Encyclopedia
Please note this is a comparison between Version 1 by Basheer Wasef Shaheen and Version 2 by Lindsay Dong.
Industry 4.0 is the latest technological age, in which recent technological developments are being integrated within industrial systems. Consequently, maintenance management of current industrial manufacturing systems is affected by the emergence of the technologies and features of Industry 4.0 which need to be integrated efficiently.
- maintenance management systems
- Industry 4.0
1. Introduction
The changes occurring now in every aspect of our daily lives result from technological advancement. This advancement has resulted in the emergence of new concepts, such as the Internet of Things (IoT), cloud computing, big data, artificial intelligence (AI), and cyber–physical systems (CPSs). These new technologies have paved the way for new innovative opportunities and more significant development in socioeconomic life [1]. This new stage of technological development is often referred to as “Industry 4.0” [2].
Industry 4.0 was first introduced by the German government to maintain mass production effectiveness and efficiency [3]. Industry 4.0 was proposed to respond to an increasing market demand around the globe, which implies more challenges, including intensified competition with leading industrial economies, such as the United States, and other developing economies, such as China and India. In addition to this, Industry 4.0 has come into being since many European countries are struggling to maintain their leading economic positions while facing various challenges, such as ageing communities, resource limitations, demographic and social changes, and the dynamic nature of world markets [4][5][4,5]. Industrial development has gone through four major advancement stages. The first industrial revolution was initiated in the 18th century and was characterised by mechanisation enabled by the power generated by steam and water. The second started at the beginning of the 20th century and was characterised by the utilisation of conveyor-based mass production supported by the emergence of electricity. The third revolution began in the 1960s and continued afterwards, witnessing the deployment of the programable logic controller (PLC) and the integration of computerised systems in machines to elevate automation and mass production [3][6][7][3,6,7]. The fourth development stage, namely, “Industry 4.0”, or the fourth industrial revolution, occurring now, is characterised by the utilisation of new technologies [8], which are resulting in new intelligent manufacturing systems [4]. Parallel with the development of industry, maintenance management has continuously developed to cope with the new requirements of industrial revolutions [8][9][8,9]. The origin of maintenance goes back to the first industrial revolution. During this period, maintenance meant using a machine until it failed and then repairing it. During the second industrial revolution, more complex machines were born, which needed more care and more complex maintenance activities. This age was characterised by frequency-based maintenance [10]. After the Second World War, in the 1950s, manufacturing sectors rapidly grew and were characterised by high competitiveness. Japanese engineers started to work on more proactive maintenance strategies to keep machines working and reduce downtimes; that is, a preventive maintenance strategy was applied. All technicians were encouraged to schedule general maintenance activities and report any other noted observations of machines. This method succeeded in reducing downtimes but resulted in high costs. Later, in the 1960s, more proactive maintenance strategies were developed, such as reliability-centred maintenance, risk-based maintenance, and total productive maintenance, due to the development of new manufacturing technologies. In the 21st century, the complexity of manufacturing systems is increasing, and maintenance has become a crucial and integrated part of the overall production system, which demands more concentration on the reduction of the associated costs, as well as increases in productivity, quality, and profits. More complex knowledge is needed to achieve these goals. Thus, the predictive maintenance strategy is being improved by better decision-support systems, which lead to fewer machine failures and fewer downtimes. Furthermore, another strategy that needs very complex knowledge is prescriptive maintenance, which is aligned with software support, AI, complex sensory systems for instant condition-based monitoring, IoT, big data, augmented reality (AR), and other integration-assisting tools in the total integration of maintenance management with Industry 4.0.
