Introductory Chapter: Trends of Maintenance in the Industry 4.0 Era

Abstract

No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter. No abstract is required for Intro chapter.

1. Introduction

In order to increase the efficiency of production and service activities, companies need to pay increasing attention to ensuring the functionality of logistics and technological resources of systems and processes, which can be effectively supported by a well-functioning maintenance solution. In the field of planning and management of maintenance activities, a number of typical research directions can be identified that contribute significantly to improving the efficiency and reliability of production and service activities. The literature discusses a wide range of maintenance policies, such as predictive maintenance, corrective maintenance, preventive maintenance, condition-based maintenance, emergency maintenance, risk-based maintenance, planned maintenance, prescriptive maintenance, failure finding maintenance, predetermined maintenance, adaptive maintenance, total productive maintenance, or proactive maintenance. The various types of maintenance strategies have a great impact not only on the core processes of the value chain including the production processes but also significantly influence the efficiency of connected purchasing, distribution, and inverse or reverse processes (see Figure 1).

Figure 1.

Maintenance in the value chain.

This book provides the reader with an overview of the latest trends in maintenance activities to improve the efficiency of production and service systems. As the chapters illustrate, maintenance activities can improve the efficiency of production activities in a way that not only increases the level of customer service but also the sustainability of systems. Appropriate maintenance strategies are also becoming increasingly important in supply chains, as the proper performance of maintenance activities in road, rail, water, and air transport can make a major contribution to improving the reliability of global supply chains. The chapters in this book will help you understand these potential benefits.

2. Industry 4.0 technologies in maintenance

Jardine et al. [1] show in their research focusing on machinery diagnostics and prognostics implementing condition-based maintenance that the operation of condition-based maintenance is generally based on data acquisition, data processing, and maintenance decision-making, where the data acquisition and data processing are significantly influenced by diagnostics and forecasting of resources’ status. This research also highlights the importance of up-to-date sensor systems, multi-sensor solutions, which can lead to big data problems to be solved. Maintenance policy can also significantly influence the availability, flexibility, cost efficiency, and performance of maintenance, as Bányai shows in her researches focusing on maintenance strategy optimization from energy and cost efficiency point of view [2, 3]. Wang also discusses in a survey of maintenance policies of deteriorating systems [4] that there is a wide range of maintenance strategies, which can be applied depending on the objective of the production and service process focusing on age replacement, periodicity, failure limit, repair time, or prevention aspects. Imperfection is a very important characteristics of maintenance. Pham and Wang [5] analyze 40 different mathematical models of maintenance policies focusing on the imperfection, and they summarize that reliability measures and the state of the systems are significantly influenced by the chosen maintenance strategy. Therefore, suitable, appropriate optimization models have to be used to find the optimal maintenance strategy and maintenance policy. Ahmad and Kamaruddin also discuss and compare different maintenance strategies in an overview of time-based and condition-based maintenance in industrial applications [6]. They conclude that time-based and condition-based maintenance have their own characteristics, and the integration of IT solutions (data determination, data acquisition, data collection, data transfer, data analysis, and decision-making) depends on the chosen strategy. However, time-based maintenance policy is widely known, but they show the importance of condition-based maintenance, which needs state-of-the-art IT solutions, including Industry 4.0 technologies.

Machine learning and the application of artificial intelligence shows significant potential in improving the performance of maintenance solutions. Carvalho et al. discusses in their research [7] the application of machine learning technologies in predictive maintenance and shows that Linear Regression, Gaussian Process regression, Bayesian Network, Support Vector Machine, Multi-Gene Genetic Programming, Artificial Neural Network, Convolution Neural Network, Long Short-Term Memory Network, Deep Learning, Hierarchical Clustering, Gradient Boost, and k-means are the most widely used machine learning methods in predictive maintenance. These machine learning methods can significantly improve the design and operation of maintenance solutions, especially in the case of robust, large-scale industrial applications. As a practical scenario shows, Analytic Hierarchy Process is a very appropriate methodology to choose the optimal maintenance strategy because maintenance strategy or maintenance policy selection can be described as a complex, multi-tier decision-making problem. As the research by Bevilacqua and Braglia show [8], Analytic Hierarchy Process can be used to arrange the characteristics and parameters of maintenance strategies and policies, and the comparison of different maintenance policies can be based on the pairwise comparison of judgment aspects. The production and service processes can be described either as deterministic or as stochastic systems. In the case of maintenance, because of the uncertainties of resources and the environment, the decision-making models are generally uncertain models, and the systems can be described as stochastically deteriorating systems, where inspection frequency and maintenance degree significantly influence the efficiency of the chosen maintenance policy, as shown by Alaswad and Xiang [9]. As Cho and Parlar describe in their research [10], the availability, efficiency, sustainability, and flexibility of maintenance are significantly influenced by a wide range of subprocesses and sub models to be optimized, including repair models, inventory models, replacement models, and inspection models. The performance of maintenance operations and the maintenance policy depends on both technological and logistics aspects, where logistics aspects include not only inventory management problems but also purchasing, distribution, and transportation problems.

