Tumbling self-piercing riveting (T-SPR) extends conventional self-piercing riveting by superimposing a controlled tumbling motion of the punch on its linear movement. While T-SPR improves the versatility of conventional SPR, it also increases the complexity and enlarges the space of adjustable process parameters. This work presents a Bayesian optimization framework for a data-driven optimization of these parameters. Two sequential single-objective optimization campaigns are conducted on a steel-aluminum material stack to optimize mechanical and geometric joint quality. The approach identifies parameter sets that substantially improve joining performance, achieving large reductions in joining force together with increases in tensile shear strength and energy absorption. The results demonstrate that Bayesian optimization enables sample-efficient exploration and optimization of the T-SPR parameter space.
Product authentication technologies play a significant role in combating global counterfeiting, yet their environmental impacts remain underexplored. This study proposes a framework to assess the environmental impacts of three authentication technologies: Quick Response (QR) codes, Radio Frequency Identification (RFID), and image-based solutions. The assessment uses primary data for image-based solutions and secondary data for QR and RFID solutions, enabling quantification of environmental impacts from hardware production and technology use. In addition, a break-even analysis is conducted for annual production under low- and high-throughput scenarios. The results show that the image-based solution is associated with a higher initial environmental footprint due to hardware-related material and energy demands. However, this technology outperforms QR codes and RFID in the use stage, as it does not require physical information carriers. A qualitative functionality discussion provides further technology comparison: image-based systems excel in accuracy, QR codes in cost efficiency and scalability, and RFID in bulk processing speed. These findings provide support for manufacturers seeking to balance sustainability and scalability. By incorporating hardware production and use stages in the assessment, this study addresses a critical gap in the assessment of authentication technologies and offers actionable insights for their sustainable implementation.
Global competitive trends, technological disruptions, sustainability regulations, and an increasingly complex and dynamic environment are threatening the future viability of production systems. These factors are increasing the need for adaptability, technological advancement, and sustainable operations. In order to remain competitive, producing companies must adapt themselves to these conditions while taking target dimensions, such as quality and costs, into account. The three paradigms of lean, digitalization, and sustainability provide solutions for designing a production system, which is fit for the future. However, there is currently a shortage of transparent information on the degree of implementation and the specific measures required to achieve the objective of a future-proof production system (FPPS). A maturity model (MM) for FPPSs provides producing companies with a means of assessment and, therefore, is a valuable tool for addressing these challenges. In this article, an MM for FPPSs is designed that enables a holistic, multidimensional assessment of a production system in relation to the three paradigms of lean, digitalization, and sustainability. The resulting MM comprises nine areas, 25 overarching elements, and 86 elements. Consequently, the design of the MM allows the three paradigms to be assessed at different abstraction levels, using five maturity levels, thus enabling the derivation of specific measures. Furthermore, transparency of the maturity stages is achieved through the visualization of the assessment results using radar charts. Future work should focus on the industrial evaluation and application of the MM.
An established method to enhance the lightweight potential and the crashworthiness of hot stamped components is locally tailoring the part properties according to the functional purpose of the respective zones. Tailored carburization is hereby an alternative approach which offers several advantages compared to established ones, e.g. high geometrical flexibility. The process-adapted austenitization, as a lately introduced and innovative approach, showed the potential to increase the ductility of the not carburized zones while maintaining the high strength in the carburized ones. This enables an even higher grading and expands the applicability. Besides the component properties the joints between different parts of materials of dissimilar types are another crucial aspect regarding manufacturability and crash performance. Tumbling self-piercing riveting (T-SPR) is a pioneering and versatile mechanical joining process, which is a promising development of the regular process to enhance the joining capability. The combination of the tailoring approach and the mechanical joining process offers the potential to overcome current limitations in joining hot stamped parts with a component of a dissimilar material. Based on this hypothesis, the aim of this study is to evaluate this potential by investigating the impact of the material condition and combination as well as the joining parameters on the joining process, joint development and mechanical load capacity. Hereby, joints with high load capacities up to 7 kN under shear tension and an energy absorption of up to 20 J were achieved for the highest investigated tumbling velocity and a fully non-carburized upper blank.
Rising energy costs and workforce shortages compel smart factories to modernize scheduling. Dispatching rules and single-paradigm methods cannot jointly deliver responsiveness, sustainability, and Pareto optimality under stochastic disruptions. This study proposes a co-evolutionary framework tightly coupling reinforcement learning (RL) with MOEA/D via bidirectional knowledge transfer: RL trajectories seed evolutionary populations, while elite non-dominated solutions refresh RL experience. Evaluated across four benchmarks, the proposed approach reduces makespan by up to 15.6
