Exploring Virtual Geometry Assurance for the Nominal Buck Verification Process at Polestar Master’s Thesis in Production Engineering HUGO KÅHRE SAMUEL PERSSON DEPARTMENT OF INDUSTRIAL AND MATERIALS SCIENCE CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2025 www.chalmers.se www.chalmers.se Master’s Thesis 2025 Exploring Virtual Geometry Assurance for Nominal Buck Verification Process at Polestar A work at Polestar in partnership with PE Geometry about exploring the virtual possibilities for the nominal buck verification process HUGO KÅHRE SAMUEL PERSSON Department of Industrial and Materials Science Chalmers University of Technology Gothenburg, Sweden 2025 Exploring virtual geometry assurance possibilities for nominal buck verification pro- cess at Polestar. © HUGO KÅHRE, 2025. © SAMUEL PERSSON, 2025. Examinator and Supervisor: Kristina Wärmefjord, Department of Industrial and Materials Science Master Thesis 2025 Department of Industrial and Materials Science Chalmers University of Technology SE-412 96 Gothenburg Telefon +46 31 772 1000 ii Abstract This thesis investigates the feasibility of digitalizing the nominal buck verification process at Polestar using virtual geometry assurance software such as RD&T. The nominal buck, a physical aluminum structure representing the ideal chassis geometry, is currently employed in Polestar’s verification process to evaluate the fit and align- ment of interior and exterior vehicle components. Although the process is reliable with a high accuracy when it comes to asserting the geometric state of a part, the process is costly, time-consuming, and environmentally demanding. The objective of this study is to assess whether a virtual alternative can replace or complement the physical verification approach to improve efficiency, accuracy, sustainability, cost- effectiveness and flexibility. Through a combination of literature reviews, industry interviews, and a detailed case study, the project explores the capabilities of RD&T for simulating geometric variation and part assembly. The digital method involves 3D scanning of physical parts, creating virtual fixtures, and applying statistical variation analyses to predict assembly behavior and compliance with tolerances. The research also includes a comparative analysis between the physical and digital approaches, highlighting ad- vantages such as reduced lead times, lower material consumption, increased process flexibility, and enhanced data traceability. The results suggest that a virtual verification process using RD&T can signifi- cantly improve the robustness and efficiency of the geometry assurance processes at Polestar, with additional sustainability benefits including waste reduction and streamlined global collaboration. However, challenges such as software limitations, implementation complexity, and required changes in workflow and skill sets must be addressed. This thesis contributes to the advancement of digital manufacturing practices and supports Polestar’s strategic sustainability and innovation goals. iii Acknowledgements We would like to express our deepest gratitude to our examiner, Kristina Wärme- fjord, for her exceptional guidance, continuous support, and remarkable patience in addressing our many questions throughout the thesis. Her expertise and en- couragement have been instrumental in shaping both the direction and outcome of this work. Our sincere thanks also go to Lars Lindqvist, developer of the RD&T software, whose persistent efforts and technical insights were crucial in overcoming challenges related to our simulation models. This thesis would not have reached its current level without his invaluable assistance. We are grateful to Andreas Stenlund, site manager at the PE Geometry office in Gothenburg, for introducing us to the field of geometry assurance and providing foundational knowledge that helped us understand the core concepts. His genuine interest in our project and generous sharing of expertise helped us stay on the right track. We also extend our heartfelt thanks to Niklas Nylén, geometry assurance engineer at PE Geometry and consultant for Polestar, for his steadfast support and technical input throughout the thesis. His clear feedback, access to critical re- sources, and helpful discussions significantly contributed to the progress of our work. A special thanks to Charlie Berner, manager at the ME department at Polestar, for serving as our supervisor during the project. His help in keeping the project on schedule and connecting us with key contacts at Polestar has been greatly appreci- ated. We also wish to acknowledge Elliot Edholm, PE Geometry technician, for his sup- port during the scanning process and introduction to the PolyWorks software. The case study would not have been possible without his hands-on expertise. The au- thors also thank Peter Edholm, CEO of PE Geometry, for his ongoing support and for facilitating contact with several experts in the field, which proved invaluable to the completion of this thesis. Finally, we are thankful to everyone else who contributed to this project in various ways. Your support and encouragement have made this project possible. Hugo Kåhre Samuel Persson Gothenburg, 2025 iv v List of Acronyms Below is the list of acronyms that have been used throughout this thesis listed in alphabetical order: 3D Three-Dimensional BB Black Box BIW Body In White CAD Computer-Aided Design CAT Computer-Aided Tolerancing CMM Coordinate Measuring Machines CM4D Coordinate Measurement for Dimensional Data CNC Computer Numerical Control CSV Comma-Separated Values DFA Design for Assembly DFM Design for Manufacturing DFMA Design for Manufacturing and Assembly DMIS Dimensional Measuring Interface Standard DP Design Parameter FEA Finite Element Analysis FR Functional Requirement MIC Method of influence-coefficients MMC Maximum Material Condition MP Measure Point MSV Machine Vision Systems PCA Process Capability Analysis PDP Product Design Process PP Pilot Production PS Project Start PSDS Polestar Development System RD&T Robust Design and Tolerancing RDM Robust Design Methodology SPC Statistical Process Control TT Tooling Trials VF Virtual Fixture VP Verification Prototypes vi Contents List of Acronyms v 1 Introduction 1 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Polestar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2.1 Polestar 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 PE Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.4 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.5 Research Objectives & Questions . . . . . . . . . . . . . . . . . . . . 4 1.6 Scope and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.7 Contributions of the Study . . . . . . . . . . . . . . . . . . . . . . . . 5 1.8 Outline of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2 Literature Study 7 2.1 Polestar’s Production System . . . . . . . . . . . . . . . . . . . . . . 7 2.2 The Nominal Buck . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.3 Geometry Assurance in Manufacturing . . . . . . . . . . . . . . . . . 9 2.3.1 Core Methodologies in Geometry Assurance . . . . . . . . . . 11 2.3.1.1 Product Design . . . . . . . . . . . . . . . . . . . . . 11 2.3.1.2 Geometric Tolerances . . . . . . . . . . . . . . . . . 15 2.3.1.3 Statistical Process Control (SPC) . . . . . . . . . . . 15 2.3.1.4 Perceived Quality . . . . . . . . . . . . . . . . . . . . 16 2.4 Geometry Assurance Tools . . . . . . . . . . . . . . . . . . . . . . . . 17 2.4.1 Measurement Technology . . . . . . . . . . . . . . . . . . . . . 18 2.4.1.1 Coordinate Measuring Machine . . . . . . . . . . . . 18 2.5 Software for Digital Verification . . . . . . . . . . . . . . . . . . . . . 20 2.5.1 RD&T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.5.2 Virtual Fixture . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.5.3 Meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3 Methodology 26 3.1 Methodology Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.2 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.2.1 Literature Study . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.2.2 RD&T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.2.3 Interviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 viii Contents 3.2.3.1 Qualitative Interviews . . . . . . . . . . . . . . . . . 27 3.2.3.2 Quantitative Interview . . . . . . . . . . . . . . . . . 29 3.3 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.3.1 Step-by-Step Guide for RD&T, Meshing, PolyWorks, and Scan- ning Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.4 Comparative Analysis - Physical and Digital Verification Methods . . 30 3.5 Impact of Virtual Implementation . . . . . . . . . . . . . . . . . . . . 30 4 Result 31 4.1 Findings from Industry Interviews . . . . . . . . . . . . . . . . . . . . 31 4.1.1 Current Physical Nominal Buck Process . . . . . . . . . . . . 31 4.1.2 3D Scanning Technology . . . . . . . . . . . . . . . . . . . . . 36 4.1.3 Virtual Initiatives at External Organizations . . . . . . . . . . 37 4.1.4 Research Within Virtual Geometry Assurance . . . . . . . . . 38 4.2 Quantitative Interview Findings . . . . . . . . . . . . . . . . . . . . . 40 4.3 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.3.1 Virtual Verification Methodology . . . . . . . . . . . . . . . . 44 4.3.2 Virtual Verification Using RD&T . . . . . . . . . . . . . . . . 47 4.3.2.1 Summary of Steps and Necessary Files & Data . . . 47 4.3.2.2 Step 1: Virtual Fixture - Import CAD and create positioning system . . . . . . . . . . . . . . . . . . . 48 4.3.2.3 Step 2: Virtual Fixture - Optimize virtual fixture sphere location using the Scan Planning toolbar . . . 55 4.3.2.4 Step 3: Black Box - Import scan data . . . . . . . . 60 4.3.2.5 Step 4: Virtual Assembly - Using the scan data to analyze part variation . . . . . . . . . . . . . . . . . 64 4.3.2.6 Output of Virtual Verification Method using RD&T 86 4.4 Comparative Analysis: Physical versus Virtual Verification . . . . . . 87 4.4.1 Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 4.4.2 Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 4.4.3 Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.4.4 Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.4.5 Flexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.4.6 Summary of comparison . . . . . . . . . . . . . . . . . . . . . 91 4.5 Impact of Virtual Implementation . . . . . . . . . . . . . . . . . . . . 92 4.5.1 Implementation Feasibility . . . . . . . . . . . . . . . . . . . . 94 4.5.2 Automation of Process . . . . . . . . . . . . . . . . . . . . . . 96 5 Discussion 97 5.1 Interpretation of Results . . . . . . . . . . . . . . . . . . . . . . . . . 97 5.1.1 Technological and Economical Research Questions . . . . . . . 97 5.1.2 Sustainability Research Questions . . . . . . . . . . . . . . . . 99 5.2 Implementation Challenges & Constraints . . . . . . . . . . . . . . . 100 5.3 Recommendations for Future Work . . . . . . . . . . . . . . . . . . . 