Special issue “Towards Digitalized Manufacturing 4.0”

In light of Industry 4.0, Manufacturing has become internationally interlinked. The ongoing transformation of the manufacturing landscape offers fertile ground for better and more efficient manners, with which engineers are capable of managing all the different processes within a plethora of industries. The increase of automation, as a result of the immense digitization and digitalization, in production processes and smart systems of machines, has enabled the efficient communication with each other (Machine to Machine – M2M) and with operators (Human to Machine – H2M) through the Internet of Things (IoT). Consequently, it becomes apparent that manufacturing and production companies should digitalize all their horizontal and vertical integrations, in order to secure their sustainability in the industrial digital ecosystems of the future. However, new business models and new working spaces in manufacturing require a greater need for technical knowledge, digital and soft skills development. By extension, scientific research needs to enhance sharing and advance specific knowledge in smart manufacturing systems/networks, flexible processes and automated machine tools that characterize Industry 4.0, while taking advantage of Cyber-Physical Systems, Internet of Things, Cloud Manufacturing, and Big Data Analytics.

The scientific community in the field of manufacturing is expected to generate world class breakthroughs and help in turning digital innovation into new manufacturing knowledge. The manufacturing education paradigm has to change rapidly in order to provide industry with a new generation of engineers owning a digital culture. The CIRP CMS 2021 covered a wide spectrum of research works with more than three hundred publications. The key research topics covered in the Conference and by extension this Special Issue include among others the following:

• Industry 4.0 & Cyber-Physical Systems

• Machine 4.0 & Operator 4.0

• Internet of Things and Smart Sensing

• Cloud Manufacturing

• Social Media in Manufacturing

• Advanced Human-Robot Collaboration

• Smart systems via Big Data Analytics

• Product Service System Engineering

• Adaptive Self-learning Systems

• Lean Manufacturing and Design

• Virtual, Digital, Smart Factories

• Customisation and Personalisation

• Value-adding Networks and Logistics

• Rapid Manufacturing

• Predictive Maintenance & Lifecycle Management

The first paper of this Special Issue is focused on hybrid teaching of the new generation of engineers (Mourtzis, Panopoulos, and Angelopoulos 2022). Concretely, hybrid classrooms have become increasingly prevalent amid the global pandemic. Cutting-edge digital technologies are playing a key role in shaping the new teaching paradigms. Digital Twins technology, a notable facet of Industry 4.0, enables faculty to construct simulation models aligned with course requirements. Universities grapple with the challenge of providing high-quality learning experiences while prioritizing participant health. In response, this study introduces a Hybrid Model within the Teaching Factory framework, successfully implemented, and validated. Through four hybrid case studies, Engineering students remotely guided laboratory personnel, achieving successful manufacture and assembly of customized projects. The study also incorporated training webinars in Maintenance and Computer-Aided Manufacturing (CAM). The findings of this research work have unveiled new pedagogical dimensions, introducing sustained support and cyber safety as integral constructs beyond technology, and learning environments. Among the key elements of the proposed method is collaboration, which is emphasized at various levels, with a spotlight on cyber safety for ensuring a secure digital learning environment.

The second paper is entitled ‘A human-oriented design process for collaborative robotics’, and it focuses on efficiently realizing the potential of collaborative robotics in human-robot interaction (Papetti et al. 2022). The latter poses a great challenge, particularly in the design of integrated production systems for mass customization within the framework of Industry 4.0. Therefore, in this study a set of methodologies are proposed aimed at facilitating engineers in the human-centric design process, ensuring satisfactory performance while prioritizing worker safety and health. The design criteria are formed in five key pillars, namely i) safety, ii) ergonomics, iii) effectiveness, iv) flexibility, and v) cost. Following the implementation of the proposed methods, a series of alternative design solutions are extracted in order to ensure informed decision-making. In order to emphasize on the dynamic impact of design choices on system elements, the method was implemented in a leading Italian kitchen manufacturer. More specifically, the integration of a collaborative robotic cell in one of the manufacturer’s assembly lines has resulted in a more balanced production line (10% improvement), reduced risk factors (RULA score down from 5 to 3 and OCRA score from 13.30 to 5.70). Further to that, enhanced allocation of operators to high-value activities has also been achieved.

