The core concept of Software-defined Manufacturing involves the separation of software and hardware. Functionalities are not predefined but can be added later through software changes. This necessitates a suitable software and hardware architecture, which allows for reconfiguration of control software, i.e., the dynamic orchestration of real-time software. Traditional orchestration tools from the realm of non-real-time computing lack real-time support and do not provide means to configure real-time communication channels such as Time-Sensitive Networking. Thus, this work describes an orchestration tool based on Kubernetes, which is capable of orchestrating real-time containers including the necessary communication interfaces required for TSN.

Industrial control systems are hard real-time systems. The correctness of such a system is determined not only by the result, but also by the time frame in which the result is obtained. Logical controllers receive sensor inputs, onto which they perform logical operations and generate an output to affect the behavior of the system. The end-to-end response time in distributed real-time systems is composed of communication and processing time and is time-constrained. The execution time of a task depends on the node’s resources available for execution. Therefore, execution of compute-intensive control tasks with limited edge node resources is challenging. Since meeting deadlines is mandatory, additional resources must be provided. For resource-constrained edge nodes, horizontal scaling must occur at some point and additional nodes must be taken into account to satisfy the requirements. In this paper, we develop a concept for scaling real-time control tasks horizontally across multiple nodes. For this, we apply the approach of software pipelining to multiple nodes. To ensure deterministic communication, a real-time TSN network is used.

Industrial automation systems are evolving. Approaches such as software-defined manufacturing aim to overcome the barriers of rigid predefined structures in the manufacturing domain. When functionality of the production hardware is defined by software services, new types of resources have to be modeled to depict the automation system in its entirety. Since software is executed on computing resources, and these directly influence the behavior of the manufacturing system, an approach to consistently model these different types of resources is necessary. In this paper, we present a model to depict production and compute resources continuously and uniformly.

Manufacturing-X aims to build federated data ecosystems to increase industrial resilience, sustainability, and competitiveness. However, while this ambition is clear, the concrete pathway to achieve it remains fragmented. To address this gap, this paper systematically analyzes five diverse, advanced use cases within a physical federated testbed. Our novel contribution lies in identifying that, despite the great diversity of these use cases, the underlying technological barriers to overcome are remarkably similar. Building on this key finding, we propose a unified ”meet in the middle” architectural framework. This strategy reconciles bottom-up operational technology (OT) realities with top-down IT intent. Ultimately, this use-case-agnostic approach enables a step-by-step implementation towards a scalable ecosystem, offering greater real-time capability and automation potential than conventional approaches.

In response to the increasing complexity of production systems, automated methods for creating simulation models are being explored. Accurate simulation models are necessary for virtual commissioning to effectively model the real operation of a factory. However, manual model creation is labor-intensive and requires expertise. To address this challenge, this work investigates automated approaches by reviewing current methods and proposing the automated creation of simulation models from video data, especially for brownfield machines. Experimental results confirm the feasibility of using photogrammetry and Neural Radiance Field for automated mesh generation.

Industrial robots are relatively inexpensive, compared to conventional machining tools. However, they suffer in terms of positioning and tracking accuracy. To mitigate this issue, laser trackers can be used to calibrate robots or even for positioning control. The spherically mounted reflector used must be visible to the laser tracker. Therefore, the position of the reflector needs to be known precisely to ensure visibility. To evaluate and satisfy the visibility condition for laser tracker tasks, such as calibration or path tracking, this paper presents an efficient algorithm for identifying the kinematic chain, which includes the tracker.

Static configurations and slow adaption to changing requirements characterize today's production systems. In order to cope with variable external influences, these systems need to become more flexible. One basic building block to achieve more flexible automation systems from the software perspective are modular real-time architectures. These are realized as distributed real-time systems containing distributed real-time applications. The improved modularity and flexibility come at the price of greater complexity. In this paper we develop a scheduler for the requirements-based, container assignment of distributed realtime applications to computing nodes of a hardware infrastructure, which includes an automated check of relevant properties and configuration of deployment units. Furthermore, this scheduler is able to form a holistic framework for the orchestration of distributed real-time applications as a background process together with the orchestration tool Kubernetes. This makes it possible to control the entire life cycle of distributed real-time applications.

