Recent research in manufacturing increasingly focuses on energy efficiency, sustainable production, automation, and digitalization, driven by economic and environmental pressures. This progress is underpinned by digital technologies, advanced optimization methods, and smart manufacturing concepts, which collectively redefine the capabilities of modern production systems. Globalization and product individualization impose intensified competition, shortened lifecycles, and greater variability of workpieces. Conventional production systems cannot adequately respond to such dynamics. Within this context, autonomous machine tools constitute a key enabler, capable of independently planning, monitoring, and optimizing machining operations based on digital models and real-time data. This contribution reviews detailed solutions relevant for autonomous manufacturing and proposes a conceptual structuring of an autonomous manufacturing cell and its subsystems, outlines essential functional components, and discusses hierarchical control loops in the context of RAMI 4.0. Integrating industrial robots and mobile manipulators further enhances flexibility by automating material handling, thereby establishing machining cells with minimal human intervention.

Asset Administration Shells (AASs) provide standardized and semantically rich representations of industrial assets; however, their integration into machine-interpretable knowledge systems that support automated reasoning remains limited. This paper presents an architecture that derives Knowledge Graphs (KGs) from AAS models and applies multi-layer ontology-based reasoning for operational decision support within Digital Twins (DTs). A template-driven mapping transforms incoming AAS submodels into Resource Description Framework (RDF) triples, which are queried using SPARQL Protocol and RDF Query Language (SPARQL) against an ontology stack covering assets, capabilities, and decision rules. A Large Language Model (LLM) facilitates natural language interaction by generating ontology-constrained SPARQL queries and explanations, thereby improving accessibility for non-expert users. The architecture is demonstrated in a manufacturing use case, where it identifies spare tools after tool breakage in a manufacturing cell. Capability constraints, business rules, and availability data are combined to derive feasible alternatives, with the LLM assisting in interpreting candidate trade-offs. Although the case study employs simplified scoring parameters, the results demonstrate that the architecture reduces downtime and facilitates structured, explainable decision-making. The approach highlights the potential of integrating AAS-derived knowledge graphs with layered reasoning and controlled LLM assistance for operational DTs.

In volatile markets, resilient supply chains and shopfloors are essential to mitigate the impact of disruptions, such as crises, machine failures or quality issues, which result in significant costs. To address these challenges, information about products, processes and resources is required to design, test, and deploy resilience assessment and reconfiguration tools. Nowadays, this information is intended to be made available through data spaces and ecosystems, which necessitates preserving the data sovereignty of the respective companies involved. The architecture proposed by the Flex4res project accommodates these requirements. The implemented pre-pilot use case allows testing and eases the transition for companies.

In dynamic production environments, frequent changes in products and setups demand efficient machine vision systems to verify part confgura-tions accurately. This study introduces a digital twin-enabled toolchain designed to generate synthetic images for machine vision training, thereby reducing dependence on large, real-world image datasets. By employing pretrained deep learning models and an extractor-based classification method, the toolchain significantly minimizes the required number of training images without compromising accuracy. An industrial case study reveals the effectiveness of this approach, achieving reliable performance with fewer images and reducing training time, ofering a cost-effective solution for adaptable, resilient manufacturing systems.

Shop floor operations are affected by uncertainties and disruptions. These can come from the supply chain causing shortages of materials, worker unavailability due to, e.g., unforeseen pandemics, and spontaneous machine breakdowns. To address these challenges, resilient manufacturing systems are essential. However, there is a lack of practical methods for quantifying how well a shop floor withstands and recovers from such disruptions and assessing its level of resilience. Therefore, we propose a new approach for evaluating the resilience of the shop floor by assessing the production schedule and its underlying scheduler. Our Severity of Failure (SoF) method assesses each scheduled task by calculating the impact of its failure on the entire production schedule and combining it with the probability of the failure occurring. The method has been validated by assessing production schedules generated by various dispatching rules, demonstrating both its practical applicability and its potential to evaluate the resilience of different scheduling algorithms. Furthermore, the beneficial effect of buffer times on resilience was also evaluated.

Collaboration in manufacturing networks is crucial to maximize the impact of digitalization and leveraging collected data. Addressing three key challenges overcoming data silos, achieving semantic interoperability and protecting intellectual property is essential. A promising solution is the combination of Asset Administration Shells and data space technologies. This systematic literature review investigates implementations of these two concepts in manufacturing, evaluating use cases based on their technological readiness levels and the technologies employed. The findings aim to guide researchers and practitioners and identify future research directions.