As a research by Peng et al. shows [11], machine prognostics is a core problem of maintenance (especially in the case of condition-based maintenance). They suggest a novel systematic maintenance decision framework, which integrates a wide range of Industry 4.0 technologies, including feature selection based on Principal Component Analysis, Genetic Algorithm, and Support Vector Machine; data training; diagnostics, forecasting, and prognostics based on real-time sensor data; and maintenance schedule influenced by the precision of prognostics and maintenance cost function.

Augmented reality is also an important Industry 4.0 technology, which can be integrated into digital twin solutions and can play an important role in the education and training of maintenance and support maintenance operations by external experts. Palmarini et al. [12] summarize the application potentials of augmented reality in maintenance, and their research results indicate a high fragmentation of hardware, software, and solutions. This fragmentation has a great impact on the complexity of augmented reality solutions. Augmented reality solutions lead to the integration of emerging technologies and result in the novel tele maintenance paradigm. The prognostic and the forecasting play an important role in the organization of maintenance operations, therefore it is important to have up-to-date solutions to decrease of error value in forecasting and prognostic processes. As Lee et al. show in research focusing on intelligent prognostics tools and e-maintenance [13], the integration of different Industry 4.0 technologies and Internet of Things solutions leads to the transformation of conventional maintenance strategies and policies (e.g. fail and fix policy) into e-maintenance policies (e.g. predict and prevent). Swanson also discusses this topic and concludes that fire-fighting strategies will be replaced by proactive maintenance policies including preventive, predictive, and total productive maintenance [14]. Garga and Deshmukh concluded [15] that maintenance management can be represented as an integrated approach because it includes a wide range of optimization models and methods, various maintenance techniques, different scheduling algorithms and approaches, telecommunication, and IT solutions. As Muller et al. conclude [16], this integration has important aspects from a collaborative environment, pertinent knowledge, and intelligence point of view.

As the above-mentioned literature review shows, Industry 4.0 technologies play an important role in the development, improvement, and operation of maintenance, and they can significantly improve efficiency. As discussed in previous studies [17], Industry 4.0 technologies are increasingly bringing to the fore advanced maintenance strategies that enable real-time optimization of maintenance. This is based on the idea that a logical layer, which is a digital twin of the real-world processes, can be created from the physical layer represented by the resources to be maintained by means of a suitable sensor network. The sensor network naturally generates such a large amount of data from the real system that it requires the use of advanced big data technologies to process it. From the digital twin, a real-time simulation model can be generated, in which real-time simulations can be performed using a suitable simulation software with a model of the real system with current parameters, and the future state of the system and the resources that build the system can be predicted with high accuracy. Based on these predictions, real-time optimization can be done to define a maintenance strategy to perform the maintenance activities required by the predictions in a cost-effective manner.

The digital twin retrieves information from a centralized database, which is linked to the Enterprise Resource Planning (ERP) and Manufacturing Execution System (MES). The optimization results of the real-time simulation and forecasting are linked to both the digital twin, the real-world system, the ERP, and the MES (see Figure 2).

Figure 2.

Concept of real-time optimization of maintenance policy.

The Industry 4.0 maturity of companies also influences the success of digitization of conventional maintenance because the available hardware and software components, production, logistics, and business processes have a great impact on the integration of Internet of Things solutions and the transformation of conventional maintenance operations into e-maintenance and maintenance in a cyber-physical environment.

The optimization of maintenance-related design and operation problems can be characterized as NP-hard optimization problems. Veres et al. [18] conclude in their research work that the design and operation tasks of industrial environments are determined by a wide range and number of deterministic and stochastic parameters. This increased number of variables and parameters results in enormous computational time for the exact solution, and this problem highlights the importance of heuristic and metaheuristic algorithms, which are useful to solve NP-hard optimization problems, as shown in the case of different maintenance-related problems:

• differential evolution algorithm is used by Feng et al. [19] to optimize a selective maintenance problem with stochastic durations,

• an adaptive large neighborhood search approach was proposed by Liu et al. [20] for the optimization of maintenance routing and scheduling,

• Arzanlou and Sardroud apply Particle Swarm Optimization for budget allocation and scheduling of maintenance operations [21],

• Zhang et al. [22] propose an improved Ant Colony Optimization for the operational aircraft maintenance routing problem.

As the above-described heuristic and metaheuristic approaches show, maintenance optimization is a complex mathematical problem where new models and methods are available to improve the performance of design and operation tasks.

3. Conclusions

Manufacturing and service companies must make increasing efforts to meet the dynamically changing needs of their customers in a cost-effective, timely, and high-quality way. This requires a production and service infrastructure that operates reliably and continuously. An important prerequisite for this reliable operation is an effective maintenance strategy to ensure that production and service systems operate efficiently. In this chapter, the author presented the main Industry 4.0 technologies (focusing on digital twin, augmented reality, heuristic and metaheuristic optimization, and machine learning) that can increase the efficiency of maintenance activities and transform conventional maintenance into maintenance in a cyber-physical environment through the integration of digitalization and Internet of Things technologies.

Conflict of interest

The authors declare no conflict of interest.