In today’s rapidly changing manufacturing environment, it is increasingly important to have dynamic process reconfiguration through adaptation control mechanisms. Digital Product Passports are also being developed and will become mandatory, such as for batteries in the European Union by 2027. Although this may initially require additional work, it can lead to synergies from the information obtained. By using CO_2 values, e.g., from a Digital Product Passport of components, you can optimize the CO_2 content of the end product to achieve a specific target and improve competitiveness. This can be achieved by adapting processes, such as choosing between high or low dynamics to influence overall energy consumption. A framework is necessary to extract specific information from DPP and make decisions for adaptation. This paper presents a framework architecture based on OPC UA and the AAS, accompanied by an illustrative example of a battery-packing handling process. The speed of the packing process is determined by the energy consumption values of the individual cells to reduce the total energy consumption value of the battery.

The dynamic accuracy of industrial robots is significantly influenced by the elastic drive trains of the axes. Their behavior is composed of the coupled dynamics of drive control and gear mechanics. As an undesirable consequence, increasing elasticity leads to growing tracking errors. One approach to reduce tracking errors is semi-active damping. The functional principle is based on damping the gear mechanics by selective braking of an additional actuator. From a drive control perspective, this results in a more favorable system behavior, which, in turn, allows the selection of more performant control parameter values. This leads to better tracking and disturbance behavior. The aim of this paper is to transfer the semi-active damping with a low-cost additional actuator to cascade-controlled industrial robots. For this purpose, a novel semi-active control law is proposed for actuator control. A damping actuator for the first robot axis is designed, design rules are derived, and an integration concept is proposed. Finally, a H infinity synthesis methodology for simultaneous parameterization of the drive and actuator control is introduced. An experimental validation proves the effectiveness of the solution at axis level resulting in an average 17.3 % reduction in tracking errors, and in a milling experiment, reducing the average Euclidean tracking error by 42.7 %.

Due to flexibility requirements, the strictly horizontal communication of the automation pyramid is converging. This includes communication at the field level. Due to the long lifetime of machines in manufacturing, it is important to support converged communication for existing fieldbuses such as EtherCAT. This enables the implementation of brownfield approaches for gradually implementing converged networks in existing plants. EtherCAT is a widely used industrial Ethernet-based fieldbus protocol for communication between programmable logic controllers and field devices. This work analyzes the tunneling concept of EtherCAT through a Time-Sensitive Networking (TSN) network from the literature and contributes an empirical evaluation based on a test setup with multiple EtherCAT networks and EtherCAT slaves. The tunneling of EtherCAT through TSN based on Virtual Local Area Networks (VLANs) is demonstrated to be a viable option, allowing the utilization of existing EtherCAT devices without the necessity for adaptation. A comparison of the Linux features SO_TXTIME and a simple RAW_SOCKET reveals that both introduce jitter, which is compensated by the EtherCAT slaves to a few microseconds.

