Industry 4.0, characterized by the integration of technologies like the Internet of Things (IoT), Big Data, Cyber-Physical Systems (CPS), and Artificial Intelligence (AI), facilitates the collection and analysis of crucial data such as machine operation parameters (e.g., speed, power) and environmental factors (e.g., humidity) [27].

   PdM, realized through the analysis of such data, holds the promise of enhancing maintenance operations, reducing costs, improving product quality, enhancing customer satisfaction, and potentially generating new sources of income [27]. However, despite the widely recognized benefits of PdM, its practical implementation faces several challenges, with one of the main hurdles being the lack of systematic methods to assist enterprises in selecting the most suitable PdM technique for their specific requirements [27]. In practice, there are five primary types of PdM techniques employed: experience-based, model-based, physics-based, data-driven, and hybrid approaches [27]. Each technique has unique requirements concerning hardware, software, and information, leading to variations in implementation cost, time, and capability spectrum [27]. The capability spectrum of PdM techniques encompasses tasks such as accurate component state prediction, fault prognosis, and asset service life extension [27]. Therefore, the selection of the most appropriate PdM technique for a given scenario is crucial for optimizing costs, time, and ensuring long-term financial viability and return on investment [27].

   Experience-based PdM techniques draw on the knowledge and expertise gained through practical experience. Human experts develop specific rules and insights over years of maintaining technical systems [29]. These techniques leverage tacit knowledge to identify faults, describe component wear and tear, and predict failures [28]. While experience-based techniques offer explicative results and can be implemented at modest costs, their prognostic capabilities are limited [28].

   Data-driven PdM techniques, on the other hand, utilize sensor data from IoT-enabled manufacturing systems to monitor operational parameters and assess component wear [27]. Various machine learning approaches such as Artificial Neural Networks (ANN), Support-Vector Machines (SVM), andDecision Trees (DT) are employed to build predictive models with high accuracy in fault prediction and estimating remaining useful life (RUL) [29-33].

   Model-based PdM techniques rely on mathematical models to describe the degeneration of systems or components. These techniques use residuals as features, comparing measured results with expected behavior to diagnose faults and predict remaining service life [34]. While model-based techniques offer high prediction accuracy and reusability, they are

   computationally expensive and require advanced mathematical understanding for development [34].

   Physics-based PdM techniques, grounded in the laws of physics, simulate component wear and tear using physics behavior models. These techniques provide highly accurate predictions but are limited in their ability to describe certain wear phenomena and may be influenced by external factors such as temperature and pressure [35].

   Hybrid PdM techniques combine multiple approaches to address the complexity of modern manufacturing systems. By integrating different PdM techniques, hybrid approaches aim to improve overall effectiveness, although they may entail higher implementation costs [28,36].

   Despite the general characteristics outlined above, the successful implementation of PdM depends on various factors, including the specific machinery involved, types of wear and tear to be monitored, implementation costs, staff training expenses, long-term strategies, and return on investment considerations [27]. To aid decision-makers in industry, a decision-based framework has been proposed to facilitate the selection and implementation of the most appropriate PdM strategy and solutions for individual needs [27].

   PdM techniques play a vital role in the maintenance and optimization of industrial equipment in the Industry 4.0 era. Each technique offers distinct advantages and challenges, requiring careful consideration based on specific operational requirements. The development of systematic methods and decision-making frameworks is essential to support enterprises in effectively implementing PdM strategies and maximizing the benefits of predictive maintenance.

3. DATA ANALYTICS AND BIG DATA ANALYTICS

   The significance of Data Analytics and Big Data Analytics in influencing industry futures is growing as technology advances. These approaches are the compass that leads companies through the complexity of the data-driven environment; they are more than just catchphrases. Organizations that adopt these analytics methodologies will be able to make sense of the present while also laying the foundation for a day when data will be used strategically to drive innovation, competitiveness, and long-term success.

