FLASH-COMP: AI-based quality control for composite manufacturing

FLASH-COMP aims to revolutionise the production of high-value composite parts, with a primary focus on achieving zero-defect manufacturing, through the development of a state-of-the-art quality control solution. By testing this technology on the manufacture of yacht hulls and aeroplane wings, the project will demonstrate how waste can be minimised and production standards raised by integrating artificial intelligence into any quality control process that will eventually see products made right first time.

Composite manufacturing, particularly liquid resin infusion (LRI), faces significant challenges in ensuring product quality. Manual inspection processes often lead to post-production defects, resulting in material wastage and increased costs. FLASH-COMP addresses this issue by implementing advanced inspection and monitoring techniques at key manufacturing stages, such as preforming and infusion, to retrieve key process parameters. This data powers an AI-based tool that accurately predicts and identifies defects in real-time, enabling workers to take immediate corrective actions.

By eliminating defects during production, FLASH-COMP will reduce the need for excess raw materials, often deployed in LRI processes, and it minimises trimming requirements during the finishing stages. With this technology, the process of detecting and rectifying defects is seamlessly integrated into production, ensuring that every final product is right first time.

Now, with the project reaching its mid-way point, three significant milestones have been achieved that will enable the technology to be implemented at the two pilot sites and for rigorous testing to begin. As well as the development of the technology, work has also been completed on setting up evaluation processes designed to assess the readiness of any company to adopt ZDM technology and begin manufacturing their products right first time.

Establishing a robust data management infrastructure

At the core of FLASH-COMP lies a bespoke data management system created by the team at MSI in Spain, which has created a novel centralised data space conforming to international standards. This allows the project to lay the foundation for its Decision Support System (DSS).

This infrastructure enables seamless integration of data from various inspection tools along a production line and uses AI algorithms, providing workers with actionable insights for process optimisation. Essentially, the data space serves as a central hub containing all necessary data, applications, algorithms, and AI systems for the whole FLASH-COMP solution, all of which will be integrated into the FLASH-COMP DSS.

Data governance rules for the space are also being developed and these will establish different user profiles and regulate access to sensitive data to ensure proper control over data usage and commercial exploitation. These stringent data governance rules ensure the ethical handling of sensitive information.

The data space also offers functionalities such as monitoring data flow and simulating the entire data processing pipeline, from collection to presentation, at the two FLASH-COMP pilot sites. Additionally, it incorporates an Asset Administration Shell (AAS) Registry Server for developing digital twins,
enabling MSI to create digital replicas of real hardware and inspection tools. The data being collected by the partners’ various inspection and monitoring tools is being supplemented by synthetic data so that AI-generated simulations can be developed.

An illustrative example of data flow at this first phase involves the monitoring of composite material flow using cameras. By analysing data on bubble formation in the material, for example, quality standards can be quickly assessed and potential risks to product quality identified and corrected.

Enhancing bubble detection and active control strategies

One of the primary challenges in composite manufacturing is the formation of bubbles during the resin infusion process. FLASH-COMP addresses this issue by deploying advanced matrix cameras and AI systems for real-time detection, followed by mitigation strategies to avoid the inclusion of bubbles.

Also during the infusion process in LRI, one of the main aspects which is monitored is the flow front. FLASH-COMP is now able to track this using these cameras and implement active control strategies to minimise defects and optimise the infusion process.

The fact that the matrix camera is a very common technology is a real plus for the project as it makes the whole solution a lot more accessible and affordable. With the addition of AI systems and detection technologies information, from this system can now be used for active control strategies for the infusion process. This will allow workers to see when the resin is not flowing properly, which often leads to defect formation. Based on this information corrective strategies to improve the infusion process can be implemented.

Additionally, FLASH-COMP is exploring other innovative techniques, such as LIDAR scanning, to identify regions of interest for high-resolution monitoring, further enhancing process efficiency.