Large industrial robots are inexpensive compared to their workspace. Due to flexibility, the positioning accuracy is compromised. This paper presents a simple yet effective static calibration method for robots using a laser tracker that can be applied to processes with no change in external load and without high processing forces. The calibration method is based on a novel elastokinematic model that captures the nonlinear gearbox and link compliance and also accounts for the effects of pose dependent joint friction and nonlinearities due to the gravity compensation mechanism. Furthermore, a feedforward concept is presented that separates kinematics from compliance to reduce computation time. The calibration method was implemented on industrial hardware and experimentally validated on a KUKA KR210-2. In comparison to state of the art compliance modeling, accuracy is improved by 23.1 %. Finally, the feedforward model reduces the mean positioning error in the entire workspace by 54.9 % compared to the purely kinematically calibrated model, emphasizing the need for compliance compensation. Note to Practitioners-Industrial processes such as bonding, welding, quality inspection, and aircraft assembly require high path and positioning accuracy. This is often accomplished by programming paths offline. Due to inaccuracies, these programs often require manual corrections, which is undesirable in an era when production aims for greater customization. Another possibility is to utilize additional, costly sensors. Especially for large industrial robots, the inaccuracy increases as the lever arms and component masses increase due to the additional load on the drivetrains. Therefore, robot manufacturers specify repeatability rather than absolute positioning accuracy, because these effects depend on the robot's pose. To overcome this issue, a practical compensation technique that minimizes the pose dependent error is presented. This is achieved by using a laser tracker to identify an improved model of the robot. The calibration procedure can then be automated without the need for any information other than that provided in the robot's manual. Finally, the compensation algorithm is numerically inexpensive, enabling it to be implemented in robot control when there is a possibility of manipulating the path planning angles. A short step by step explanation is given on how to implement our methods for other robots. With that, it is possible to expand the spectrum of robotic applications without external sensors, where no large process forces act on the axes.

The use of digital twins enables the trend toward a flexible and dynamic production system. As an aspect of digital twins, virtual commissioning models are used to test machines in the engineering phase, before they are even built. These models, however, are no longer used after the engineering phase. In this paper, a concept is presented, for utilizing these models during the production phase in order to optimize the production process. In addition, simple geometrical faults can be detected in the virtual commissioning model and optimized before the control program of the product is deployed on the real machine. (c) 2024 The Authors. Published by Elsevier B.V.

The migration of field-level automation systems to an edge-cloud hosted on premises can reduce the operation costs of a manufacturing line and increase its flexibility thanks to computation scalability and use of cloud-native operation concepts. Nonetheless, due to the novelty of the concept, a transparent life-cycle costs (LCC) assessment for edge-cloud automation systems (ECASs) is complex to achieve. In particular, uncertainties concerning costs related to the engineering and operation of an ECAS, such as commissioning delays or unexpected failures, increase the risks of investments towards its adoption. In this paper we propose a sensitivity analysis approach based on Monte Carlo (MC) life-cycle simulations. Through the approach, design parameters of an edge-cloud are investigated concerning their impact on the magnitude and distribution of LCC. The use of the approach is tested on a real-world case study to optimize the choice of the edge-cloud components and reduce the economic risks for its adoption. The proposed approach is expected to be integrated in early stages of the design process of a production line to identify cost-viable designs of ECASs, thus fostering the transition towards smart manufacturing systems. (c) 2024 The Authors. Published by Elsevier B.V.

In today's global market, manufacturers must be highly adaptable, producing profitably smaller batches. Disin-tegrating hardware and software and embracing modern IT concepts like containerization and Service-oriented Architecture enable this flexibility. While offering rapid reconfiguration, these approaches add complexity. Orchestration tools aid in analyzing response times, workload mapping, and parameter selection for containerized real-time applications. However, current tools struggle with managing distributed, heterogeneous, real-time apps. This paper proposes an orchestration solution based on Compositional Performance Analysis and Genetic Algorithm to optimize task allocation across diverse computing resources. The approach enables the orchestration of heterogeneously scheduled, container-based. distributed real-time systems.

Digital twins have a wide range of applications throughout the life cycle of computerized numerical controlled (CNC) machines. One particular manifestation of digital twins is the creation of models derived from virtual commissioning. These models actively accompany a physical machine throughout the engineering phase, but are typically discarded afterward. The objective of this study is to demonstrate the effectiveness of operation-parallel real-time simulation using virtual commissioning models for the early detection and elimination of defects in workpieces produced by CNC machines. The simulation model interfaces directly with the control drive, producing a virtual workpiece simultaneously with the actual manufacturing process in real-time, thus enabling necessary optimizations. Simultaneously, a virtual comparative model operates in the background to detect errors in the workpiece during runtime. This approach provides the advantage of eliminating the requirement for post-production quality inspections since the operational-parallel simulation conducts the inspection function. In addition, defects in the workpiece can be detected early so that the machine can be stopped in the event of a tolerance violation. At last, a framework is proposed to enable a knowledge-based model derived from identified defects. This framework can identify errors in early engineering stages, including the CAM planning phase, by using the knowledge from the production phase.