   5.1 The Tools and Technologies of Big Data Analytics Many techniques and technologies are used by businesses to extract insights from large datasets. While Apache Spark speeds up data processing, Apache Hadoop enables distributed storage and processing. NoSQL databases are adept at managing a wide range of data kinds. The potential of Big Data Analytics is further enhanced by machine learning algorithms and artificial intelligence, which allow for real-time analysis and decision-making.

4. DATA COLLECTION AND PREPROCESSING

   Preprocessing and data collecting are the hidden heroes of the data analytics process. This important stage ensures that the data being examined is relevant and polished for best outcomes, laying the groundwork for solid analyses.

   1. Data Collection: The initial phase entails gathering information from various sources. Semi-structured data from APIs, unstructured data from social media, and structured data from databases may all fall under this category. The

      strategic choice of data sources is made in accordance with the particular goals of the analysis. The source material needs to be thorough and representative, regardless of whether it's sensor data, financial transactions, or consumer behavior.

   2. Data Preprocessing: After obtaining the raw data, the preparation phase begins. To prepare the data for analysis, this entails organizing, converting, and cleaning it. Among the tasks are standardizing formats, eliminating duplicates, and addressing missing values. In order to maintain consistency, data normalization and scaling could be required, particularly when working with variables that have different scales.

   3. Data Analysis: In Industry 4.0, manufacturing equipment and machinery are equipped with various sensors that continuously collect data on performance metrics such as temperature, vibration, pressure, and more. Data analysis is the foundation of PdM, providing the raw material.

   4. Data Post Processing: Once the models are trained and validated, they can be deployed to make predictions in real- time using data post processing. These predictions indicate the likelihood of a machinery failure or maintenance requirement within a certain time window. Maintenance decisions can then be made based on these predictions, such as scheduling preventive maintenance activities or triggering alerts for imminent failures.

5. FEATURE ENGINEERING FOR PREDICTIVE MAINTENANCE

   In predictive maintenance, feature engineering is crucial to converting unprocessed sensor data into useful insights. This method improves the accuracy of predictive models by combining temporal featuressuch as the duration since the last maintenance and the failure historywith sensor data and environmental factors. In the end, firms may maximize operational efficiency in industrial settings by proactively maintaining equipment, minimizing downtime, and anticipating equipment problems through the carefully planned provision of relevant features.

6. RESULTS AND DISCUSSION

   The integration of predictive maintenance (PdM) strategies empowered by artificial intelligence (AI) and data analytics within the manufacturing sector represents a significant paradigm shift, revolutionizing maintenance processes and optimizing operational efficiency. This section presents the results and discussions derived from an extensive literature review spanning from 2014 to 2024, focusing on the revolutionary impacts of predictive maintenance, the role of AI algorithms, and the integration of Industry 4.0 technologies.

   Figure 6. Common approaches for PdM

   1. Revolutionizing Predictive Maintenance

      The transition from reactive maintenance methods to predictive maintenance (PdM) strategies has been pivotal in reshaping maintenance practices within the manufacturing industry. Traditional reactive maintenance approaches often led to unscheduled downtime and escalating expenses. However, the adoption of predictive maintenance, enabled by cutting-edge technologies such as AI and data analytics, has completely transformed this landscape. Predictive maintenance emphasizes proactive measures by leveraging real-time data analysis to detect equipment problems before they occur. This proactive approach not only minimizes downtime but also maximizes equipment durability and operational effectiveness.

      The evolution towards predictive maintenance is driven by the recognition of data as a potent prediction tool when utilized effectively. By harnessing data analytics and AI algorithms, manufacturers can gain valuable insights into equipment conditions, enabling them to anticipate maintenance needs and optimize maintenance schedules. Moreover, predictive maintenance allows businesses to extend the lifespan of their equipment, reduce energy usage, and lower maintenance expenses, thereby enhancing overall operational efficiency.

   2. AI Integration in Predictive Maintenance

      The integration of artificial intelligence (AI) algorithms plays a crucial role in enhancing he efficacy of predictive maintenance strategies. Various AI algorithms, including neural networks, logistic regression, support vector regression, random forests, and linear regression, have been employed to analyze real-time data and predict equipment failures. These algorithms offer distinct advantages and disadvantages, and their suitability varies based on the specific requirements of manufacturing operations.