Moving on, the concept of intelligent workplaces which are capable of adapting to human behavior has emerged in the context of cognitive production systems. In that context the authors of the third manuscript which is entitled ‘Explainable human activity recognition based on probabilistic spatial partitions for symbiotic workplaces’ recognized and analyzed crucial details of human actions through continuous probability density estimates in various workplace configurations (Belay Tuli and Manns 2023). Concretely, efforts are being directed toward methods for the identification of human emotions and actions, aiming to enhance the cognitive abilities of smart machines, particularly robots in collaborative workspaces. The recognition of human activities and the anticipation of subsequent operations can simplify robot programming and improve overall collaboration efficiency. However, there is a need for explainable models in human activity recognition that are adaptable, robust, and interpretable. In this research work a novel approach called Human Activity Recognition based on Probabilistic Partition (HAROPP) is proposed. The authors have compared its performance with a geometric-bounded activity recognition method. Moreover, in an attempt to test the applicability of the proposed method, three scenarios, including standalone, one-piece flow U-form, and human-robot hybrid workplaces, are explored. The results indicate that spatial partitions relying on probabilistic density exhibit a 20% reduction in data frames and a 10% increase in spatial areas compared to geometric bounding boxes. On average, HAROPP accurately detects human activities in 81% of cases for a predefined workplace layout. HAROPP shows promise in scalability and applicability for cognitive workplaces with a digital twin, contributing to advancing the cognitive capabilities of machine systems and creating human-centric environments.

The fourth paper is entitled ‘Evaluation of a production system’s technical availability and maintenance cost – development of requirements and literature review’ (Sielaff, Lucke, and Sauer 2023). Essentially this research contribution elaborates on the theoretical aspect of industrial maintenance and equipment availability in the context of Industry 4.0, which is characterized by increasingly intricate production environments. Industrial maintenance can be realized as a determining factor for the efficiency of manufacturing systems and networks, having a direct impact on costs and technical availability of machinery and components. Maintenance managers face the challenge of maintaining necessary production system availability while minimizing resource utilization. Therefore, in an attempt to address this challenge, new methods for evaluating the interdependence between the technical availability of production systems and the associated maintenance resources, are required. In that sense, through the literature review investigation in this research work, such methods are examined and evaluated. To evaluate the methods, specific requirements are established, emphasizing the inclusion of maintenance strategies. Maintenance strategies are crucial as they underpin both component availability and the required maintenance resources. Thirteen requirements are identified, and 21 distinct methods are scrutinized. Notably, only one of the proposed methods fulfills all requirements, while others fall short in accommodating various combinations of maintenance strategies and their impacts on the production system.

Next, the fifth paper is entitled ‘A data-driven digital twin framework for key performance indicators in CNC machining processes’ (Vishnu, Varghese, and Gurumoorthy 2023). In this research work, the authors propose a data-centric digital twin (DT) framework designed to forecast Key Performance Indicators (KPIs) within a CNC machining environment. The anticipated KPIs serve as valuable insights for decision-makers involved in the CNC machining process, aiding in the optimal selection of cutting parameters to achieve specific KPI goals. The primary beneficiaries of this framework are the process planner during the planning phase and the machine operator during machining. Given that cutting parameters significantly impact crucial KPIs like machining time, quality, and energy consumption, their accurate selection can enhance overall performance in CNC machining operations. The study focuses on energy and surface roughness as the chosen KPIs for predictive modeling within the proposed DT, utilizing experimental data obtained from CNC milling processes. The paper further elucidates the selection of predictive modeling methods at different stages of CNC machining and outlines their outcomes.

