Nicolas Leberruyer recently defended his Licentiate Thesis titled “Facilitating the Adoption of AI-driven Zero Defect Manufacturing in Production Systems.” His research delves into integrating AI into manufacturing to ensure zero defects in production, a subject driven by his desire to have a tangible impact on the final customers by meeting and exceeding their expectations.
In the realm of modern manufacturing, the quest for perfection and sustainability is ever-evolving, driven by innovative technologies and groundbreaking research. Nicolas Leberruyer, a recent licentiate graduate, stands at the forefront of this movement with his thesis on AI-driven Zero Defect Manufacturing (ZDM). His work not only promises to revolutionize quality management in production systems but also aims to leave a lasting impact on customer satisfaction and environmental sustainability. In an engaging discussion, Nicolas delves into his motivations, the challenges faced, and the transformative potential of AI in ensuring flawless production processes, offering invaluable insights into the future of manufacturing.
Leberruyer explains that traditional quality management approaches rely heavily on statistical analysis of individual measures to reduce process variations. In contrast, AI-driven Zero Defect Manufacturing (ZDM) employs real-time analysis of multiple variables from both the process and products to predict and ensure quality. This method adds an extra layer of quality assessment on top of existing quality methods.
AI-driven ZDM significantly enhances sustainability in production systems. By preventing defects, it minimizes scrap and rework, thus reducing waste. Additionally, it supports the introduction of new, sustainable processes by providing recommendations to maintain desired quality levels. Furthermore, it empowers workers by assisting them in evaluating the quality of complex products and offering solutions to address quality issues.
Leberruyer shares that among the four AI-driven ZDM applications developed, only one was successfully deployed in production. This application aids operators in assessing abnormal sounds from gearboxes during the final quality test, a crucial step in ensuring the reliability of the product.
His thesis includes a case study conducted in collaboration with a company that produces heavy-duty automotive transmission components. He highlights that abnormal noise from transmission components has been a persistent issue for automotive companies. With the rise of electric vehicles, there is an increased need to identify and mitigate the causes of such noises more effectively.
“Effective communication with all stakeholders is crucial; without it, even the most advanced AI model might remain unused, confined to a local computer”
During a two-year case study, Leberruyer took on multiple roles, including data scientist, project leader for AI-driven applications, mentor, and data analyst. This experience emphasized the importance of collaboration and communication for successful AI integration in production systems. Effective communication with stakeholders was found to be crucial, and collaborating with Volvo and Indtech proved highly beneficial for sharing experiences and gaining different perspectives from industrial PhD students.
“We hope this app will ease the operator’s workload”
A practical example of AI-driven ZDM’s benefits in production systems is the deployment of an application to assist operators in evaluating gearbox sounds during final quality tests. This application aims to reduce the stress on operators, who must manually judge each gearbox test within a minute, thus streamlining their workload and improving overall efficiency.
Leberruyer emphasizes the importance of early engagement in identifying data needs and understanding implementation limitations. He advises starting with available resources and improving iteratively based on needs. One common challenge is managing expectations, particularly in explaining that the solution’s performance is not yet perfect but will improve over time.
Leberruyer envisions significant long-term impacts of widespread AI-driven ZDM adoption, including cost savings, enhanced customer satisfaction, and more flexible product updates. He believes that AI-driven ZDM can serve as a safety net in production, ensuring that products not only meet quality requirements but also fulfill customer desires.
The future of AI-driven ZDM research involves developing standardized processes for integrating AI into production systems. Current approaches are highly customized, and there is a need for defining roles and responsibilities, managing AI performance, and presenting AI model results understandably.
Reflecting on his transition from a decade-long career at Volvo Construction Equipment (VCE) to pursuing a PhD, Leberruyer cites his motivation to learn new things and meet new people. His proximity to both Volvo and the university facilitated this shift, allowing him to enjoy the best of both worlds.
“I realized that successfully integrating AI into production systems requires communication with everyone involved”
Comparing the journey of implementing AI-driven ZDM to a famous movie plot, Leberruyer likens it to “Modern Times,” which depicts how industrial processes can dehumanize workers. He sees the development, integration, and use of AI-driven ZDM as a collaborative effort involving workshop workers, quality engineers, process engineers, IT, product development engineers, and AI specialists. This collaboration fosters creativity and empowers the entire organization to ensure customer satisfaction.
Leberruyer’s licentiate thesis has provided a solid foundation for his PhD journey. His experience in manufacturing, gained through collaboration with Volvo colleagues and discussions with supervisors and fellow PhD students, has equipped him to define the focus of his future research and make a meaningful contribution.
To conclude on a lighter note, if AI-driven ZDM were to have a superhero persona, Leberruyer envisions it as an invisible hero that aids factory workers in carrying out their daily tasks seamlessly and efficiently, ultimately improving the overall productivity and quality of the production system.
Nicolas Leberruyer’s work on AI-driven Zero Defect Manufacturing is set to revolutionize production systems. His vision of an AI-enhanced manufacturing future, where defects are almost eliminated, sustainability is prioritized, and workers are empowered, paints an inspiring picture of what lies ahead. Through his dedication and pioneering research, Leberruyer is reshaping manufacturing for a more efficient world.
Volvo Construction Equipment (Volvo CE) is a world-leading manufacturer of construction machinery such as wheel loaders, articulated haulers, and excavators. Volvo CE has a long track record of collaboration with MDU.