      Neural networks, for instance, exhibit strong pattern recognition capabilities, making them well-suited for complex predictive maintenance tasks. Logistic regression, on the other hand, is valuable for binary classification problems, where the objective is to predict whether a failure will occur or not. Support vector regression and random forests are effective for both classification and regression tasks, offering robust predictive capabilities. Linear regression, while simpler compared to other algorithms, can still provide valuable insights into equipment performance trends.

   3. Industry 4.0 Integration

      The integration of Industry 4.0 technologies further enhances the capabilities of predictive maintenance in the manufacturing sector. Industry 4.0, characterized by the convergence of cyber-physical systems, the Internet of Things (IoT), and advanced data analytics, facilitates real- time data collection, analysis, and decision-making. This digital interconnectedness enables manufacturers to monitor equipment conditions comprehensively, anticipate maintenance needs, and optimize operational processes.

      In the context of Industry 4.0, predictive maintenance (PdM) stands as a pivotal strategy leveraging digitized sensor data and advanced analytics to continuously monitor the condition of machine components or processes. The application of PdM has been noted to effectively minimize maintenance costs while maximizing the operational lifespan of devices. Industry 4.0 technologies such as IoT, Big Data, Cyber- Physical Systems (CPS), and AI facilitate the collection and analysis of crucial data, enabling proactive maintenance interventions.

   4. Challenges and Opportunities

   Despite the transformative potential of predictive maintenance and AI integration within the manufacturing industry, several challenges persist. One of the main hurdles is the lack of systematic methods to assist enterprises in selecting the most suitable PdM technique for their specific requirements. The selection of the appropriate AI algorithm and PdM technique depends on various factors, including equipment types, maintenance objectives, and available resources.

   Furthermore, the practical implementation of predictive maintenance faces challenges such as data availability, algorithm complexity, and integration with existing systems. Obtaining labeled data for training AI models can be challenging, particularly for rare or unpredictable equipment failures. Moreover, the complexity of AI algorithms and the need for specialized expertise pose additional barriers to implementation.

   However, these challenges also present opportunities for innovation and advancement within the manufacturing sector. Research and development efforts focused on overcoming data availability issues, improving algorithm performance, and streamlining implementation processes can drive further progress in predictive maintenance and AI integration. Moreover, collaboration between academia, industry, and technology providers is essential to develop standardized methodologies, best practices, and decision support frameworks for predictive maintenance implementation.

7. CONCLUSION

In conclusion, the integration of predictive maintenance strategies, artificial intelligence, and Industry 4.0 technologies represents a revolutionary approach to maintenance optimization within the manufacturing sector. By leveraging real-time data analytics and AI algorithms, manufacturers can proactively monitor equipment conditions, anticipate maintenance needs, and optimize operational processes. Despite challenges such as data availability and algorithm complexity, ongoing research and collaboration efforts hold promise for further advancements in predictive maintenance and AI integration. Ultimately, the adoption of predictive maintenance strategies empowered by AI and Industry 4.0 technologies is poised to enhance operational efficiency, reduce maintenance costs, and drive competitiveness within the manufacturing industry.

REFERENCES

1. Rüßmann, M.; Lorenz, M.; Gerbert, P.; Waldner, M.; Justus, J.; Engel, P.; Harnisch, M. Industry 4.0: The Future of Productivity and Growth in Manufacturing Industries; Boston Consulting Group: Boston, MA, USA, 2015.

