This work introduces a graph-based CAD assistant that predicts the next modelling operation in parametric design sequences. Real CATIA V5 models from the automotive domain are converted into directed acyclic graphs capturing feature dependencies, enabling learning directly from structural design data. A four-layer Graph Attention Network achieved a top-5 prediction accuracy of 94%, outperforming a frequency-based non-parametric baseline. The results show that graph representations and attention-based message passing provide a strong foundation for context-aware modelling assistance.

A blended learning approach was introduced to extend guideline-based training from key users to the wider end-user community in order to implement a new CAD/PLM environment. By combining asynchronous self-study of guidelines with synchronous, trainer-led sessions, the programme fostered procedural understanding and consistent modelling practices, as well as learner engagement. The results demonstrate how scalable blended learning strategies can bridge the gap between industrial training and academic education.

This paper presents a characterization approach for analysing geometric variability in industrial 3D model datasets to support the preparation of synthetic datasets for machine-learning applications. By implementing pairwise Hausdorff distances and manifold-based embedding techniques, the study identifies variability ranges required for generating representative synthetic data and demonstrates how targeted augmentation can effectively reproduce real data’s variability, ultimately leading to more reliable and robust NN model performance.

This study investigates how clinicians and technicians describe the preparation and use of CAD models in collaborative design sessions for mass personalised products. Semi-structured interviews with ten professionals were analysed using thematic analysis. The findings revealed three themes: the input data required before modelling, the additional information that supports interaction with the CAD model, and the role-specific ways in which contributors evaluate it. These insights guided the development of an initial parametric CAD model intended to support future collaborative work.

During the transition to CAD/PLM software, key users underwent guideline-based training aligned with company workflows. This practical approach, which linked tool functions to real design practices, accelerated the acquisition of skills, ensured modelling consistency, and improved understanding of digital engineering. The study identifies key users as knowledge multipliers and reveals how such methods develop competence. The findings emphasise the significance of problem-solving training and the relevance of guideline-based methods for industrial practice and design education.

This study examines how ChatGPT support influences verbal communication in synchronous collaborative CAD activities. Using a verbal protocol analysis of teams solving an embodiment design task, the results show that ChatGPT-supported teams communicated less, devoted less verbal communication to problem- and analysis-related communication, and shifted toward process and solution synthesis, indicating a shortened design co-evolution cycle in which teams move more quickly toward generating solutions. Future work should integrate these findings with broader teamwork and taskwork analyses.

This study examines how CAD geometry variations affect finite element (FE) crash simulations for automotive front rail assembly and motivate the use of combined impact measures that better reflect the physical response. Based on these insights, we outline a machine learning formulation that links geometric modifications to their simulation effects. The study centers on geometric representation, employing UVbased graph encodings to capture local shape changes and provide the basis for advancing and validating the full prediction pipeline.

The ideating phase of product design is critical, as decisions made here influence the rest of the product’s lifecycle. Usually, early preliminary designs in engineering are created with pen and paper, which are incompatible with the subsequent digital design process. In an effort to find a modeling tool for early designs that provides the creative flexibility of freehand sketching but also the further processability of digital models, this research investigates natural modeling in virtual reality (VR). To do so, a VR modeling method allowing the intuitive creation of preliminary designs as simplified computer-aided design (CAD) models is presented. The main contribution is the evaluation of this natural VR modeling method against freehand sketching in an extensive user study.

Increasing adoption of additive manufacturing (AM) makes software support for design for additive manufacturing (DfAM) more relevant. This paper presents a novel, knowledge-driven design support tool for AM that leverages a central knowledge base to provide extensible and powerful DfAM support early in the development process. The approach was implemented using Python for the knowledge base and as a plugin for Siemens NX. It offers automated design checks, optimizations, and further information through an integrated Wiki. Evaluation confirms the feasibility and benefits of the approach.

This paper investigates the use of Large Language Models (LLMs) in engineering complex systems, demonstrating how they can support designers on detail design phases. Two aerospace cases, a system architecture definition and a CAD model generation activities are studied. The research reveals LLMs' challenges and opportunities to support designers, and future research areas to further improve their application in engineering tasks. It emphasizes the new paradigm of LLMs support compared to traditional Machine Learning techniques, as they can successfully perform tasks with just a few examples.