Now these technologies have been developed in the lab, they are being deployed at the pilot sites, where the AI algorithms will be further tested and a strategy for scaling up will be developed.

Pioneering ZDM frameworks

Beyond these two developments in the DSS and in defect detection, FLASH-COMP is also advancing access to ZDM deployment by developing a comprehensive evaluation framework that will both assess a company’s readiness to adopt these digital solutions as well as provide a roadmap for their implementation. This important work is key to enabling the uptake of new technology that will eventually deliver better, cheaper and more sustainable products.

Project partners from POLIMI, INEGI, and ESTIA have published the first iteration of this sustainability evaluation framework that encompasses the methodologies and tools needed for assessing the readiness levels of companies to adopt ZDM technologies.

At its core, the evaluation framework revolves around three key assessments: Zero-Defect Manufacturing Maturity, Efficiency and Eco-efficiency, and Human Factors, each playing a crucial role in fostering the adoption of ZDM technologies.

The Zero-Defect Manufacturing Maturity Assessment serves as a compass, guiding companies through the stages of ZDM adoption. Built on the 6Ps digital maturity assessment model, previously developed by POLIMI, the assessment evaluates six dimensions: Products, Processes, Platform, People, Partnership, and Performance. These dimensions are clustered into technical and socio-business pillars, providing a comprehensive view of the maturity of an organisation’s  strategy, processes, people and technologies for ZDM implementation.

The maturity levels for each of these dimensions, ranging from Initial to Exploited, offer a roadmap for companies to gauge their progress towards achieving zero defects. From basic quality control processes to the integration of advanced practices, skills and technologies like machine learning and digital twins, each level signifies a milestone in the journey towards ZDM excellence.

Meanwhile, as efficiency and eco-efficiency are the cornerstones of sustainable manufacturing, the FLASH-COMP assessment methodology also encompasses both economic and environmental considerations. Through processes like Multi-Layer Value Stream Mapping and Eco-efficiency Analysis, companies can identify inefficiencies, minimise costs, and mitigate environmental impacts.

Evaluating key performance indicators (KPIs) across different phases of manufacturing processes allows companies to streamline operations. By meticulously analysing processes before and after technology implementation, this approach aims to optimise resource allocation, minimise waste, and maximise added value. This holistic approach not only enhances competitiveness but also fosters responsible stewardship of resources. The FLASH-COMP methodology helps identify those KPIs, so that companies can assess where they are and monitor progress towards implementing digital technologies that will allow them to deliver ZDM products to the market. Initially, FLASH-COMP will lay the groundwork by defining KPIs and configuring the tools.

Finally in terms of this readiness assessment methodology, FLASH-COMP is considering the human element involved in adopting new technologies. People remain indispensable in manufacturing and this is a central pillar in FLASH-COMP’s approach to ZDM. The Human Factors Assessment within the project focuses on enhancing ergonomic design and usability of ZDM technologies for the workers involved as well as ensuring workers on the ground develop the necessary skills to make the most out of the new ZDM methods and technologies. By prioritising user feedback and incorporating ergonomic recommendations, companies can then enhance employee satisfaction, productivity, and overall system performance.

Conclusion

In essence, the FLASH-COMP goal is twofold: to streamline manufacturing processes and empower human capital. Now at the mid-way stage of its project lifespan, FLASH-COMP is looking to build on these key developments and its focus on implementation and impact assessment. The two industrial testbeds will serve as platforms for validating the innovations and through rigorous data collection and analysis, the project will gain valuable insights into process optimisation and performance evaluation. As part of its exploitation strategy, companies looking to move towards ZDM will be able to leverage these insights, and drive their own tangible improvements in quality, efficiency, and sustainability. FLASH-COMP will represent a significant step forward for composite manufacturing. The evaluation methodologies and digital tools so far established by the project lay the foundation for ZDM, paving the way for a future where defects are not just minimised but eliminated altogether.