1.1. Maintenance Management
According to the European Committee of Standardization [11], maintenance management “includes all activities that determine the maintenance objectives, strategies, and responsibilities, and implementation by such means as maintenance planning, maintenance control, and the improvement of maintenance activities and economics”. Maintenance was considered a nightmare in past decades due to the application of corrective maintenance only [12]. In this process, maintenance mainly involved repairing and replacing things when needed, without planning, scheduling, or optimisation, combined with a lack of awareness of machine downtimes and behaviour. After that, maintenance activities became an independent function in most factories instead of being a production sub-function [13]. According to Duffuaa and Raouf [14], maintenance management systems consist of three major functions: planning, organisation and control. As depicted in Figure 1, planning activities include strategic maintenance alliances, in which the maintenance department should have its strategic maintenance plans comply with the strategic objectives of the company, such as outsourcing, organisation, and support.Figure 1.
Maintenance is not just about retaining and restoring a unit, but includes another important aspect, namely, optimising the overall cost. The goal is to find the general optimum for any process based on predetermined requirements and targets. For this purpose, maintenance strategies need to be addressed and selected. Accordingly, the main maintenance strategies in any modern manufacturing system include but are not limited to the following:
Corrective maintenance (CM): Also known as run-to-failure maintenance or breakdown maintenance. This concept is based on fixing things after a failure occurrence [13]. It can be carried out promptly after the failure occurrence or postponed for further repair or replacement actions.
Preventive maintenance (PM): Ensures that equipment does not break unexpectedly. To achieve this, maintenance must be performed regularly to maintain the stability of machines [15].
Risk-based maintenance (RBM): In which the available resources for maintenance tasks are prioritised toward the machines or other assets the failure of which poses the most significant threats to the whole system [16]. Based on the associated risk analysis of the potentially failed asset, maintenance plans and schedules are updated, then they are monitored through other maintenance strategies, such as condition-based maintenance [17][18][17,18].
Condition-based maintenance (CBM): A maintenance action and decision-making program which recommends suitable actions according to condition monitoring information by utilising prognostic methods for more reliable and cost-effective maintenance [19].
Reliability-centred maintenance (RCM): An approach that was originally developed for the aircraft industry and used to generate a cost-effective maintenance schedule by utilising estimated parameters for the reliability of the system [20]. For applications that have safety issues, the minimisation of costs and downtime as the main goals are usually achieved by eliminating the chance of failure occurrence by striking a balance between safety and availability, along with cost-effective maintenance.
Predictive maintenance (PdM): Represents an optimised trade-off between maintenance and performance costs. In addition, it measures efficiency and productivity and predicts remaining useful life before failure happens; it includes health condition monitoring and prognostics for future system behaviour and helps in decision-making processes [21][22][21,22]. It has been found that predictive maintenance is an effective strategy that can reduce the downtime of machines by 30–50% and extend their lifetime by 20–40% compared to traditional strategies [23][24][23,24].
Prescriptive maintenance: An advanced version of predictive maintenance supported by further decision-making mechanisms. Prescriptive maintenance goes one step further than PdM; it inspects not only the equipment to be maintained but also its environment and the correlation between them [25].
Opportunistic maintenance (OM): According to [26][27][26,27], OM is a systematic method of collecting, investigating, and preplanning activities for the generation of maintenance tasks to be implemented given opportunity [28]. A typical example is when a complex machine is disassembled to replace a broken component and it might be worth replacing other components that are close to end of life.
1.2. Industry 4.0
1.2.1. Industry 4.0 Technologies
The Internet of Things (IoT) represents the system in which elements in the material world, such as machines, equipment, and devices, communicate with each other and with cyber elements, such as software and data [1][29][1,34]. The main characteristic of IoT is the strongly decentralised and heterogeneous digital information exchange between devices connected in a network. IoT offers the possibility of providing an instant response to any request from surrounding objects or environments [30][35].