Design sprints complement traditional teaching methods, especially in project-based learning courses. While this approach can potentially change Computer-Aided Design (CAD) usage, it is still underexplored. Therefore, this study explores the influence of design sprints on embodiment-focused CAD activities in project-based learning by examining differences in patterns of CAD user actions, focusing on design space and action types. The case involves two higher-graded and two lower-graded student design teams monitored with a non-invasive method across a two-day design sprint event.

This study envisions a unified paradigm for design for automated disassembly. The goal is to integrate disassembly insights related to precious material recovery with the design phase for sustainable lifecycle management.Targeting plastic products with embedded electronics, the collaboration between design and robotic engineers aims to program a robot for disassembly for the LEGO® motor (45603) as demonstration, emphasizing a disassembly map as a vital tool. By considering the limitations and strengths of robots, this research pioneers a design for disassembly framework.

With recent advancements in Virtual reality (VR), 3D design in VR has gained significant interest from both academia and industries. However, the development of these VR CAD tools is either skewed towards the creative industry or simply mimicking conventional CAD. This paper presents three different tools, analyzes them, and compares their capabilities over various performance parameters. The paper finally suggests where these tools can be used in the design process and some critical pathways for developing VR-based CAD modeling software for practical use in the engineering design industry.

The presented study investigates differences in engineering designers' CAD performance when modelling from two types of projections in technical drawings – isometric and orthographic. The results revealed significant differences in the percentage of correctly replicated components' size and shape, indicating better CAD outcomes when generating CAD models from the orthographic projection. In addition, a comparison of duration, as well as the number and type of sketch entities, sketch relations, and CAD features, showed that CAD modelling processes were similar in both conditions.

This article presents an fNIRS experiment investigating cognitive differences between physical and digital prototyping methods in designers (N=25) engaged in open and constrained design tasks. Initial results suggest that physical prototyping yields increased hemodynamic response (i.e., brain activity) compared to digital design, and that constrained design yields increased hemodynamic response compared to open design, in the prefrontal cortex. Further work will seek to triangulate results by investigating potential correlations to design processes and design outputs.

In HMLV manufacturing, assembly mistakes by operators are common due to the ever increasing product variability and complexity. If mistakes can be detected early-on in the design process, product designers can reduce the possibility for mistakes. We present an algorithm to automatically detect and evaluate potential misplacements of parts that need to be fastened. Evaluation starts from a product CAD and returns the risk of misplacement as well as visual feedback on possible misplacements. An implementation with FreeCAD of our algorithm is illustrated on different use cases.

In the field of 3D model reconstruction, manifold methods have been developed that derive CAD models from 3D scan data. Opposed to classical CAD modelling, where surface and solid modelling exist, a further diversification of modelling techniques is observed, caused by different methods to build up the geometry. This research introduces a new classification, the so-called Level of Complexities. It can be applied to the complete Reverse Engineering process chain and lays the foundation for further research on how to match requirements arising from all process steps and downstream applications.

In objected-oriented design, "smells" are symptoms of code violating design principles. When a deadline is looming, decisions can affect the long-term quality of a code or CAD. Given this and the similarities between object-oriented code and CAD models, this paper introduces a set of CAD smells. These smells are derived from a top-down review of potential CAD smells mapped against the reported code smells that violate abstraction, modularity, encapsulation, and hierarchy principles. This list was further reviewed considering CAD systems and specific examples (some illustrated in the paper).

Manually exploring the solution space for different variants of a product for a given set of requirements is ineffective regarding product development time and adaptation to dynamic customer requirements. Variant generation coupled to optimization algorithms offers possibilities to search the solution space in an automated way. This paper provides a framework to build a generative parametric design environment for functional assemblies by implementing analysis as well as synthesis methods in computer-aided tools. The procedure is presented using the example of a coffee machine.

The integration of additive manufacturing processes into the teaching of students is an important prerequisite for the further dissemination of this new technology. In this context, the DfAM is of particular importance. For this reason, this paper presents an approach in which a connection is made between methodical product development and practical implementation by AM. Using a model racing car as an example, students independently develop significant improvements of particular assemblies. A final evaluation shows that the students have significantly improved their skills and competencies.