102 6 Conclusion 103 ix Contents References 104 A Interview Questions - PE Geometry I B Interview Questions - Polestar III C Interview Questions - Outside Organizations V D Quantitative Form Questions VII E Scanning Procedure and PolyWorks IX E.1 Step 1: Scanning Preparation - Setting Up the Sphere Fixtures . . . . XI E.2 Step 2: Scanning Procedure - Open PolyWorks and Gather Scan DataXIII E.3 Step 3: Postprocessing - Import CAD Part and Create Features . . . XVII E.3.1 If there are no suitable objects located on the scan surface to use as features . . . . . . . . . . . . . . . . . . . . . . . . . . . XXV E.4 Step 4: File Export - Export Necessary Files to RD&T . . . . . . . . XXXIV F Meshing Procedure XXXVIII F.1 Step 1: Import CAD model into Hypermesh . . . . . . . . . . . . . . XXXVIII F.2 Step 2: Creating the mesh . . . . . . . . . . . . . . . . . . . . . . . . XXXIX F.3 Step 3: Exporting the mesh . . . . . . . . . . . . . . . . . . . . . . . XLII x 1 Introduction 1.1 Background In the beginning of the 20th century, mass production became a central theme in global manufacturing and thus also interchangeability where tolerances needed to be specified (Söderberg et al., 2016). Following the second world war the overall quality of products began to increase in Japan and then increased even further in the Western societies by the 1980’s. By this time, all tolerancing were done physi- cally using measuring tools and formulas. In pace with technological advancements, machines and software were developed to simulate different tolerancing processes to make it cheaper, easier and faster. Today, many companies — particularly in the automotive manufacturing industry — use these software tools to simulate various processes, such as variation analysis, contribution analysis, and stability analysis of their parts and assemblies. In modern terminology, these processes are typically grouped under the field of geometry assurance, particularly within the automotive industry. Consequently, geometry assurance will be the central focus of this thesis. Polestar currently manufactures an aluminum bodywork model of cars in develop- ment called a nominal buck as the first step in their verification process. This is done to test how different interior and exterior parts fit when they are assembled together, as well as how fasteners should be placed in the car. All of this is currently done physically, leading to substantial cost and material usage, as well as extensive lead times. Depending on the standard of manufacturers, tolerances can differ, which with the current process are discovered at a quite late stage in the production chain. Additionally, if parts are within tolerances but screw-holes are slightly misplaced, fasteners occasionally need to be moved or re-engineered, resulting in even further delays. To potentially reduce lead times, increase efficiency, and decrease costs, Polestar wants to explore the possibility to conduct nominal buck tests as these in virtual environments instead, with help of Computer-Aided Tolerancing (CAT) tools such as Robust Design & Tolerancing (RD&T). To give the project a clear objective aligned with sustainability initiatives, research questions are formulated around economic, technological, and environmental as- pects. In discussions about digitalization initiatives within production environments, two key themes frequently emerge: cost reduction and increased efficiency. Addi- tionally, waste reduction and changes in employment are also central concerns. By digitalizing the verification process, the need for a physical nominal buck is signif- 1 1. Introduction icantly reduced, leading to less material waste and contributing to both economic and environmental sustainability. However, digitalization may also affect workforce requirements, potentially altering the number of personnel needed. To explore these considerations in depth, six research questions have been developed, as outlined in Chapter 1.5. 1.2 Polestar The history of the Polestar brand started in 1996 when a small Swedish team called Flash Engineering started building and racing purpose-built Volvo vehicles which were used in the Swedish Touring Car Championship (O’Steen, 2024). They later became Polestar Racing and became an official partner to the Volvo Cars brand. Volvo Cars fully acquired Polestar in 2015. In 2017, Polestar branched out and became their own independent company, as they released their first car model - Polestar 1 (Rabe, 2017). With this release they positioned themselves as a premier maker of electric and plug-in hybrid luxury cars (O’Steen, 2024). Polestar 1 was manufactured in Chengdu, China, and the last model was manufactured in 2021. An example of a Polestar 1 vehicle can be seen in Figure 1.1. Figure 1.1: The Polestar 1 with a total system output of 609 horsepower (HP) and 1000 Nm torque (Polestar, 2025b). Polestar currently have three models in their supply - Polestar 2, Polestar 3, and Polestar 4. Manufacturing is outsourced, as Polestar 1 and Polestar 2 (their 5-door liftback) have both been manufactured in China (Chengdu and Luqiao respectively) (Polestar, 2025c). Polestar 3 (the electric SUV) on the other hand, is manufactured in South Carolina, the United States (Polestar, 2025d). Polestar 4 is their SUV 2 1. Introduction coupé and is produced in Hangzhou, China, with additional production planned in Busan, South Korea, in 2025 (Polestar, 2025e). All manufacturing is therefore outsourced. However, their main office is located in Gothenburg, where 1156 em- ployees are currently working (Allabolag, 2025). Polestar has sustainability as one of their core priorities, with clear sustainability goals for the next couple of years. By 2030, they hope to manufacture a completely climate neutral vehicle, and halve their greenhouse emissions per sold vehicle (Polestar, 2025a). By 2040 they hope to achieve climate neutrality throughout their whole value chain. Polestar is the main stakeholder in this project. 1.2.1 Polestar 5 The specific vehicle of interest during this thesis is Polestar’s coming vehicle - Polestar 5, which is planned to be sold on the market by 2026. The Polestar 5 is produced in Chongqing, which means the nominal buck (which is explained in chapter 2.2) is also located there. All parts, data, and other references used dur- ing this thesis will refer to the Polestar 5. Figure 1.2 shows a prototype of the vehicle. Figure 1.2: Prototype of the Polestar 5 at the 2023 Goodwood Festival of Speed, United Kingdom (Polestar, 2025f). 3 1. Introduction 1.3 PE Geometry PE Geometry are leading experts in the field of geometry assurance (PE-Geometry, 2025). The company was founded in 1999 by the current CEO, Peter Edholm, in Gothenburg. PE Geometry Inc, daughter company of PE Geometry, opened up in 2016 in North Carolina in the United States to serve their American customers. PE Geometry has their main office in Mölndal, and their employees work mostly as consultants for other companies, mainly customers within the automotive indus- try. For example, many geometry assurance processes at Polestar are outsourced to PE Geometry, which means that they play a crucial role in this project as both supervisors and stakeholders. PE Geometry states themselves that for them, "geom- etry assurance includes all processes that serve to create a geometrically robust and well defined product, both functionally and aesthetically". This is usually achieved by using advanced Three-Dimensional (3D) simulation software which allows for variation analysis in great detail. 1.4 Problem Statement In Polestar’s current design process, all tolerancing are performed either physically or with Computer-Aided Design (CAD) software. This entails that the digital model which is constructed is a nominal model, meaning it is a perfect model without deviations. As such, when simulating variations in software like RD&T the model represented will always be perfect. This will not be the case for manufactured parts. Since there is always noise and disturbances in machines used for production, the parts will always have slight variations compared to the nominal model. To enable a better understanding and more realistic simulation of parts and how they are situated as opposed to each other, the parts used in simulation model needs to be as close to the real life model as possible. Furthermore, because of the earlier mentioned variation in parts, engineers cannot be sure that all parts fit within the tolerances specified. This creates a problem for engineers to quickly analyze if batches are within tolerances or not. 1.5 Research Objectives & Questions To be able to evaluate the results of this thesis, certain research objectives and questions need to be posed. These questions will be used to keep track of the goals for the report and finally clearly establish the conclusions of the thesis. There are two types of questions posed. The first is technical/economic questions which relate to the main problem at Polestar. Secondly, there are sustainability questions that relate to how Polestar would be sustainably impacted by solving the issue. Technological & Economic Questions: • RQ1: What are the advantages/disadvantages of simulating nominal buck geometry assurance within a virtual space like RD&T? 4 1. Introduction • RQ2: How do you ensure the most realistic simulation for nominal buck verification? • RQ3: Can all parts be effectively simulated using a virtual nominal buck? • RQ4: How would a transition to virtual geometry assurance impact Polestar? Sustainability Questions: • RQ5: How would waste be reduced by digitalizing the nominal buck verifica- tion process? • RQ6: How would employees working with nominal buck verification be im- pacted by digitalizing the nominal buck process? 1.6 Scope and Limitations The aim of this master’s thesis is to evaluate the feasibility and potential benefits of digitalizing the nominal buck verification process using software tools such as RD&T. The outcome will guide Polestar’s decisions regarding if virtual modeling of nominal buck verification is viable and how they should proceed with implementing it. Certain limitations have had to be considered during the project. The most prevalent limitation is that the only geometry assurance software available is RD&T. This will not make it possible to compare different variation simulation software. Another limitation is the fact that the nominal buck is located at the Polestar production plant in Chongqing, China, making it impossible to carry out any tests with the nominal buck. Three plastic parts were used during the case study to demonstrate the digital method developed during this project. However, the project group only had access to these parts for one week, which was also a limitation. 1.7 Contributions of the Study This thesis will contribute in several ways to Polestar’s working methods regarding geometry assurance for nominal buck verification. Increasing Polestar’s knowledge regarding virtual geometry assurance, realistic tolerance simulations and advantages as well as disadvantages will come to light. Furthermore, this thesis will contribute to research on how companies should and can develop their working methods toward more realistic virtual simulation within robust design and tolerancing software. 1.8 Outline of the Study This thesis is organized into six chapters. Chapter 1 introduces the background, objectives, and scope of the study. Chapter 2 presents a literature study with a focus on geometry assurance, as well as important physical and digital technologies in relation to this. Chapter 3 describes the research methodology, detailing the 5 1. Introduction analytical framework, interview process, and case study design. Chapter 4 presents the results of the interviews, the description of the case study, and the comparative analysis between the physical and digital verification methods. Chapter 5 provides a discussion of the findings in relation to the research questions and broader industry implications. Finally, Chapter 6 concludes the study with a summary of key insights, limitations, and suggestions for future work. 6 2 Literature Study This chapter presents the literature study that was conducted to find crucial infor- mation within the field of geometry assurance. Geometry assurance is the concept of ensuring that products are within quality tolerances, and includes all activities and tools needed to do so (Söderberg et al., 2016). Furthermore, related topics to this field such as current methods of geometry verification are also explored, which will pave the way for constructing a digital method at later stages. 2.1 Polestar’s Production System Polestar is developing and producing vehicles according to their core production system - the Polestar Development System (PSDS). The PSDS is used to design, develop, and produce products that are then delivered to their customers, which is achieved by balancing time, cost, and quality. The PSDS provides a generic frame- work of key milestones, gateways, phases, and deliverables to support the delivery of their vehicle program. The PSDS consists of distinct areas, each one specializing in governing crucial aspects of Polestar’s business processes. An overview of the PSDS with its various departments can be seen in Figure 2.1. Figure 2.1: An overview of the PSDS with its important areas. Work is conducted within Manufacturing, Logistics & Launch in this project. 