2. Jasiulewicz-Kaczmarek M, Gola A. Maintenance 4.0 Technologies for Sustainable Manufacturing – An Overview. IFAC-PapersOnLine. 2019;52(10):91-96. doi:10.1016/j.ifacol.2019.10.005

3. Ran Y. A Survey of Predictive Maintenance: Systems, Purposes and Approaches. arXiv.org. Published December 12, 2019. http://arxiv.org/abs/1912.07383

4. Achouch M, Dimitrova M, Ziane K, Karganroudi SS, Dhouib R, Ibrahim H, et al. On Predictive Maintenance in Industry 4.0: Overview, models, and challenges. Applied Sciences. 2022;12(16):8081. doi:10.3390/app12168081

5. Statista. Global predictive maintenance market size 2020-2030. Statista. Published August 29, 2023.

   https://www.statista.com/statistics/748080/global-predictive- maintenance-market-size

6. Predictive Maintenance Market Size, share & Trends Analysis Report by component, by solution, by service, by deployment, by enterprise size, by monitoring technique, by end-use, by region, and segment Forecasts, 2023 – 2030. Published September 4, 2023.

   https://www.grandviewresearch.com/industry-analysis/predictive- maintenance-market

7. Predictive maintenance market. Published August 28, 2023.

   https://www.futuremarketinsights.com/reports/predictive- maintenance-market

8. Gawde S, Patil S, Kumar S, Kamat P, Kotecha K, Abraham A. Multi- fault diagnosis of Industrial Rotating Machines using Data-driven approach : A review of two decades of research. Engineering Applications of Artificial Intelligence. 2023;123:106139. doi:10.1016/j.engappai.2023.106139

9. Li Z, Wang Y, Wang K. Intelligent predictive maintenance for fault diagnosis and prognosis in machine centers: Industry 4.0 scenario. Advances in Manufacturing. 2017;5(4):377-387. doi:10.1007/s40436- 017-0203-8

10. Silvestri L, Forcina A, Introna V, Santolamazza A, Cesarotti V. Maintenance transformation through Industry 4.0 technologies: A systematic literature review. Computers in Industry. 2020;123:103335. doi:10.1016/j.compind.2020.103335

11. Vaidya S, Ambad PM, Bhosle S. Industry 4.0 a glimpse. Procedia Manufacturing. 2018;20:233-238. doi:10.1016/j.promfg.2018.02.034

12. Mohamed AMO, Menezes BC, AlAnsari T. Interplaying of food supply chain resilience, industry 4.0 and sustainability in the poultry market. In: Computer-Aided Chemical Engineering. ; 2021:1815-1820. doi:10.1016/b978-0-323-88506-5.50281-3

13. Galati F, Bigliardi B. Industry 4.0: Emerging themes and future research avenues using a text mining approach. Computers in Industry. 2019;109:100-113. doi:10.1016/j.compind.2019.04.018

14. Oesterreich TD, Teuteberg F. Understanding the implications of digitisation and automation in the context of Industry 4.0: A triangulation approach and elements of a research agenda for the construction industry. Computers in Industry. 2016;83:121-139. doi:10.1016/j.compind.2016.09.006

16. Butt J. A Strategic Roadmap for the manufacturing industry to implement Industry 4.0. Designs. 2020;4(2):11. doi:10.3390/designs4020011

17. Moktadir MdA, Ali SM, KusiSarpong S, Shaikh MdAA. Assessing challenges for implementing Industry 4.0: Implications for process safety and environmental protection. Process Safety and Environmental Protection. 2018;117:730-741. doi:10.1016/j.psep.2018.04.020

18. Sharma M, Kamble SS, Mani V, Sehrawat R, Belhadi A, Sharma V. Industry 4.0 adoption for sustainability in multi-tier manufacturing supply chain in emerging economies. Journal of Cleaner Production. 2021;281:125013. doi:10.1016/j.jclepro.2020.125013

19. Stricker N, Pfeiffer A, Moser E, Kádár B, Lanza G, Monostori L. Supporting multi-level and robust production planning and execution.