Cloud computing comprises online resources, such as servers, applications, and networks, to offer regular services that require more investment and resources to operate locally. Cloud computing is commonly used nowadays for its efficiency, cost-effectiveness, stability, and high-power availability, if needed [31][38]. Moreover, cloud computing is one of the main infrastructures for big data. Big data is related to the development of the Internet and connectivity, which have generated production-related data in large volumes, with high velocity, variety, and veracity. Such data require more sophisticated systems that can handle, analyse, and transform them into useful knowledge. Data obtained with IoT devices are analysed, only meaningful information is extracted, and knowledge is transferred efficiently to support business activities [1]. Simulation has become an important tool in Industry 4.0 contexts. It is a powerful computational tool for designing, analysing, and understanding the behaviour of complex systems; it plays a key role in the successful implementation of digital manufacturing [32][33][39,40]. Artificial intelligence (AI) is considered one of the primary keys to transforming manufacturing systems in the era of Industry 4.0. Through the integration of IoT, big data, and AI tools, manufacturing systems are able to make factual decisions by real monitoring and analysis of their processes through real communication modules to coordinate and monitor all activities between machines, people, sensors, and other parts of the manufacturing system [34][43]. Meanwhile, machine-learning (ML) techniques are used in real-life scenarios to predict the future behaviour of systems [35][44]. The most suitable algorithm can be chosen based on the given computational power, memory resources, and the number and quality of the data to be analysed [36][45]. Cyber–physical systems (CPSs) comprise environments in which the physical world, including machines, warehousing, and whole manufacturing systems, are transformed into the cyber world through network devices [3], where both the cyber and the physical parts interact. Exhaustive connectivity, supported by supercomputing power, has enabled such systems to gather and process real-time data and control production processes instantly.1.2.2. Industry 4.0 Features
Interconnection refers to the interconnection between different elements in a value chain. An example is interconnection between machines that handle similar work to coordinate the flow of products and avoid downtime or production delays. The product can inform the machine about the kind of operation to perform. Such interconnection elevates intelligent production. Similarly, products are smart and connected; once a product is produced in a machine, the next machine is well-informed jointly with the conveyor or the logistics robot responsible for transporting the product to the location of the next production process [37][53].
Interoperability is a crucial feature in Industry 4.0; it is the ability of two different systems to communicate with each other and make mutual use of their functionalities according to basic and common technological standards [6]. Therefore, machines, products, suppliers, and customers are integrated through a common language. Interoperability is vital for the effective operation of IoT, as every machine should have an interoperability standard that makes communication with other machines possible [29][34]. Integration is the process of combining all elements of the production system, including machines, products, and control systems, using sensors and actuators, and connecting these elements with other key players, such as customers, suppliers, logistics, transportation, maintenance, and production management [38][33].2. Maintenance Management in the Context of Industry 4.0
Various technologies are being used in manufacturing systems, but new concepts of integration can transform production and other related systems, such as maintenance. The full integration of such technologies can transform manufacturing cells and other supporting systems, such as maintenance engineering, to operate as fully integrated and automated systems with higher performance and greater efficiency.