7 2. Literature Study As can be seen, one area is Manufacturing, Logistics and Launch, where the area of manufacturing is handled by the department of Product & Process Manufacturing Engineering. They have the responsibility to work in partnership with the busi- ness to agree on product and process solutions, to allow the delivery of vehicles according to quality agreements, and to enable profitable manufacturing strategies. The department also has the responsibility of planning production and preparing plants for new products and processes, which also involves the tools and equipment necessary to do so. Product & Process Manufacturing Engineering consist of four function areas, one of them being Manufacturing Engineering Geometry Assurance. This department governs all processes related to verifying the geometries of parts assembled in a Polestar vehicle. This involves, for example, developing methods for part verification, which is currently done with help of a physical nominal buck. It is within this area that the project group will conduct work to improve Polestar’s current verification process, and to assess the possibility to perform that verifica- tion with help of digital tools, either as a total replacement or an additional tool that may help improve the company’s business operations. Polestar is also working according to four distinct phases during their development process. The first one is the Project Start (PS), a phase of early production runs to test new manufacturing processes and identify issues. The next phase is the Verification Prototypes (VP), which is when prototypes are built to verify the design and engineering aspects of their vehicles. The third phase is the Tooling Trials (TT), which consist of trials to test production tools and equipment. The fourth and last phase is the Pilot Pro- duction (PP), which consist of initial production runs using the finally developed production processes, tools, and equipment. 2.2 The Nominal Buck The nominal buck is a framework of the car chassis where all dimensions are nominal (see Figure 2.2). Other names for the nominal buck is a master buck, body buck, etc. and the naming varies between automotive manufacturers. The nominal buck is called "nominal" because all dimensions of the nominal buck are close to perfect and closely resembles CAD data of the vehicle. Common material is either aluminum or steel, and Polestar uses aluminum for their nominal buck. A nominal buck is therefore constructed for each car model. Purchased or manufactured vehicle parts, both exterior and interior parts, are then physically inserted into the nominal buck, to visualize the fit of the parts in a nominal environment. To achieve this, fixtures and fasteners are used to hold the parts in place using attachment plates during measurement and analysis. Attachment plates are used, because when small design changes need to implemented to an already manufactured nominal buck, the attach- ment plates can be re-machined instead. The nominal buck offers the possibility to analyze part dimensions and determine if they are within tolerances, both in terms of functionality and aesthetics. The perceiver, or worker, can then assume that the parts are either okay and deliver these to final assembly, or not okay and send them back to the supplier for reworking. The nominal buck is therefore a physical verification tool and an important step in the vehicle production chain. 8 2. Literature Study Figure 2.2: One of Polestar’s nominal bucks. The photo has been brought from internal sources (Polestar, 2025). 2.3 Geometry Assurance in Manufacturing Geometry assurance can be described as a number of activities, all contributing to minimizing the effect of geometric variations on a final product (Söderberg et al., 2016). This methodology is the base of the project and it is from this source that all other important methodologies and concepts will stem from. During the early twentieth century with the introduction of mass production, manufacturers real- ized that two products of the same kind may never be completely identical, due to the stochastic nature of manufacturing. Because of this, tolerance requirements are needed, as they direct focus to certain parts of a product that are more sensitive to geometric variations. Geometry assurance processes are implemented to ensure both functional and aesthetic qualities of a product (Rosenqvist et al., 2016). A product’s realization phase can be divided into three distinctive phases - the concept phase, the verification phase, and the production phase (Söderberg et al., 2016). This re- alization phase is visualized in Figure 2.3 below. Assuring that products are within geometrical tolerances is important within all these phases, to prevent that geomet- rical variations are detected at a much later stage in the product chain, where costs and environmental impact of remanufacturing can be problematic. 9 2. Literature Study Figure 2.3: The different phases of a product’s realization phase, including typical ac- tivites within these (Söderberg et al., 2016). The concept phase is the start of the product’s life cycle, when both product and production are conceptualized (Söderberg et al., 2016). At this stage, geometry as- surance can be initiated by simulating product variations, which can give hindsight of future outcomes. Simulation is a robust tool in such situations, as realistic data can be implemented to assure that the model is realistic in the sense that it mimics the real world product and production. Statistical tolerance analysis can be used to predict how the product’s geometry may vary during production. During this stage, the visual appearance and functionality of the product can be optimized. For example, ensuring that the product is geometrically robust, which could be achieved by having an uncoupled design. In an uncoupled design, each output or Functional Requirement (FR), is controlled by only one output or Design Parameter (DP). An uncoupled design is much easier to control than an coupled design, as geometrical variations in one part (in an assembly) will not affect another part in that assembly. An uncoupled design therefore ensures that the design is geometrically robust, in other words, that the design is unsensitive to variations. However, some authors mention that it may be difficult to create digital models of a complex manual as- sembly operation, as some assembly factors (such as assembly complexity), are not included in the digital models (Rosenqvist et al., 2016). The authors also mention that a study demonstrated that the correlation between digital CAT simulations and actual outcome in a production environment is low. Furthermore, assembly complexity seems to have a major impact on product quality. In the verification phase, product and production are physically implemented and tested (Söderberg et al., 2016). As product design already has been concluded the important step of this stage is to focus on the efficiency of the production. The Computer Numerical Control (CNC) machines are programmed to manufacture the product according to design parameters, a process which needs to ensure that criti- 10 2. Literature Study cal parts of a product are within tolerances. Lastly comes the production phase, which is when all production processes are adjusted and product is in full production (Söderberg et al., 2016). To ensure product quality, geometrical dimensions and functionality need to be measured, especially after final production processes before product is being packaged and sent out to customers. The nominal buck process is used during the production phase, as it is a verification tool before parts are sent to final vehicle assembly. Thus it is within this phase that this master thesis project is focused. As the objective of this project is to digitalize an activity in Polestar’s geometry assurance process, it is worth mentioning the importance of data collection. As part of Industry 4.0 initiatives the continuous collection of data offers predictability and a higher rate of steadiness (Barton et al., 2024). Collected data can be measured throughout time and analyzing this creates an overview of production. This is one of the key advantages of digital tools compared to traditional physical measuring methods. 2.3.1 Core Methodologies in Geometry Assurance A comprehensive understanding of geometry assurance requires familiarity with sev- eral key methodologies and concepts. Initially, it is essential to recognize how a product’s design can influence its sensitivity to geometric variation—both as an individual component and within a complete assembly. Additionally, the role of ge- ometric tolerances must be considered, alongside the purpose of Statistical Process Control (SPC) processes, which is commonly employed by manufacturing organiza- tions to enhance process stability and quality control. Lastly, the notion of perceived quality from the customer’s perspective is crucial, as it provides insight into the ex- tent of geometric variation a product can exhibit before it negatively impacts the user’s perception and overall satisfaction. All of these concepts will be explained in this chapter. 2.3.1.1 Product Design The importance of the product’s design must be considered when seeking to increase product quality. This is important to consider when studying the geometry assur- ance process at Polestar. Below is information regarding robust design principles and design for assembly & manufacturing. Robust Design Principles Variation in products affect their performance, which ultimately can lead to product failure, dissatisfied customers, the need for increased quality control, and added de- velopment and service costs, all of which impact the profit of the supplier (Ebro and Howard, 2016). Variation inflicts the intention of consistent behavior. Therefore, the design engineer should design the product in such a way that variation in func- tional behavior is minimized. The design engineer may come up with an entirely nominal product model, but as the world is always stochastic and two products of the same kind may never be completely geometrically identical, another important 11 2. Literature Study task is to minimize the variation in functional performance. The sources of vari- ation (noise sources) are commonly divided into three categories: environmental factors including variations in the conditions of use (external noise), deterioration including the aging of products (internal noise), and variation in production (unit- to-unit noise) (Torben Hasenkamp and Arvidsson, 2007). Each such noise affecting variation in performance can therefore arise in a specific phase of the product’s life cycle. There are two ways to minimize noise in products. One approach is to elim- inate them, which may be difficult as the noise factors might be unknown, costly, or even impossible to eliminate. The other approach is to design products in such a way that they are insensitive to variations, which is the core of Robust Design Methodology (RDM). The key idea is that, instead of trying to control the environ- ment, the product should be designed in such a way that it is insensitive (robust) to the environment. This methodology has its roots in the work of Japanese engineer Genichi Taguchi during the 1980s (Andersson, 1997). As RDM focuses on the ini- tial design of a product, it should be initiated early in the Product Design Process (PDP), as product design plays a key role in its potential for variation in perfor- mance. The important thing to have in mind is that products should be designed so they have a high level of geometrical robustness. Different examples of sensitive designs, and how these can be altered to be more robust, is demonstrated in Figure 2.4. As robust design principles focus on minimizing the sensitivity of products to variations in manufacturing processes, early tolerance analysis and front-loading of design processes are essential to achieve robust designs (Goetz et al., 2018). In this specific project, this is important to be aware of as different automotive parts will be studied in relation to each other, how they are interconnected and assembled. Even though the main task is to find an alternative, digital method to the current physical verification process, it is important have knowledge about the fact that geometrical robustness plays a key role in variation. If a specific model of a part has problems of variation between units, it may be because of machining settings, but it may also be that the design is not geometrically robust. 