    CIRP Annals. 2015;64(1):415-418. doi:10.1016/j.cirp.2015.04.115

20. Aburakhia S, Shami A. SB-PdM: A tool for predictive maintenance of rolling bearings based on limited labeled data. Software Impacts. 2023;16:100503. doi:10.1016/j.simpa.2023.100503

21. Yang L, Shami A. On hyperparameter optimization of machine learning algorithms: Theory and practice. Neurocomputing. 2020;415:295-316. doi:10.1016/j.neucom.2020.07.061

22. Attoui I. Novel Fast and Automatic Condition Monitoring Strategy Based on Small Amount of Labeled Data. IEEE Journals & Magazine

    | IEEE Xplore. 2022;52(2):1266-1275. doi:10.1109/tsmc.2020.3018102

23. Attoui I, Boutasseta N, Fergani N. Novel machinery monitoring strategy based on TimeFrequency Domain similarity measurement with limited labeled data. IEEE Transactions on Instrumentation and Measurement. 2021;70:1-8. doi:10.1109/tim.2020.3011874

24. Arun P, Lincon SA, Prabhakaran N. An Automated method for the analysis of bearing vibration based on spectrogram pattern matching. Journal of Applied Research and Technology. 2019;17(2). doi:10.22201/icat.16656423.2019.17.2.805

25. A transfer learning Framework for anomaly Detection using a model of normality. IEEE Conference Publication | IEEE Xplore. Published November 4, 2020.

    https://ieeexplore.ieee.org/abstract/document/9284916

26. Aburakhia S. Similarity-Based Predictive Maintenance Framework for rotating machinery. arXiv.org. Published December 30, 2022.

    https://arxiv.org/abs/2212.14550

27. Li J, Schaefer D, Milisavljevic-Syed J. A Decision-Based Framework for predictive maintenance technique selection in Industry 4.0.

    Procedia CIRP. 2022;107:77-82. doi:10.1016/j.procir.2022.04.013

28. Jiménez JJM, Schwartz S, Vingerhoeds RA, Grabot B, Salaün M. Towards multi-model approaches to predictive maintenance: A systematic literature survey on diagnostics and prognostics. Journal of Manufacturing Systems. 2020;56:539-557. doi:10.1016/j.jmsy.2020.07.008

29. Zhang L, Li X, Yu J. A review of fault prognostics in condition based maintenance. Proceedings of SPIE – the International Society for Optical Engineering. Published online November 13, 2006.

    doi:10.1117/12.717514

30. Wöstmann R, Barthelmey A, West N, Deuse J. A retrofit approach for predictive maintenance. In: Springer eBooks. ; 2019:94-106. doi:10.1007/978-3-662-59317-2_10

31. Van Cuong S, Shcherbakov M, Tran VP. Data-Driven Framework for Predictive Maintenance in Industry 4.0 concept. In: Communications in Computer and Information Science. ; 2019:344-358. doi:10.1007/978-3-030-29743-5_28

32. Bezbaruah PGR& SJD& SSN& PB& I. Implementation of Predictive Maintenance Systems in Remotely Located Process Plants under Industry 4.0 Scenario. ideas.repec.org. Published online 2020.

    https://ideas.repec.org/h/spr/ssrchp/978-3-030-36518-9_12.html

33. Shamayleh A, Awad M, Farhat J. IoT based predictive maintenance management of medical equipment. Journal of Medical Systems. 2020;44(4). doi:10.1007/s10916-020-1534-8

34. Ran Y. A Survey of Predictive Maintenance: Systems, Purposes and Approaches. arXiv.org. Published December 12, 2019.

    https://arxiv.org/abs/1912.07383

35. Schmidt, Wang. Predictive Maintenance: Literature Review and Future Trends. University of Wolverhampton, UK: The Choir Press. 2015;1:232-239.

36. Alaswad S, Xiang Y. A review on condition-based maintenance optimization models for stochastically deteriorating systems. Reliability Engineering & System Safety. 2017;157:54-63. doi:10.1016/j.ress.2016.08.009

37. Zolotová I, Papcun P, Kajáti E, Mikuf M, Mocnej J. Smart and cognitive solutions for Operator 4.0: Laboratory H-CPPS case studies. Computers & Industrial Engineering. 2020;139:105471. doi:10.1016/j.cie.2018.10.032

38. Alqahtani AY, Gupta SM, Nakashima K. Warranty and maintenance analysis of sensor embedded products using internet of things in industry 4.0. International Journal of Production Economics. 2019;208:483-499. doi:10.1016/j.ijpe.2018.12.022

39. Olaf JM, Hanser E. Manufacturing in times of digital business and industry 4.0 – the industrial internet of things not only changes the world of manufacturing. In: Lecture Notes in Mechanical Engineering.