Cable-Driven Parallel Robots (CDPRs) can be used where a large workspace is required or where high payloads need to be handled. However, certain applications may avoid CDPRs due to drawbacks such as potential cable collisions. To improve the usability of CDPRs, the fixed frame anchors can be mounted on reconfiguration axes. This results in more complex kinematics that are challenging to control. This paper introduces a gradient descent approximation technique for reconfigurable CDPRs (RCDPRs) to identify, with low computational effort, appropriate configurations for a given platform pose to achieve a larger workspace. A workspace study shows a wrench-feasible workspace enlargement up to 28

Digital twins with a skill-based description are increasingly finding their way into industrial production. Their models link properties of a product with available process capabilities that are realized by resource skills. A matchmaking derives capable resources, e.g. for modular assembly systems. For capability-based planning in manufacturing, integration with computer-aided design (CAD) and computer-aided manufacturing (CAM) programs is missing. This paper presents an automated toolchain with focus on subtractive manufacturing using skill-based digital twins. Product-Process-Resource (PPR) models are used and enriched with subtractive manufacturing information and further relations. A configurator allows the design of the individual digital product twin, which can synchronously be generated as a 3D model using a CAD service. Necessary subtractive manufacturing steps including capable resources are then determined in a planning environment and fed into the CAM system. This allows an individual generation of machining code based on PPR information, which can be validated on a simulation model. The result is an automated tool chain of different software services, which are composed by a common description of the PPR model and the software services themselves. By defined models and interfaces the toolchain is not restricted to subtractive manufacturing but enables the integration of computer-aided everything (CAx) services. (c) 2024 The Authors. Published by Elsevier B.V.

Solving current challenges in the manufacturing industry, for example, ever-increasing productivity, growing un-certainty and flexibility, as well as the digitalization, requires a shift in paradigm of how industrial production is thought, engineered and implemented. Software-defined Manufacturing (SDM) replaces the production-system-centric approach by a production-solution-focused approach. Solutions are engineered virtually using software tools and defined by standardized model. The physical implementation of a software-defined solution is performed using services, that deploy the models on resources, representing available assets. The consistency between models and implementations of solutions enables continuous processes. SDM is an open ecosystem-based approach, combining key technologies, models and services. In this paper, a reference architecture for the SDM ecosystem is presented. Based on the definition of resources and solutions, the SDM-core is introduced connecting both. The core consists of data models and essential services, including asset management, orchestration and access management. The holistic architecture also includes human roles and external entities. The reference architecture is validated using representative use cases that were identified based on a large-scale use case study. The architecture is considered as a foundation for SDM and will be followed by a series of detailed publications regarding resources, technologies, use cases and applications.

Industrieroboter nehmen eine zentrale Rolle in der flexiblen Automatisierung von Produktionssystemen ein. Das Engineering dieser immer komplexer werdenden Systeme wird allerdings erschwert, da spezialisierte und oftmals herstellerabhängige Steuerungs- und Simulationslösungen verwendet werden. In diesem Beitrag wird aufgezeigt, welche Anforderungen offene Steuerungs- und Simulationsplattformen in Zukunft erfüllen müssen, um ein durchgängiges Engineering von der Anforderungsphase bis in den Betrieb zu ermöglichen. Anhand einer beispielhaften Realisierung wird aufgezeigt, wie eine solche Lösung schon heute mithilfe offener industrieller Komponenten umgesetzt werden kann.