12 2. Literature Study Figure 2.4: Examples of sensitive design and how these can be designed to be more robust (Ebro et al., 2012). Design for Manufacturing & Assembly Design for Manufacturing (DFM) is a discipline where the goal is to design prod- ucts that are easy and cost-effective to produce (Bogue, 2012). Design for Assembly (DFA) is a closely related discipline, but with the goal of making the product(s) easier and more efficient to assemble. Together they are often referred to as Design for Manufacturing and Assembly (DFMA). These both disciplines are important when considering product quality. Products that are made easier to both manufac- ture and assemble are in need for less quality control processes. One typical DFA guideline is to minimize the number of parts in an assembly (Oh and Behdad, 2016). The idea is that more parts require more steps in an assembly, and thus requires more time, precision, and resources. It also requires more frequent contact with suppliers. One important DFA principle is therefore to remove unnecessary parts from an assembly, or joining parts together in the inital design. A case of such is demonstrated in Figure 2.5 below. There are many great examples of DFA ini- tiatives in historical manufacturing settings (Bogue, 2012). Ford Motor Company have used many DFMA tools throughout their operations, training their engineers 13 2. Literature Study in the DFA methodology. Ford is also requiring their vendors to conduct DFA anal- yses before submitting bids on subcontracted products. This means they make sure that all parts coming into their final assembly have successfully went through DFA assessments. Their operations involve gathering product designers, manufacturers, and suppliers, for regular meetings where both the conceptual design for their future products and products currently being manufactured are discussed. Ford’s imple- mentation of DFMA tools has resulted in significant benefits, including enhanced product quality and increased customer satisfaction. A key takeaway from these disciplines is the critical importance of the initial design of each component. In the case of a Polestar vehicle, which consists of thousands of individual parts, every com- ponent must be designed with efficient manufacturing and assembly in mind. These components are subsequently analyzed in later stages using CAD and simulation software. Integrating this approach within manufacturing processes ensures that a higher proportion of components will meet the required standards during nominal buck verification - whether the verification is conducted digitally or through physical methods. Figure 2.5: The Figure shows how the initial design of a product can be more or less suited for assembly. The design to the left have many screws and assembly steps, while the design to the right require much less number of screws and assembly steps. The ultimate goal is to assemble the body to the back cover, however, the differences in assembly process affects time, cost, and flexibility considerably (Disher, 2020). 14 2. Literature Study 2.3.1.2 Geometric Tolerances Geometric tolerances are fundamental to geometry assurance, as they govern ac- ceptable deviations from ideal geometry and are critical for maintaining product quality and functionality (Haberhauer, 2014). They provide a standardized frame- work that ensures mutual understanding between designers, manufacturers, and suppliers regarding geometric and dimensional accuracy. See Figure 2.6 for exam- ples of geometric tolerances. Figure 2.6: Examples of geometric tolerances (Pairel et al., 2007). In modern product development, precise tolerancing is vital due to increased mar- ket competition and the need to reduce defects and production costs (Armillotta and Semeraro, 2011). Well-defined tolerances not only guide manufacturing but also directly influence downstream processes such as assembly feasibility, inspection complexity, and overall production efficiency. Moreover, they serve as documented assurance for customers, confirming that the product meets both functional and aesthetic requirements. 2.3.1.3 Statistical Process Control (SPC) Statistical Process Control (SPC) is a term that covers a broad scope, and it includes many different activities intented to control a manufacturing process. The main ob- jective of these activities is to control quality and decrease variation in production (Montgomery et al., 1994). In short, SPC looks for signals that represent assignable causes, which can be thought of as external disturbances that increase variation. It is often assumed that process data can be seen as a set of statistically independent observations that fluctuates around a constant mean. Each such observation is then considered to be either special cause noise or normal cause noise, depending on if they are within the upper and lower tolerance limits. In order to control the process, the special cause noise, or unnatural noise, need to be located and eliminated. An 15 2. Literature Study example of a control chart demonstrating this can be seen in Figure 2.7. SPC is usually implemented as a "top-down" initiative, starting with upper management, going downward to improve general quality within the company. The purpose of SPC is to monitor production processes, which identifies certain signals that may arise, and eliminating these signals to minimize variability and maximize control. This includes any type of action a company takes to monitor, analyze, and address certain processes in their production systems. SPC operations can include tools such as control charts, Process Capability Analysis (PCA), or digital tools such as Machine Vision Systems (MSV), computer-based SPC systems, or even predictive maintenance (which has gained much attention in recent years). For example, in the automotive industry, this could include a company doing physical measurements of all parts coming in to final assembly - a constant monitoring of geometries and di- mensions. This data can then be stored, which with physical measurements often is done manually, but it enables the possibility to track historic data, and assumptions and conclusions can be drawn about the process throughout time. Certain parts may have higher variability, and once the root cause of such varieties are addressed, the problem has been acknowledged and eliminated. This whole chain of event is a typical SPC process. This subject is of importance, as a digitalization of an ge- ometry assurance process also involves digitalizing the SPC process. Once parts are measured and verified digitally, this opens up the possibility for storing and com- municating data about variations effectively. This can be productively automated which reduces both cost and time. Figure 2.7: A control chart which categorizes data points based on normal cause varia- tion and special cause variation (CQE, 2025). 2.3.1.4 Perceived Quality Perceived quality is the expression customers feel toward purchased products. This is derived through touch, sight, sound, and scent, and it largely affects how cus- tomers feel about their suppliers (eg. brand loyalty). Research has shown that the choice to buy a new car is largely affected by several factors such as the brand value and the visual perception of interior and exterior parts (Tovillo et al., 2024). 16 2. Literature Study The visual perception includes things such as color, which is explored deeply by Tovillo & colleagues. However, it also includes material quality and geometrical quality. The geometric outcome is an important factor that influences the perceived quality of the final car (Lindau et al., 2015). In another report, the authors has constructed a Perceived Quality Framework to characterize different units of quality perceptions, and rank them in importance according to what customers deemed to be most important (Stylidis et al., 2019). This framework can be seen in Figure 2.8. The visual quality took up 70% of the importance in perceived quality. Geometric variation involves gaps, flushes, and reflection alignments, and how these are per- ceived in interior and exterior parts. One important condition is "split lines", which are sometimes called gaps, which is commonly defined as the "relation between two mating parts over a specified distance". In the eyes of customers, such gaps can cause dissatisfaction. Geometry assurance is necessary in order to provide parts to assembly that has the required functionality, but also to look the intended way visually. The result of this is that a comprehensive geometry assurance process is needed in order to eliminate risks of failures, accidents, and customer dissatisfaction. Figure 2.8: Perceived Quality Framework (Stylidis et al., 2019). 2.4 Geometry Assurance Tools Even though the importance of digitalization has been underscored by several au- thors in recent years, it seems like most companies are yet to implement it in their geometry assurance processes (Lindau et al., 2020). For Body In White (BIW) sheet metal assemblies, physical verification of geometries is still the dominant method used in pre-series built to verify the outcome, before the start of the production phase. BIW is the stage in vehicle manufacturing in which the car’s body frame has been joined together, before painting, trimming, and final assembly. The tech- nologies and tools needed to achieve digital verifying methods are known, due to research in the field. There are also examples of companies who are in an early phase of implementing such tools on an production scale. This section will explain 17 2. Literature Study how a digitalizaiton of the geometry assurance process could be executed by using a range of different tools. 2.4.1 Measurement Technology As mentioned in earlier sections, there are never two parts of the same kind that are completely identical. Companies therefore have to regularly measure the dimensions of their parts to ensure they are within tolerances. There are various ways of doing so but the most common today is using Coordinate Measuring Machines (CMM). 2.4.1.1 Coordinate Measuring Machine CMM is widely used in industry for precise dimensional measuring, especially in mass production and automotive manufacturing (Ali, 2010). CMM measure object geometry by using Measure Points (MP) (Hocken, Pereira, et al., 2012). The MP are then stored in measurement file, which can be of different types such as CM4D, CSV, and DMIS. The two most common types of measurument tools are probes and articulated arms. These machines can be manual or automatic. Probe Probing can be done with either handheld devices or bridge-type CMM. Handheld probing is performed by an operator with a probe-mounted stylus (Ali, 2010). There are different probes of this type but the most common are touch-sensitive and react to any type of contact with another surface. By touching specific points of the object surface, discrete points are extracted resulting in a mapping of 3D space. Bridge- type-CMM, see Figure 2.9, are 3-axis machines that move in X, Y and Z direction in addition to the probe being able to rotate 360 degrees (Artkin, 2023). To create dimensional measurement, the obtained points are used to create geometries. To efficiently measure parts, the part measured need to be positioned in relation to the main axis. Similarly to the handheld stylus type probing, bridge-type CMM acquire its coordinates by touch-trigger of the probe. Moreover, if a CAD model of the part being measured is available, the scanning can be done automatically. 18 2. Literature Study Figure 2.9: 3-axis probe-mounted bridge-type CMM (Wikimedia, 2009). Articulated Arm In recent years, articulated arm CMM has become more popular. This technology can be used in combination with both laser scanners and probes. By mounting a laser scanning device on the end of a robot arm with several joints, part surfaces can be efficiently scanned (Artkin, 2023). The arm is manually maneuvered by an operator. Because of its flexibility, which is similar to a human arm, the scanner can efficiently examine parts which are inaccessible to traditional CMM methods. Additionally, this technology captures several thousand points at a time, making it preferred in SPC applications because it provides comprehensive data for detecting variations and defects more accurately than traditional methods. It works by transmitting a laser beam onto a surface which is then reflected back to the scanner. The scanning location is then measured as the separation between the object and the scanner. The most widely used scanners are interferometers and Absolute Distance Meters (ADM) in which Spherically Mounted Retroreflector (SMR) are the most prevalent. 