    ; 2019:11-17. doi:10.1007/978-3-319-99353-9_2

40. Karkalos NE, Markopoulos AP, Davim JP. General aspects of the application of computational methods in Industry 4.0. In: SpringerBriefs in Applied Sciences and Technology. ; 2018:1-10. doi:10.1007/978-3-319-92393-2_1

41. Dinardo G, Fabbiano L, Vacca G. A smart and intuitive machine condition monitoring in the Industry 4.0 scenario. Measurement. 2018;126:1-12. doi:10.1016/j.measurement.2018.05.041

42. Roy R, Stark R, Tracht K, Takata S, Mori M. Continuous maintenance and the future Foundations and technological challenges. CIRP Annals. 2016;65(2):667-688. doi:10.1016/j.cirp.2016.06.006

43. Ooijevaar TH, Pichler K, Di Y, Hesch C. A Comparison of Vibration based Bearing Fault Diagnostic Methods. International Journal of Prognostics and Health Management. 2019;10(2). doi:10.36001/ijphm. 2019.v10i2.2726

44. Peres RS, Rocha AD, Leitão P, Barata J. IDARTS Towards intelligent data analysis and real-time supervision for industry 4.0. Computers in Industry. 2018;101:138-146. doi:10.1016/j.compind.2018.07.004

45. Qiao G, Weiss BA. Industrial robot accuracy degradation monitoring and quick health assessment. Journal of Manufacturing Science and Engineering-transactions of the Asme. 2019;141(7). doi:10.1115/1.4043649

46. Xia T, Xi L. Manufacturing paradigm-oriented PHM methodologies for cyber-physical systems. Journal of Intelligent Manufacturing. 2017;30(4):1659-1672. doi:10.1007/s10845-017-1342-2

47. Yan J, Meng Y, Lu L, Lin L. Industrial big data in an industry 4.0 environment: challenges, schemes, and applications for predictive maintenance. IEEE Access. 2017;5:23484-23491. doi:10.1109/access.2017.2765544

48. Kumar A, Chinnam RB, Tseng F. An HMM and polynomial regression based approach for remaining useful life and health state estimation of cutting tools. Coputers & Industrial Engineering. 2019;128:1008- 1014. doi:10.1016/j.cie.2018.05.017

49. Ansari F, Khobreh M, Seidenberg U, Sihn W. A problem-solving ontology for human-centered cyber physical production systems. CIRP Journal of Manufacturing Science and Technology. 2018;22:91-106. doi:10.1016/j.cirpj.2018.06.002

50. Sénéchal O. Performance indicators nomenclatures for decision making in sustainable conditions based maintenance. IFAC-PapersOnLine. 2018;51(11):1137-1142. doi:10.1016/j.ifacol.2018.08.438

51. imon J, Trojanová M, Zbihlej J, Sárosi J. Mass customization model in food industry using industry 4.0 standard with fuzzy-based multi- criteria decision making methodology. Advances in Mechanical Engineering. 2018;10(3):168781401876677. doi:10.1177/1687814018766776

52. Kumar A, Chinnam RB, Tseng F. An HMM and polynomial regression based approach for remaining useful life and health state estimation of cutting tools. Computers & Industrial Engineering. 2019;128:1008- 1014. doi:10.1016/j.cie.2018.05.017

53. Subramaniyan M, Skoogh A, Salomonsson H. A data-driven algorithm to predict throughput bottlenecks in a production system based on active periods of the machines. Computers & Industrial Engineering. 2018;125:533-544. doi:10.1016/j.cie.2018.04.024

54. Mourtzis D, Vlachou E. A cloud-based cyber-physical system for adaptive shop-floor scheduling and condition-based maintenance. Journal of Manufacturing Systems. 2018;47:179-198. doi:10.1016/j.jmsy.2018.05.008