2.1. Aligning Maintenance Management and Industry 4.0 Technologies—Trending Concepts and Integration-Assisting Tools
As explained earlier, Industry 4.0 is an umbrella term for several newly developed technologies, such as IoT, cloud computing, big data, simulation, AI, and CPSs. The adaptation of Industry 4.0 is vital at many managerial levels in manufacturing systems. To facilitate this, an implementation strategy is needed to digitalise manufacturing systems and their support systems, such as maintenance planning and scheduling. This can be achieved by the successful integration of new ICT technologies and big data capabilities through CPSs, which can enable significant improvements in maintenance throughout manufacturing systems. The transformation from current maintenance systems to digital maintenance complying with Industry 4.0 requirements needs recommendations and instructions to be generated, as in Fusko et al. [39][64]. Moving forward to digital or smart factories, the following Industry 4.0 factors present the main triggers of such transformation: real-time data collection through sensory or condition-monitoring systems, data-processing methods to ensure the accuracy and quality of collected data, and, finally, prediction models to prevent failures and update information. Smart and predictive maintenance are the main concepts used in smart factories [40][108], where many production and maintenance tasks need to be managed simultaneously, such as data collection and evaluation, resource availability, production, maintenance, and quality control. To increase the effectiveness of production processes in terms of maintenance, reduce the workforces required, and increase the effectiveness of management and planning processes, total productive maintenance (TPM) practices were digitalised by Tortorella et al. [41][65], who derived five case-based research propositions.2.1.1. Predictive Maintenance (PdM)
The predictive maintenance approach is the focal point of recent AI applications in the context of Industry 4.0. ML algorithms help detect and predict failures before their occurrence to avoid unplanned shutdowns and predict the remaining useful life (RUL) of equipment. Machine-learning methods also support the scheduling of maintenance activities through combined IoT technology to reduce downtime and maintenance costs and increase machine availability [42][43][44][45][63,109,114,117]. PdM can also be implemented using digital twins [46][115]. To connect the predictive maintenance concept with Industry 4.0 technologies, Li et al. [47][75] introduced a framework for predictive maintenance to analyse and predict faults in a machining centre. The framework included data acquisition, data pre-processing, fault diagnosis, and prognosis based on ANN, performance analysis, and maintenance schedule optimisation. Similarly, Tran Anh et al. [48][82] presented a PdM strategy applied in an automotive manufacturing company to cope with Industry 4.0 requirements, focusing on its impact on maintenance optimisation, in addition to the financial situation. ML algorithms and data-driven modelling are widely used for failure prediction [49][116]. A PdM model using the ML technique (Bayesian Filter) was developed by Ruiz-Sarmiento et al. [50][79] to predict the gradual degradation of machinery in a rolling process and then apply maintenance actions. Similarly, Paolanti et al. [51][80] used a random forest approach for PdM. The required data were collected from the sensory system and the machine’s PLC, while the communication protocols were set using Microsoft Azure. The results showed that the developed PdM predicted machine statuses efficiently and with high accuracy. Kiangala and Wang [52][83] suggested an experimental design for a PdM framework to detect the deterioration in conveyor motors in small manufacturing firms. To classify abnormalities into those that are production-threatening and those that are not, an ML classification model was built using time-series imaging and a convolutional neural network (CNN) to increase the accuracy of classification combined with parameterised rectifier linear units to improve the performance of the model. Moreover, principal component analysis (PCA) was applied to the multivariate time series to reduce the dimensions to two channels. The experimental results showed that this PdM framework was better in terms of performance and accuracy than the traditional approaches. Moreover, data-driven models can be used to develop architectures for PdM. Thus, Calabrese et al. [53][81] used such a model combined with machine-learning algorithms in an industrial woodworking machine to predict failure probabilities and the RUL of the machine. The proposed method was tested and showed high effectiveness in reducing machine downtime. Lastly, it was deployed in a big data framework to monitor multiple connected machines. It is worth mentioning that the prediction is a valuable source of information that can be used in maintenance management systems. PdM is considered the central concept of integration; in fact, different concepts connected with it were found in the literature. Ferreira et al. [54][66] utilised a CPS in sheet-metal-working machinery to achieve a full proactive maintenance system; three approaches were employed: (1) component failure detection using various detection means and models; (2) component failure prediction, depending on data collected from