The demand for reconfiguring production systems is on the rise due to shorter innovation and product life cycles, coupled with economic volatility. Concurrently, the industrial automation landscape is witnessing the emergence of cyber-physical production systems, offering notable potentials like self-organization capabilities (Müller et al. 2022). Therefore, the authors of the sixth paper which is entitled ‘Architecture and knowledge modelling for self-organized reconfiguration management of cyber-physical production systems’ focus on efficient knowledge modelling in order to develop an effective reconfiguration management methodology. The authors initialize with a presentation and discussion of reference architectures, architectural patterns, basic principles, and knowledge modeling and management approaches. Furthermore, an evaluation of the above-mentioned is performed with a primary focus on reconfiguration management and emphasizing UML/XML-based and ontology-based approaches. As a result of the investigation and the evaluation, the authors proceed with the proposal of a novel approach featuring a multi-agent system and the MAPE-K concept for reconfiguration management, complemented by a service-oriented architecture for deterministic plant control within a layered framework. In order to model knowledge, a UML (Unified Modelling Language) information model has been utilized, seamlessly integrated into the system through XML files. The accompanying tool support enables users to describe system components with ease and adhere to the information model’s schema and restrictions through an intuitive GUI.

The seventh paper is entitled ‘Physics-based modelling of robot’s gearbox including non-linear phenomena’ (Aivaliotis, Kaliakatsos-Georgopoulos, and Makris 2023). In the context of this research work, advanced simulation techniques, i.e. Digital Twin, are the main area of focus. Under the light of Industry 4.0 the above-mentioned technology has been extensively applied in control, safety, and maintenance, in view of enhanced added value with increased accuracy. This study focuses on refining the accuracy of digital twins by addressing the critical parameter of friction torque in robot gearboxes, crucial for simulating robot dynamics. The framework outlined in this paper encompasses a comprehensive approach which is based on i) a literature review in the field of friction modeling approaches, ii) identification of LuGre model parameters through a metaheuristic Genetic Algorithm, and iii) validation through a 6-axis industrial robot. The proposed methodology, executed in OpenModelica, demonstrates efficiency in capturing the non-linear behavior of the gearbox using the LuGre model. The optimization problem for parameter estimation is solved in Matlab, utilizing data from the robot controller. The validation of the proposed methodology extends to the comparison of the torque responses of the digital model versus the real-world counterpart for a specific position signal.

The issue closes with the eighth manuscript entitled ‘Predictive maintenance: assessment of potentials for residential heating systems’ (Aurora and Rabe 2023). Predictive Maintenance (PdM) solutions are imperative for the prediction and minimization of equipment downtimes. However, the high implementation costs associated with PdM solutions, both for manufacturing systems and field-operating systems raise questions about the conditions under which the benefits justify the investment. This publication serves as an extended and more elaborated version of a previous scientific publication, in which the authors proposed a methodology for systematically assessing the potentials and benefits of PdM solutions in manufacturing, applying this methodology to Residential Heating (RH) systems necessitates adaptation due to the unique context of RH. Unlike manufacturing systems, RH systems are situated at customers’ sites, posing challenges related to customer and system feedback as well as data availability. In the context of this research work aspects of optimization, transfer, and adaptation of the methodology to RH systems, is carried out, providing baseline information on maintenance in RH. Furthermore, a real-life case study in the RH field, supporting BOSCH Thermotechnology’s reliability department has been utilized in order to validate the proposed methodology and most importantly to quantify the potentials of PdM solutions.

The conference organizing committee would like to express their gratitude to the scientific committee who supported the conference and this special issue with the provision of their valuable feedback and recommendations for improving the manuscripts published in this special issue. Further to that, the conference organizers would also like to extend their gratitude to the contributors who supported the publication of this special issue. With regards to the IJCIM journal, the organizing committee would like to acknowledge the support of Professor Stephen Newman who serves as the journal’s Editor-In-Chief.

Disclosure statement

No potential conflict of interest was reported by the author(s).