55. Frieß U, Kolouch M, Friedrich A, Zander A. Fuzzy-clustering of machine states for condition monitoring. CIRP Journal of Manufacturing Science and Technology. 2018;23:64-77. doi:10.1016/j.cirpj.2018.09.001

56. Sahal R, Breslin JG, Ali MI. Big data and stream processing platforms for Industry 4.0 requirements mapping for a predictive maintenance use case. Journal of Manufacturing Systems. 2020;54:138-151. doi:10.1016/j.jmsy.2019.11.004

57. Goodall PA, Sharpe R, West AA. A data-driven simulation to support remanufacturing operations. Computers in Industry. 2019;105:48-60. doi:10.1016/j.compind.2018.11.001

58. Kumar SM, Purohit BS, Manjrekar V, Singh V, Lad BK. Investigating the value of integrated operations planning: A case-based approach from the automotive industry. International Journal of Production Research. 2018;56(22):6971-6992.

    doi:10.1080/00207543.2018.1424367

59. Rødseth H, Schjølberg P, Marhaug A. Deep digital maintenance. Advances in Manufacturing. 2017;5(4):299-310. doi:10.1007/s40436- 017-0202-9

60. Masoni R, Ferrise F, Bordegoni M, Gattullo M, Uva AE, Fiorentino M, et al. Supporting Remote Maintenance in Industry 4.0 through Augmented Reality. Procedia Manufacturing. 2017;11:1296-1302. doi:10.1016/j.promfg.2017.07.257

61. Gattullo M, Scurati GW, Fiorentino M, Uva AE, Ferrise F, Bordegoni

    M. Towards augmented reality manuals for industry 4.0: A methodology. Robotics and Computer-Integrated Manufacturing. 2019;56:276-286. doi:10.1016/j.rcim.2018.10.001

62. Scurati GW, Gattullo M, Fiorentino M, Ferrise F, Bordegoni M, Uva AE. Converting maintenance actions into standard symbols for Augmented Reality applications in Industry 4.0. Computers in Industry. 2018;98:68-79. doi:10.1016/j.compind.2018.02.001

63. Aschenbrenner D, Maltry N, Kimmel J, Albert M, Scharnagl J, Schilling

    K. ARTab – using Virtual and Augmented Reality Methods for an

    improved Situation Awareness for Telemaintenance**funded by the Bavarian Ministry of Economic Affairs, Infrastructure, Transport and Technology in its R&D program Bayern digital. IFAC- PapersOnLine. 2016;49(30):204-209.

    doi:10.1016/j.ifacol.2016.11.168

64. Koch PJ, Van Den Broek MK, Dbska P, Thormann MA, Tetzlaff AJ, Bøgh S, et al. A skill-based robot co-worker for industrial maintenance tasks. Procedia Manufacturing. 2017;11:83-90. doi:10.1016/j.promfg.2017.07.141

65. Chong S, Pan G, Chin JK, Show PL, Yang TC k., Huang CM. Integration of 3D Printing and Industry 4.0 into Engineering Teaching.

    Sustainability. 2018;10(11):3960. doi:10.3390/su10113960

66. Mueller E, Chen X, Riedel R. Challenges and Requirements for the Application of Industry 4.0: A Special Insight with the Usage of Cyber- Physical System. Chinese Journal of Mechanical Engineering. 2017;30(5):1050-1057. doi:10.1007/s10033-017-0164-7

67. Fusko M, Rakyta M, Furmannová B, Dulina , Gao M, Grznár P. BASICS OF DESIGNING MAINTENANCE PROCESSES IN INDUSTRY 4.0. MM Science Journal. 2018;2018(01):2252-2259.

    doi:10.17973/mmsj.2018_03_2017104

68. Hesser DF, Markert B. Tool wear monitoring of a retrofitted CNC milling machine using artificial neural networks. Manufacturing Letters. 2019;19:1-4. doi:10.1016/j.mfglet.2018.11.001

69. Alonso A, GarcíaPozo A, Cantera JM, De La Vega F, Hierro JJ. Industrial Data Space Architecture Implementation using FIWARE. Sensors. 2018;18(7):2226. doi:10.3390/s1807222