19 2. Literature Study Figure 2.10: Scanner-mounted articulated arm CMM. 2.5 Software for Digital Verification This section describes current and possible future software and technologies that can be used for digital verification of geometry assurance. As this thesis is limited to testing in RD&T, this will be the only digital software environment described. 2.5.1 RD&T RD&T is a CAT software tool developed to simulate manufacturing and assem- bly variations before production begins (RD&T, 2025). The software has been used over 20 years within the automotive industry, working with original equipment man- ufacturers, suppliers and consultants. It works by applying statistical variation in tolerances on parts or assemblies which is then visualized. By using RD&T, several different concepts and assemblies can be quickly analyzed and compared, greatly reducing time needed for the design process. The main task of RD&T is ensuring robust designs that have a minimal variation when manufactured and are mainly used for non-rigid analysis. There are 3 basic functionalities in RD&T, namely vari- ation analysis, contribution analysis, stability analysis, and engineering document generation. To create model in RD&T which can be simulated, certain steps need to be per- formed. One of the most crucial steps to acquire the correct geometrical impact on the part in question is creating appropriate locators. Locators (or fixture loca- tors) are reference points or physical features used to precisely position and secure a part during assembly, testing, or inspection. The locators are then used to create 20 2. Literature Study a position system for the part. Positioning systems are digital representations of real-world alignment, used to simulate and validate product design, manufacturing processes, and testing procedures. This is used to lock the position of the part or sub-assembly in space to be able to perform different analyses. Non-Rigid Variation Simulation Parts that are produced for the automotive industry is more often than not, sup- posed to be flexible. Another word for parts that are flexible are non-rigid parts (Lindau et al., 2015). Non-rigid analysis of parts and assemblies aims to simulate parts according to their flexibility and bounce-back geometry. Using tools such as the variation analysis in RD&T, a more realistic result can be achieved for how the BIW parts will act (Wärmefjord, Söderberg, Lindau, et al., 2016). To achieve this analysis, Monte Carlo simulation is used. The simulation works by first utilizing a Finite Element Analysis (FEA) solver to create a initial mesh model of the non-rigid model. By studying the nodes in this mesh, displacement is determined. This in combination with the Method of Influence (MIC) makes it possible to find a linear relationship between part deviation and spring-back deviations allowing for thou- sands of simulations to be simulated quickly. The end result of a variation analysis presents the amount of parts which will be within tolerances in a normal distributed curve in regards to the upper and lower design limits. It also shows the deviation of the assembly or part as seen in Figure 2.11. Figure 2.11: Variation analysis of part (RD&T, 2025). 21 2. Literature Study Contact Modeling In RD&T, contact modeling plays a vital role in ensuring the validity of assembly simulations, particularly when simulating interactions between flexible components. Contact analysis in RD&T involves identifying surface or edge interactions between adjacent parts and determining areas where physical interference or excessive clear- ance may occur (Lindau et al., 2015). The software supports automated contact detection through a specified range of values and tolerances, allowing for detailed visualization of surface interactions. Thus, RD&T facilitates the creation of critical gap and flush measures at contact points, providing a foundational understanding of how parts fit together within an assembly environment. Material Temperature Temperature-dependent material properties are a critical consideration in variation simulations involving dissimilar materials, particularly in assemblies that include plastic and metal components. Wärmefjord, Söderberg, and Lindkvist, 2016 high- lights that parameters such as Young’s modulus and thermal expansion coefficients must be defined for each part to ensure accurate results. These properties directly influence the stress distribution and holding forces in the final assembly. For exam- ple, plastic components typically experience greater thermal expansion than metals, which can lead to stress and geometric deformation. Moreover, Young’s modulus itself is temperature dependent plastic becomes more compliant at higher tempera- tures and stiffer at lower ones. As shown in their case study, increasing the temper- ature results in decreased holding forces due to material softening, whereas lower temperatures led to increased forces. These findings underscore the importance of integrating thermal effects into RD&T simulations when evaluating clip fasteners or similar joining methods. Accurate simulation of these interactions supports robust design decisions. Stability Analysis The stability analysis function in RD&T allow the user to evaluate the geometrical robustness of an assembly or part design (RD&T, 2025). Using this analysis, the positioning of locators can be automatically or manually optimized, ensuring an optimal geometrical sensitivity. Stability analysis allows for early design changes, making later design faults less likely. The sensitivity is displayed using color coding of the parts visualized in Figure 2.12. 22 2. Literature Study Figure 2.12: Stability analysis of a part (RD&T, 2025). Contribution Analysis From the variation and stability analysis general knowledge of how much the parts deviate can be acquired as previously shown. However, they do not display which tolerances play the most vital roles in the design. To get an understanding of each tolerances impact, the contribution analysis can be used (RD&T, 2025). The contri- bution analysis ranks the impact of tolerances in critical dimensions in percentage. Using this analysis, each tolerance impact on the design can be changed accordingly, increasing or decreasing design limitations. See Figure 2.13. Figure 2.13: Contribution of tolerances in critical measures (RD&T, 2025). 23 2. Literature Study 2.5.2 Virtual Fixture Virtual fixture refers to the process of creating virtual fixtures for part or assemblies in digital softwares like RD&T. This method makes it possible to analyze the parts free state shape with regards to spring-back effect (Lindau et al., 2020). This is not possible in traditional testing as the part usually is over-constrained by several more clamps than its amount of degrees of freedom, effectively bending the part towards its nominal state. Additionally, the traditional way of clamping does not take into account the forces applied from the clamps resulting in a loss of spring-back data after release. Performing a realistic simulation of part contribution to variation in assembly becomes much more difficult because of this. Using virtual fixtures and having an understanding of a part’s free state is therefore essential when analyzing non-rigid digital assembly models. An article (Lindau et al., 2020) thoroughly de- scribes the steps necessary to create a realistic digital fixture representation, these being the following: 3D-Scanning of Parts One of the most prevalent steps is creating a efficient and accessible, ergonomically sustainable 3D-scanning station for the operator that performs the scanning. To es- tablish a standardized way of 3D-scanning the following steps should be conducted. 1. Three adjustable sphere supports rigged on a horizontal measurement base plate so so that part can be positioned and rest close to parallel with clearance from sharp radius’s. 2. Position part on the supports. 3. Scan part to acquire resting geometry. 4. Remove part to scan the resting points coordinates as well as small portions of base plate. This procedure will produce three resulting point-clouds. A point-cloud of the part geometry which describes shape, a point-cloud that describes position of the sphere supports, and a point-cloud which describes the plane perpendicular to the gravity field. 2.5.3 Meshing To accurately simulate the physical behavior of components in digital environments, meshing is essential. This process involves making a CAD model into a network of interconnected nodes and elements, forming the foundation for FEA. Each node solves equations in relation to each other, predicting deformations, stresses, and dis- placements under applied loads, such as those induced during assembly or welding. Different mesh types are used depending on the application: 1. volume meshes (e.g., tetrahedral or hexahedral elements): model thick or solid components and are ideal for plastic components. 24 2. Literature Study 2. surface meshes: thin shell or mid-surface meshes are ideal for sheet metal or lightweight structures. The quality of meshes is crucial in realizing an accurate result. Automated meshing tools like ABAQUS, HyperMesh, and ANSYS Meshing can generate a initial model but manual refinement is often required for intricate features to ensure precision. It is currently not possible to create meshes in RD&T. The meshing process used during the case study later in this project is explained in Appendix F. 25 3 Methodology In this chapter the methodology for the conducted project is clarified. The method- ology design had an mixed method, which started with a literature study to un- derstand the theory of geometry assurance and important topics within this field. In parallel, a comprehensive data collection was coordinated, with qualitative and quantitative interviews that were performed to get practical knowledge about work- ers’ opinions about possibilities regarding the digitalization of the geometry assur- ance process. This step was also important to get direct information about the current nominal buck verification process, as the workers at Polestar’s manufac- turing plant in Chongqing, China, were interviewed. The questions posed in the qualitative and quantitative interviews can be found in Appendix A, Appendix B, Appendix C, and Appendix D. Once enough information had been collected about geometry assurance and the current nominal buck verification process at Polestar, a case study was performed and documented. This process was described as a step- by-step guide in RD&T for virtually verifiying part geometries, which can be used by Polestar workers. The scanning process and data processing in PolyWorks, as well as the process of creating meshes, are explained in Appendix E and Appendix F. A comparative analysis was also performed where the digital solution was com- pared to the current physical process, which gave direct hindsight about the possible outcomes of implementation. 3.1 Methodology Design The methodology had a mixed design including literature study, interviews, a de- scription of the case study and proposed method, and a comparative analysis. The project therefore had a theoretical approach initially, but during the case study there were many practical elements, such as when the parts were 3D scanned. 3.2 Data Collection The data collection was based on a literature study and interviews, with both qual- itative and quantitative elements. 3.2.1 Literature Study In the study of the literature, all the information extracted regarding the geomet- ric assurances in manufacturing formed the knowledge base for the project. This 26 3. Methodology included both theoretical frameworks and definitions within this field, as well as spe- cific technology that can be used to measure part geometries. This was necessary to understand related technologies and concepts, which opened up the possibility to develop a case study. The literature was drawn from websites such as Web of Science, Scopus, Taylor & Francis journals, etc. using keywords such as "digitaliza- tion", "geometry", "assurance", "virtual", "tools", "nominal", "buck", among others, and combining these in convenient ways in the search string to find the appropriate material. 