Digital twins are representations of real-world systems in the digital world, relying on physical system data to optimise, manipulate, detect, and interact. The requirements of digital twins diverge depending on the application in focus. The complexity ranges from simple to highly complex use cases, which increases setup time and reduces maintainability. Therefore, different representations and views of the digital twin are necessary, e.g., file type conversions or abstractions of geometry. Standards that try to solve these problems are ISO 21597, by providing containerisation and linkage of data, and ITU-T Y.3090 and ISO 23247, by providing frameworks to support the creation process of digital twins. Based on these standards, this paper aims to enable a holistic view of digital twin applications in flexible manufacturing cells to reduce the implementation effort when the system is installed or changes. We identify a digital twin's most relevant features, parameters, and assets. This feature set is categorised regarding the frequency of changes. The main observable manufacturing elements are machine tools, robots, peripheral devices like autonomous mobile robots (AMRs), manufacturing utilities like tools, and processes. Various information systems depend on or interact with these manufacturing cells, e.g., CAM, PLM, MES, tool management, fleet control, and cell control. While applying the framework of ISO 23247 to a flexible manufacturing cell, we found that life cycle changes are only considered for products and not for the digital twin itself. Therefore we try to emphasise life cycle changes by linking the manufacturing cell to a life cycle meta-layer, simplifying the design, deployment, and updates of digital twin applications. This linkage remains valid through changes, reducing application maintenance effort by allowing re-instantiation of the applications. Therefore, a multidimensional representation of the physical layer can address rapidly changing real-time data to hardly changing life cycle information.

Reconfigurable manufacturing systems (RMS) are complex systems that regularly change the system behavior. One method to deal with this is virtual commissioning (VC). The new behavior needs to be implemented in the individual machines and the RMS-control and tested and verified afterward. Although various concepts for modeling behavior exist, simulations often lack interpreting and testing. This paper proposes a methodology for simulating and testing the functionality of RMS based on different behavior descriptions. A Proof of concept shows the first implementation of this methodology using a CAE platform combined with python.

With an increasing number of smart networked devices being used for data acquisition in production, reliable connectivity and fast communication are becoming increasingly important. Since mobile networks have been developed accordingly, 5G technology is also being used more frequently for machine-to-machine communication. The main advantages of 5G are very high data rates combined with low latency, which are essential for appropriate data acquisition in production. Although 5G is widely used in mobile phones, users are not familiar with 5G communication within the production domain. Also, companies that want to use 5G devices in their production environment often lack knowledge regarding implementation, connectivity issues, or best practices. The smart clamping pallet presented in this paper demonstrates the current possibilities of industrial communication with 5G. Various sensor systems such as GPS, acceleration sensors, and force measurement are installed in the pallet, enabling precise localization and in-process control of the production process. The fully automatic vice integrated into the pallet, which works independently from any wired connection, represents the basis for future highly flexible and collaborative manufacturing processes. Particularly concepts of distributed manufacturing can be implemented based on this. This setup enables a knowledge transfer for users and companies and hands-on training within the production domain. In the future, the pallet will be used in the TU Wien Pilot Factory Industry 4.0 (TU PF) to demonstrate and familiarise learners and other interested individuals with the possibilities of today’s industrial communication. In particular, networking and digitization can be understood based on this industrial use case for 5G. This use case meets the three pillars of the TU PF concept: (1) product piloting, (2) demonstration, and (3) knowledge transfer and education.

Additive Manufacturing (AM) has developed rapidly during the last decade. Various technologies such as Selective Laser Melting (SLM) or Direct Energy Deposition (DED) have demonstrated new opportunities in flexible product design and resource efficient process development. While many companies are entering the AM market, manufacturers still struggle to produce high-quality, certifiable parts with these new technologies. Modelling and simulation capabilities can support these companies by providing detailed information about part stiffness, stresses, temperatures, distortions, and other multi-physics-oriented problems regarding AM technologies. However, the application of these capabilities in industrial environments, especially small and medium enterprises (SMEs), has to be increased. To succeed in the industrialisation of new AM technologies, it is necessary to support companies in building-up knowledge and educating current and future design and manufacturing engineers in the area of simulation. The presented project aims to develop an innovative training program in the context of the TU Wien Pilot Factory Industry 4.0 (TU PF) learning factory in Vienna and the international EIT Manufacturing education community. The goal is to develop industry oriented learning paths and specifically designed learning nuggets, which support design and manufacturing engineers with little or no experience in AM to become proficient in the application of modelling and simulation tools for AM processes. The training concept is based on a hybrid learning approach, combining online self-learning content with practical experience using respective simulation software and a live-training with the AM hardware available at the TU PF. This paper presents an overview of the envisioned concept, the learning paths and nuggets, and the available software and hardware infrastructure.

Current manufacturing systems are faced with a rapidly changing environment caused by customization, which leads to smaller lot sizes. The answer to this challenging situation are agile production systems, which are able to react to new tasks very quickly. One major aspect for realizing such systems is a flexible control architecture for orchestration of machine tending tasks. The aim of this paper is to compare different approaches currently developed and applied in research and industry, e.g. centralized cloud control or head control by a master device. For evaluation, suitable requirements are determined with respect to current limitations and implementation effort.