the sensory system, and then prediction of the RULs of the machine and its components, as well as potential failures before their occurrence; and (3) component failure diagnosis to identify the root causes of problems. These approaches utilise empirical models to help technicians diagnose problems. Pilot implementation of the framework was realised, and the result was a CPS integrated with the maintenance system (combined predictive and proactive maintenance). In terms of asset management, as a central part of a maintenance management system, Toeh et al. [55][89] presented an integrated predictive-maintenance-based machine-learning model in fog computing to manage assets (physical, virtual, and human resources), using a genetic algorithm (GA). Fog Workflowsim was used to simulate time and costs and to evaluate the performance of the GA, along with other methods, such as MinMin and MaxMin. Logistic regression as a supervised machine-learning algorithm was used to build the predictive maintenance model that reached 95.1% training and prediction accuracy. To support low-skilled technicians in performing maintenance tasks and procedures by utilising AR and computer vision (CV) techniques, Konstantinidis et al. [56][94] described a model called the Augmented Reality Maintenance Assistant (MARMA) that is able to generate maintenance instructions and support technicians. Additionally, integrating Industry 4.0 technologies and maintenance practices in the aeronautical industry is widely discussed to address how Industry 4.0 impacts maintenance. “PdM4.0” is a new concept defined by researchers using different approaches to implement PdM in Industry 4.0 environments. Sahba et al. [57][76] proposed a novel framework for PdM based on the advanced Reference Architecture Model Industry 4.0 (RAMI 4.0) that aims at reducing the maintenance and operation-associated costs in broadcasting chains. RAMI 4.0 is a model demonstrating features of technical assets and the concepts of Industry 4.0. The proposed framework, called PdM4.0, succeeded in increasing the stability of the broadcasting system and decreasing maintenance costs. Finally, in terms of reducing associated maintenance costs, various studies have focused on developing low-cost PdM models. Sezer et al. [58][71] developed an Industry 4.0 (CPS) architecture for low-cost predictive maintenance for small and medium manufacturing enterprises. In the developed CPS architecture, the temperature and vibration variables of a CNC machining process were measured. Accordingly, a regression tree model was used to predict the quality of the machined parts and then to reject or accept them based on the quality threshold and the correlations between temperature, vibration, and roughness.2.1.2. Maintenance 4.0
Recent studies have investigated “Maintenance 4.0”, which term is used to describe the latest trends in maintenance management to meet the integration requirements of Industry 4.0 and the sustainable development aspects [59][98]. It was found that Maintenance 4.0 involves the use of advanced analytic methods not only to predict failure (predictive maintenance) but also to avoid such failures and optimise maintenance schedules and resources (prescriptive maintenance). In other words, Kumar and Galar [60][99] stated that Maintenance 4.0 uses advanced technologies to perform predictive analytics and generate solutions through the integration of maintenance practices that deal with data collection, processing, analysis, visualisation, decision making, and Industry 4.0 technologies and features. On the other hand, Maintenance 4.0 enables the effective use of ICT platforms, such as ERP and CMMS, to manage the whole maintenance management system at all levels [61][118]. Smart and sustainable maintenance is the key element in maintenance 4.0, where the integration of digital technologies enables instant access to the real-time detailed information required and manages asset life cycles [62][119].2.2. Applications
The integration process of Industry 4.0 technologies and maintenance systems can be found in many application areas in the automotive, aerospace, and machining industries. In addition, chemical industries, including oil and gas, services such as broadcasting chains, and transportation and wind energy industries are also involved in such integration. The main applications of such integration are summarised in Table 12.
Table 12.
Classification results of the “Applications” category. Source: Authors’ elaboration.
| Article | Automotive Industry | Chemicals, Oil, and Gas | Machining | Aircraft Industry | Railway Transportation and Wind Energy | Services |
|---|---|---|---|---|---|---|
| [54][66] | X | |||||
| [63][67] | X | |||||
| [58][71] | X | |||||
| [64][73] | X | |||||
| [65][74] | X | |||||
| [57][76] | X | |||||
| [66][77] | X | |||||
| [50][79] | X | |||||
| [51][80] | X | |||||
| [48][82] | X | |||||
| [52][83] | X | |||||
| [67][85] | X | |||||
| [68][87] | X | |||||
| [69][88] | X | |||||
| [70][90] | X | |||||
| [71][92] | X | |||||
| [56][94] | X | |||||
| [72][95] | X | |||||
| [73][96] | X | |||||
| [74][97] | X | |||||
| [75][100] | X | |||||
| [44][114] | X | |||||
| [49][116] | X | X | ||||
| Total | 23 | |||||
3. Conclusions
Figure 25 summarises the connections between Industry 4.0 technologies and features and maintenance integration concepts found in the examined literature and the maintenance management system model adopted.