3.2.2 RD&T PE Geometry provides courses for training personnel in software such as RD&T, both internally as well as externally to outside organizations. Their former educa- tion programs was therefore used to familiarize the project group with RD&T to understand the various modules, which was necessary before the case study. 3.2.3 Interviews The interviews contained two parts - qualitative and quantitative interviews. The qualitative interviews were held with industry personnel to gather information re- garding certain topics of interest, which will be described below. The quanitatative interviews were made to gather a larger amount of data of people’s opinions regard- ing possible virtualizations within the field of geometry assurance. These questions were posed to geometry assurance engineers & specialists, which also is described below. 3.2.3.1 Qualitative Interviews The qualitative interviews were performed by interviewing a few elected employees from either PE Geometry, Polestar, or an outside organization. In total, eight peo- ple were interviewed for the qualitative part. The selection of the interviewees was based on their knowledge regarding the technological, economic, and sustainable aspects of the current nominal buck verification process, as well as knowledge about 3D scanning methods, variation simulations & other digital tools, or general knowl- edge about the geometry assurance process. The purpose of these interviews were to gather this information directly from industry personnel, which could be useful in the analyses conducted during this project. The results from these interviews were described within different areas of importance, which can be seen in Chapter 4.1. The interviews had a structured nature, with questions specifically constructed for each interview. This is because the interviewees had different roles and specialized knowledge within these fields. PE Geometry, Gothenburg The employees at PE Geometry, Gothenburg are experts in the field of geometry assurance and 3D scanning processes, and their knowledge was necessary to extract in order to get hands-on information about current processes. Some of their em- ployees are consultants for Polestar and were also supervisors for this project. Two 27 3. Methodology people from PE Geometry were interviewed in the qualitative interviews. One being a Geometry Assurance Engineer, consultant for Volvo Trucks, but who was knowl- edgeable about 3D scanning technology in general. Therefore he was questioned regarding the 3D scanning process, its efficiency, its possibilities to implement the technology in mass production processes, and so on. This information was necessary as 3D scanning of parts was believed to be an important step in a digitalization of the nominal buck verification process. The second person is an employee at PE Geometry as well as consultant for Polestar, with the role as Geometry Assurance Expert, and he is knowledgeable about current geometric analyzing methods and the nominal buck process in Chongqing, China. Questions were asked regarding communication, file handling, different uses of softwares, etc., between the different Polestar departments in China, the United Kingdom, and Sweden. The questions posed to PE Geometry can be found in Appendix A. Polestar, Chongqing In order to get insight into the nominal buck verification process, three employees at Polestar’s manufacturing plant in Chongqing, China were interviewed. The first one is the Team Leader in geometry assurance and Audit at the plant in Chongqing. He is an expert in the overall verification process, and it was important to extract information regarding his opinions on a possible digitalization of the current process. The second interview was held with the Planner & Leader for nominal buck activi- ties at the plant. He knows information about exact lead times, possible disruptions, advantages and disadvantages of current process. The third interview was held with the Specifier of Manufacturing Requirements for nominal buck & Supplier Contact Person. Questions were asked regarding the costs of manufacturing a nominal buck, manufacturing and delivery times of such a buck, and important aspects connected to this. The questions posed to Polestar can be found in appendix B. Other people with expertise Two interviews were held with people are not working at either PE Geometry or Polestar, but who have knowledge about quality verification processes, as well as possible digitalizations. The first interview was held with a Senior Analysis Engineer at Volvo Cars in Gothenburg. They have been part of similar projects in the past, with the objective of digitalizing the geometry verification process at Volvo Cars. This has included scanning physical parts, transferring these geometries to 3D data, and then using that data to perform variation analysis using virtual tools such as RD&T. Therefore this interview was crucial for understanding the possibilities within this field. The second interview was held with two persons - a Deputy Head of Department and Professor at Chalmers University of Technology, as well as a Docent in Product Development at Chalmers University of Technology. Both are conducting their work within the Department of Industrial and Materials. One of them is also the examinator and supervisor for this project. The questions posed to these three people can be found in appendix C. 28 3. Methodology 3.2.3.2 Quantitative Interview Quantitative research utilizes two tools - experiments or surveys (Smith, 2024). Therefore a quantitative question form was created and sent out to a larger group of people. It was necessary that the respondents either had worked with geometry assurance or had the required knowledge to answer the questions. The question form was used to get people’s opinions about geometry assurance and the possibili- ties of digitalization within this field. The project workers made assumptions about important aspects within the field of geometry assurance, and structured the ques- tions accordingly. Questions were either multiple choice, one choice, or open-ended. The form was sent out to people having roles as Geometry Assurance Engineers, Geometry Assurance Specialists, or similar on Linkedin, where their job titles could be acknowledged directly. The questions posed in this question form can be found in appendix D. 3.3 Case Study Once the literature study and interviews had been performed, a specific digital method was developed further and demonstrated in a case study. This case study was continuously constructed during the project, and much of the information re- garding certain steps was brought from PE Geometry’s internal courses regarding the work procedure in RD&T. Simultaneously, discussions were conducted with in- dustry personnel at PE Geometry to understand the crucial steps in reaching a final solution which has not been described in this report. The first step of a digital geometry assurance process is transforming the physical part to a digital 3D model, which is done with help of scanning equipment. The technology of 3D scanning is explained in chapter 2.4.1. There exist different scanning methods for doing this transformation. Once the part’s geometry has been transformed to a digital file, it can be implemented in the RD&T software and used for simulation purposes. Meshes of CAD data also needed to be created for all parts involved in the final assembly. The case focused on non-rigid variation analysis of a sub-assembly and showed the steps necessary to compare nominal parts to actual produced parts. This was done to create a standardized way to perform virtual geometry assurance and showcase the advantages of using software like RD&T. The area chosen to be simulated in the case study was the B-pillar with belonging components, specifically the intermediate and lower parts of the left-hand interior design. The parts in question were of plastic material and were chosen because of their critical position in the assembly. This area is composed of several meeting points of different parts, making it a prime investigation target. Consequently, proving that simulation is possible in this area entails it is most likely possible to simulate any section of the car. 29 3. Methodology 3.3.1 Step-by-Step Guide for RD&T, Meshing, PolyWorks, and Scanning Process In parallel with performing the case study, the process was documented and ex- plained in the report as a step-by-step guide. The idea is that the guide can be used by Polestar workers to digitally verify parts using RD&T. The main objective of this guide was therefore to explain all necessary steps in RD&T. As PolyWorks must be used to gather and process the scan data, a step-by-step guide of this step was explained in a similar manner in Appendix E, which also contains a description of the practical scanning procedure. Likewise, as meshing was a necessary step to reach the final results, which was not performed in RD&T, the process for this was explained in Appendix F. The meshes were created in the HyperMesh software. 3.4 Comparative Analysis - Physical and Digital Verification Methods To gain an understanding of the aspects that both the physical and a possible virtual method would incur, a comparative analysis was constructed. The current physical process was compared to the alternative virtual approach and advantages and disadvantages was analyzed. This was done by analyzing important areas such as efficiency, accuracy, cost, sustainability, and flexibility, to showcase directly what advantages and disadvantages Polestar will face when implementing virtual tools in their geometry assurance process. Furthermore, the implementation feasibility includes a completely new virtual work- flow process based on the conclusions of the results. 3.5 Impact of Virtual Implementation Another section focuses on describing what the direct effects for Polestar if they decide to implement the proposed digital method. Data for this analysis is based on information from the industry interviews and case study, as well as internal information from Polestar. This discussion also includes a completely new virtual workflow process based on the conclusions of the results, where some areas of the previous workflow could be completely eliminated. A subchapter will also entail the implementation feasibility of the digital method for an organization such as Polestar, as well as the possibilities to automate many steps of the digital method to make it fit in a large-scale use. 30 4 Result This chapter presents the results obtained from the key stages of the methodology. These include findings from both the qualitative and quantitative interviews, the de- velopment of a step-by-step case study guide, a comparative analysis of the physical and virtual nominal buck, and an evaluation of the impact of virtual implementation. 4.1 Findings from Industry Interviews Interviews with key personnel at the factory in Chongqing, China as well as an Polestar Geometry Assurance Expert include: • Team Leader and Audit Expert in Chongqing, China (Appendix B) • Planner and Nominal Buck Leader in Chongqing, China (Appendix B) • Specifier and Supplier Contact in Chongqing, China (Appendix B) • Geometry Assurance Expert & Geometry Program Manager in Gothenburg (Appendix A) These interviews provided valuable insights into the physical nominal buck process as a whole and information regarding communication & collaboration between the different international departments at Polestar. 