Figure 25. Summary of Integration of Industry 4.0 features and technologies and the adopted maintenance management system. Source: Authors’ elaboration.
Simulation, a key technology in Industry 4.0 [76][120], can provide optimised short- and long-term decision support for maintenance planning and scheduling tasks. Additionally, simulation can help in risk evaluation, cost reduction, and performance improvement and provide a roadmap for better integration. Using simulation, advanced sensory systems and IoT can immensely support the formulation of digital twins, with which more accurate predictions can be made in cyber–physical systems. Optimal maintenance scenarios, including maintenance strategies, schedules, and load forecasting, are suggested based on the analysis of data gathered by sensory systems and transferred from shopfloor level to big data systems. Such data could include machine health conditions and other information on the surrounding environment. Data are analysed using AI algorithms, and degradation processes, failures and the remaining useful life of equipment are predicted. Prediction results are then sent away for further planning and organisation of maintenance tasks, such as updating maintenance plans and schedules and allocating maintenance technicians and spare parts; credit goes to the IIoT for making it possible to improve quality and performance based on prediction results. Opportunistic maintenance implementation based on advanced prediction, simulation tools, and/or digital twins and multi-objective optimisation, along with IIoT, can also be an ideal approach for better integration, with highly mitigated integration risks and costs, and, on the other hand, highly increased performance, productivity, resource utilisation, and planning and scheduling.
Maintenance organisation requires high levels of coordination and resource allocation. Accordingly, maintenance organisation is aided by monitoring tools, such as supervisory control and data acquisition (SCADA) systems, ERP systems, and computerised maintenance management systems (CMMSs), which are considered some of the most important integration-assisting tools in Industry 4.0 due to their substantial influence on operations and maintenance processes and their associated costs [61][118]. Resources and spare parts allocation are managed effectively in the cloud by the ERP system.
Maintenance control is handled using computerised systems and knowledge dashboards. Maintenance effectiveness can be monitored and enhanced at different levels, starting at the machine level and ending with the overall plant. Many applications of Industry 4.0 technologies can enhance maintenance control through data modelling, optimisation, and behaviour-pattern detection. AI techniques, such as machine learning, can detect unseen factors affecting machine performance, product quality, and productivity.
Feedback is carried out through software support systems, passing through maintenance control and resulting in maintenance planning and rescheduling activities. Overall system performance is evaluated and governed based on the traditional evaluation metrics of maintenance management systems. Using an integrated sensory system and SCADA provides industrial organisations with a new method for information collection for monitoring and control of the performance of manufacturing systems. AI-based predictive maintenance modules process the data. Graphical user interfaces (GUIs) show the real-time measurements and information communicated between the CMMSs, ERPs, integration-assisting tools, and SCADA systems.
Interoperability enables seamless maintenance management by providing common means of communication between machines and operators, production departments, and different maintenance systems at the other locations of companies’ plants. Conditions and operational performance data are transmitted and pipelined in a standard format, and therefore provide the various maintenance management systems with software support, such as ERP, CMMS, and SCADA, with further compatibility. Such compatibility results in smooth report generation and data exchange.
Effective integration of Industry 4.0 features and technologies with maintenance management systems can improve whole systems through decentralisation, integration, interoperability, and interconnection. For instance, the Industry 4.0 feature “Interconnection” promotes real-time measurements and data flow along a value chain and between stakeholders. Interconnection enhances responsiveness and the coordination of all contributors in maintenance tasks. Moreover, equipment and machine manufacturers can monitor the performance of their machines under different operational conditions; therefore, they can enhance future generations of machines or improve the performance of current machines remotely. Smart machines can communicate with their original manufacturers periodically, receive software updates, and improve machine performance.