4.1.1 Current Physical Nominal Buck Process In the initial stage, the factory team, consisting of 30-40 R&D engineers, 5 geometry engineers, and 4 measuring technicians begin by freezing the nominal buck design data. This means that no further modifications are allowed at this stage. The team then works together with the supplier to review the nominal buck design. This stage is vital to ensure the design can be manufactured, easily assembled and disassem- bled efficiently. From first data freeze to ready-to-manufacture model usually takes about three months. Once the ready-to-manufacture model has been obtained the suppliers can start the manufacturing. To review the suppliers process and validate results, the whole fac- tory team visit the supplier. If something is not up to standard at this point, this is notified and the supplier will need to change their process. Finally, when all details of the nominal buck is satisfactory, a final acceptance check is performed before the nominal buck is shipped to the plant. The complete nominal buck creation process 31 4. Result can be seen in Figure 4.1. Figure 4.1: Nominal buck creation process. At the factory, the nominal buck serves as a reference structure for testing and eval- uating the fit of both exterior and interior components. To support this, a variety of physical fixtures are utilized. These fixtures are custom-designed and manufactured alongside the buck to mimic the attachment interfaces found in the actual vehicle assembly. They consist of rigid metallic brackets, support frames, and locator pins that hold parts in place in a repeatable, predefined orientation. Fixtures play a critical role in ensuring consistency during measurement and visual evaluation. They enable engineers to mount components with the same geometry and constraints as in the real car. This allows for precise verification of gap and flush conditions, alignment, and fitment. The use of fixtures also helps minimize part movement during probing and 3D scanning operations, ensuring accurate and repeatable results. In addition to their use during part fitment, fixtures often in- clude built-in reference points that interface with Coordinate Measuring Machines (CMMs), simplifying the process of data acquisition. When small design modifica- tions are made, these fixtures can be re-machined to accommodate updated geome- tries. When the nominal buck has arrived at the factory, testing and measuring can be- gin. As mentioned in earlier chapters, the nominal buck is used to check the fit of exterior and interior parts. Measurements here are mostly done using an probe mounted CMM but 3D-scanning is also possible. If there is a part with irregular 32 4. Result shape, special gauges are used to measure gap and flush measures between parts. These measurements are then documented in measurement reports. During the measurement and testing process, the sample size of each batch is different depend- ing on earlier experiences. For example, if part has been outside tolerances earlier, 25-50 parts are measured to ensure validity of quality. On the other hand, if the part has been without issue earlier, only around 10 parts need to be measured. For each build phase of the car, the measurement process is performed two times to ensure accuracy. If there are deviations from the specifications, these are documented in a software called PIRA. If enough parts are out of specification, the batch is sent back to the supplier. The supplier is then notified by the team what is wrong with the batch and required to remanufacture the batch to create a better standard. If the problem is not with the suppliers product but with how parts fit together in the assembly, design changes of the nominal buck can sometimes be necessary. Modifications are usually divided into minor and major changes. Minor changes are problems which can be fixed on site in a couple of days, for example, grinding or polishing. Major changes involve needing to change the very geometry of the nominal buck itself. This entails that the nomi- nal buck will need to be sent back to the manufacturer where it is re-manufactured. These changes can take anywhere from one to three months. To visualize the part verification process in the nominal buck, a flowchart is displayed in Figure 4.2 below. Figure 4.2: Physical part verification process in nominal buck. Challenges of Physical Verification Process Along with the physical process several challenges, both logistical or geometrical, arise. This chapter aims to pinpoint mentioned challenges and explain their origin. Measurements of details in the nominal buck is a time-consuming process. First of all, engineers need to measure parts, then the engineers need to assemble the aprts 33 4. Result on the nominal buck. After this step, they need to disassemble the parts. This pro- cess needs to be done for all the interior and exterior parts of the car. Furthermore, as mentioned in the earlier chapter, this is done anywhere from 10-50 times per part, two times per build phase, doubling the time spent for the activity. This can pose significant problems for the project time-frame if not conducted efficiently. Design changes are required based on the robustness of individual parts and the nominal buck as a whole. Typically, several adjustments are needed. These modifi- cations take place after data is frozen and the design is re-evaluated. Chapter 4.1.1 discusses minor changes, which can be made onsite within a few days, and major changes, which involve shipping the nominal buck back to the supplier, a process that can take anywhere from one to three months. When working on a tight sched- ule, as is the case with automotive manufacturing, these design changes result in delayed development times. In the initial production stage, known as PS, suppliers submit measurement reports to the factory team in China. However, these reports do not guarantee that the parts are within tolerances. As a result, the team cannot be certain whether a part will properly fit the nominal buck. The final judgment is made visually by the team once the part is placed on the nominal buck, relying on prior experience. Further- more, there is no virtual verification system to check parts before shipment. These two factors significantly increase the risk of wasted time and resources. While the cost of a nominal buck is not a major problem for a company like Polestar, it can still impact the company negatively. Currently, the nominal buck cost varies depending on supplier and what is asked of the supplier. For example, in China, a nominal buck for an SUV that includes 100% of the exterior and 20% of the interior parts costs around 4 million RMB, which is roughly 5.4 million SEK. Seen in Figure 4.3 is the budget for Polestar 5 Nominal buck and fixtures as well as actual money spent on them accordingly. Note that the prices are in units of 10 000 RMB. The money spent on the nominal buck and fixtures total 3,51 million RMB, or 4.78 million SEK. Note this is only for money spent on manufacturing the fixtures and nominal buck, not total money spent on development. Figure 4.3: Budget for the nominal buck. Costs are in units of 10 000 RMB. 351 = 3 510 000 RMB. Another challenge worth noting are the ergonomical difficulties. When workers have to fit parts onto the nominal buck, certain motions like bending and reaching can be stressful for the body. Even though this is not a major problem in the physical process, it is something that would be better of eliminated. 34 4. Result Communication Regarding Nominal Buck Communication regarding geometry assurance processes is transferred between the R&D department in the United Kingdom, the Manufacturing Engineering (ME) de- partment in Gothenburg, as well as the manufacturing plant in Chongqing, China. It is of importance to ensure that design requirements are properly understood among the teams, and evaluated and implemented within manufacturing while filtering and prioritizing relevant changes to minimize disruptions. Communication with the Chinese plant is structured around project-related meet- ings, with tasks varying depending on the project’s stage. Regular weekly meetings allow for information sharing, while direct communication between R&D and the production team is encouraged to reduce unnecessary intermediaries. Over time, a strong level of trust has been established between departments, allowing for more streamlined interactions. A similar approach is taken with the UK-based teams, where collaboration is frequent but not bound by fixed meeting schedules. Instead, tasks dictate interactions, and the flexibility in communication has helped foster strong relationships. Challenges of Communication Information transfer regarding design changes are often efficient and smooth, but can be challenging, particularly if not communicated at the right time. Small adjust- ments, such as hole repositioning, are relatively easy to implement, whereas larger modifications affecting surfaces or pressing tools require significantly more effort and time. Ensuring timely information flow is therefore critical to minimizing delays in production adjustments. One notable challenge in global communication is language. Despite a strong com- mand of English within teams, differences in terminology can lead to misunder- standings, even among native Swedish speakers. These challenges are amplified when communicating across Swedish, English, and Chinese teams. Additionally, time zone differences restrict available meeting times, limiting the overlap for direct collaboration with the Chinese plant to a few hours per day. From a software perspective, the daily workflow involves a variety of tools, including PowerPoint, Teams, and TCVis. While RD&T is used to generate key measurement data, the Chinese team integrates this information into their own applications, such as the Final Demand App, which provides structured measurement instructions. Data transformation between RD&T and other formats is crucial in this process, but conversion errors have occasionally led to misinterpretations—for instance, a flush condition that should be -4 appearing as +4 after conversion. The interview also highlighted the importance of clearly defining roles and respon- sibilities. Differences in organizational structures between Polestar and its former parent company, Volvo Cars, have created challenges in aligning workflows. Histor- ically, teams have relied on legacy processes from Volvo, but variations in company structures require clarification of roles to avoid inefficiencies and duplicated efforts. 35 4. Result Currently, steps are being taken to establish clearer role definitions to enhance op- erational efficiency. 4.1.2 3D Scanning Technology Information regarding 3D scanning technology was extracted from the following interviews: • Geometry Assurance Engineer at PE Geometry, Consultant for Volvo Trucks (Appendix A) One of the key advantages of 3D scanning in geometry assurance is its ability to collect large amounts of measurement data, enabling comprehensive analysis and historical comparisons. Unlike traditional methods, 3D scanning is non-contact, meaning it does not interfere with the setup, making it possible to capture measure- ments step-by-step during assembly. Additionally, scanned parts can be virtually assembled & disassembled in simulation software, allowing engineers to predict po- tential issues before physical production begins. Some factories have also started implementing automation, where real-time scanning is used to align components such as windshields during the assembly process. The large volume of data generated from frequent scans can be difficult to manage and store efficiently. Scanning software also applies filtering algorithms that smooth surfaces, potentially masking minor but critical deviations. Accuracy can be affected by extreme angles or the positioning of the scanning arm, leading to misalignment in measurement results. Additionally, ensuring trust in the system is essential, as understanding how the software processes and filters data is crucial for achieving reliable results. Some conclusions can be drawn regarding advantages of the 3D scanning technology in geometry assurance settings. These include: • Increased data collection: Allows for comprehensive analysis and historical comparisons. • Non-contact measurement: Enables scanning without disturbing the setup, allowing for step-by-step analysis of assembled components. • Virtual assembly testing: Scanned parts can be virtually assembled to predict outcomes before physical production. • Automation potential: Some factories already use automated scanning, such as aligning windshields based on real-time scan data. There are also some challenges & limitations that can be associated with 3D scanning technologies. These include: • Data volume: Large datasets from frequent scans can be difficult to manage and store. 36 4. Result • Software filtering: Scanning software smooths surfaces, potentially consid- ering deviations as noise and therefore inaccurately filtering them out. • Accuracy concerns: Misalignment can occur due to extreme angles or posi- tioning of the scanning arm. • System trust: Understanding how the software processes and filters data is crucial for reliable results. 4.1.3 Virtual Initiatives at External Organizations One interview yielded information regarding past virtual initiatives at external com- panies, organizations, and projects. The information was valuable to get a more comprehensive understanding of the possibilities and limitations of such technolo- gies in manufacturing settings. The following interviews were used for this purpose: • Senior Analysis Engineer at Volvo Cars (Appendix C) Over the past years, significant advancements have been made in virtualizing ge- ometry assurance processes at Volvo Cars, particularly in virtual simulations. The primary focus has been on compliant simulations rather than traditional tolerance chains. However, due to recent company restructuring, the prioritization of virtual geometry assurance has decreased, leading to a reduction in dedicated personnel and a redirection of expertise to other areas deemed more critical to the organization’s overall strategy. While virtualization efforts continue, the full potential of these technologies has yet to be fully realized. For variation simulations, Volvo Cars primarily uses RD&T, complemented by other commercial software for specific structural simulations. The idea of virtual fixtures, described in chapter 2.5.2, was developed by the interviewee and their colleague at Volvo Cars. Virtual fixtures are not yet used in real manufacturing settings at Volvo Cars, but it is a long-term goal for the company to implement the technology. Additionally, 3D scanning plays a crucial role in virtual verification, with handheld arm scanners used to capture free state (a concept which is explained in chapter 2.5.2) scans of components. These scans are typically performed by suppliers, who provide STL files that are integrated into simulation models. The virtual verifi- cation process is closely aligned with the vehicle development timeline, beginning in the pre-series and prototyping phase. Before physical assembly starts, suppliers scan parts and submit the data to Volvo Cars, which then can use the data in their simulation software. Challenges of Virtual Implementations Across the industry, a common challenge is the lack of management prioritization for virtual verification, preventing it from becoming a standardized practice. There are cases when companies have made significant progress in virtual assurance by developing tools, methodologies, and supplier requirements to improve geometry as- surance. However, when this area is no longer seen as a strategic priority, resources 37 4. Result are redirected, often resulting in limited personnel to drive virtual advancements. This deprioritization slows technological progress and may explain why virtual ge- ometry assurance has not yet been fully implemented as a standard practice in many organizations. Cultural and technological resistance exists within production teams, where tradi- tional physical verification methods are more deeply ingrained. While R&D de- partments frequently rely on simulations for structural validation, manufacturing teams have been slower to integrate virtual simulations into their decision-making processes. A major challenge in virtual geometry assurance is the inconsistency of scan data provided by suppliers. The quality of these scans varies, often due to differences in measurement equipment, scanning techniques, or calibration standards. This incon- sistency affects the reliability of the data used for simulations, leading to discrep- ancies between virtual models and physical parts. A specific issue is the alignment of holes and slots, which are critical for assembly accuracy. If these features are misaligned in the scan data, simulations may produce misleading results, impacting decision-making and problem-solving. Without standardized scanning procedures across suppliers, additional verification steps are required, increasing workload and reducing efficiency. 4.1.4 Research Within Virtual Geometry Assurance Interviews with academic researchers, mentioned below, specializing in virtual ge- ometry assurance provided valuable insight into current and future aspects of the area: • Deputy Head of Industrial and Materials Science Department, Chalmers Uni- versity of Technology (Appendix C) • Associate Professor of Industrial and Materials Science, Chalmers University of Technology (Appendix C) Both interviewees are experts in virtual geometry assurance with decades of ex- perience in academia and industry (including Volvo Cars and Saab Automotive). Their work has focused on increasing accuracy and efficiency of simulations to re- duce reliance on physical prototypes. Currently their research focus on removing limitations of simulation tools such as: • Creating accurate simulation tools for simulating factors like continuous weld- ing deformations. • Reducing computational time through optimization techniques. There are many advantages of using virtual geometry assurance within automotive manufacturing. Some of the key mentions include: 38 4. Result Advantages of Virtual Geometry Assurance One of the biggest benefits of virtual geometry assurance is its ability to identify and resolve design issues in early design stages such as the concept, prototyping or pre-production phases. By simulating car models virtually before actual production is started, the time needed for part verification can be significantly decreased. Fur- thermore, multiple design iterations of a design can be quickly evaluated without the need for manufacturing, saving time, money and manpower. Another great aspect is that it enables closer collaboration between teams. Providing a virtual platform where the design is shared and can be modified in real time by design, manufactur- ing and quality assurance teams at the same time makes the process easier and faster. Virtual geometry assurance enables companies to significantly cut costs. Since parts are tested in a virtual environment, there is no need to manufacture the parts for testing. However, physical parts will still be needed toward the end of the verifi- cation process to visualize and test the final assembly of the car. Until that point, parts to test requirements and visual aesthetics will not be needed, significantly reducing cost of the process. Additionally, this ties into the sustainability of the company as well. By limiting material usage the waste produced in each production phase is decreased, making the process more sustainable as a whole. Since virtual geometry assurance allows for thousands of variation simulations to be run simultaneously, the amount of measurements can be greatly increased. This, in combination with real-world scanned data of manufactured parts can effectively pre- dict deviations from the tolerances set. These factors save both time and manpower. Challenges of Virtual Geometry Assurance One of the biggest challenges of virtual geometry assurance is the trade-off be- tween simulation accuracy and computational efficiency. Including more physical phenomenons such as welding deformation, heat effect or spring-back information improves accuracy but makes simulations more time-consuming. As such, approxi- mate methods to calculate head effect is often used, resulting in a loss of accuracy. Also, when analyzing models with more details, finer meshes are required, increasing memory usage and computation time. When conducting non-rigid simulations of parts, scans of the parts are required. This is done to acquire the free-form state of the part to get a as realistic simulation as possible in regard to spring back effect, mentioned earlier in chapter 2.5.2. The point cloud data extracted from scanning is extremely large and contain millions of data points. If this data is not optimized, performing simulations with multiple parts can overwhelm computing power. Because of this, engineers responsible need to filter and process the data so that only necessary geometric features are simulated. Also, similarly to 4.1.3, organizational resistance to change is one of the main ob- stacles to implementing simulation tools like RD&T. Many organizations face skep- ticism and reluctance when adopting new technologies due to comfort with existing practices, fear of disruption, and perceived complexity. Employees may resist due 39 4. Result to a lack of understanding or fear of failure, while management might be concerned about the potential challenges of integrating new tools into current systems and related costs. Overcoming this resistance requires clear communication of the benefits, involvement of key stakeholders early on, and tailored training programs. Leaders should foster a culture of continuous improvement and provide ongoing support to ensure smooth integration and build confidence in the new tool. By addressing these challenges, organizations can successfully implement RD&T and unlock significant benefits in efficiency and decision-making. 4.2 Quantitative Interview Findings The quantitative question form was sent out to 53 individuals and a total of 15 individuals answered the form. The form was specifically sent out to people working with geometry assurance, thus each individual is a specialist on the subject and has their own views and opinions about the virtualization possiblities within this field. Another word for geometry assurance is geometry verification, and as the term geometry assurance is quite uncommon in other industries than automotive, many questions are posed with the term geometry verification. However these two concepts are the same and can therefore be used interchangeably. As can be seen in Figure 4.4, the participants came from different sectors of work which gave valuable insight within these industries. The most common sector of work was automotive, which is expected as the subject of geometry assurance is especially common and important within this industry. Figure 4.4: The circle shows that the interviewees came from various industries. Two other questions were posed regarding the participants familiarity with virtual geometry assurance processes, and whether their processes are currently conducted physically or virtually within their respective company. The results can be seen in Figure 4.5. As can be seen, 80% are using a mix of both physical and virtual 40 4. Result verification methods. Figure 4.5: How the correspondents perform geometry assurance and their familiarity with virtual tools. A question was also posed regarding specific software that the various interviewees are using when performing analyses and calculations. The list below demonstrate which software that the respondents use during geometry assurance processes: • RD&T • Creo View • 3DCS • LK Camio • Polyworks • ATS CM4D • Crystal Ball The most common software for geometry assurance seems to be RD&T based on the results. Another aspect of interest was the effectiveness of geometry assurance processes at the different companies. Therefore the subjects were able to rate their company’s geometry assurance process from 1 to 10, where 1 represents not effective and 10 represents very effective. The participants were also posed a question regard- ing how willing they were to virtualize their geometry verification process, from 1 to 10, where 1 represents not willing and 10 represents very willing. The results of these questions can be seen in Figure 4.6. As can be seen, some participants rate their cur- rent geometry verification process as quite mediocre in their effectiveness, while they rate highly on their willingness to virtualize the process. This shows that there is a common need for virtual technologies in geometry assurance across many industries. 41 4. Result Figure 4.6: The effectiveness of current geometry verification processes at each individ- ual’s co