Periodic Reporting for period 2 - OPENZDM (OPEN PLATFORM FOR REALIZING ZERO DEFECTS IN CYBER PHYSICAL MANUFACTURING)
Reporting period: 2023-12-01 to 2024-11-30
Zero Defects Manufacturing (ZDM) is aiming at going beyond traditional lean management and six-sigma approaches with the use of artificial intelligence, IoT and Big Data for reaching the new quality paradigm in the Industry 4.0 era of Cyber-Physical Production Systems. ZDM approaches target zero defects in a production environment. Manufacturing processes in Europe and ZDM strategies considered for their optimization, so far have focused on the reduction of production defects by integrating in-line and off-line quality monitoring solutions at the shopfloor level . In the last years, research on quality control and AI-driven tools for smart manufacturing based on Industry 4.0 paradigm has provided a reliable framework for minimizing rework and promoting first-time-right schemes. However, these strategies lack a clear focus on promoting production sustainability together with product quality control, hindering the EU industry's capability for adapting to the upcoming challenges reflected in the 2030 Agenda for the UN SDGs. The most beneficial way for future zero-defect practices will be applied is through the ability to integrate value chains in the product/production lifecycle with feedback and feedforward loops.
Following the above, the main objective of openZDM is to provide an open platform to support production networks’ zero-defect processes, bringing together existing R&D solutions and creating an innovative state-of-the-art integrated solution. openZDM is focusing on the “grand challenge” of Sustainable Manufacturing, with the gravity of enabling zero-defect processes. openZDM is aiming to develop a digital platform that builds on the state-of-the-art Reference Architecture Model for Industry 4.0 (RAMI 4.0) and Asset Administration Shell (AAS) that is basically to implement intra-factory quality management practices. Apart from the RAMI4.0 and AAS based solutions for interoperability, this project considers several non-destructive inspection (NDI) methods and data-driven quality assessment techniques for online defect identification and quality assessment. In addition, the Digital Twin is one of the project’s key enabling technologies for online process adaptation and wastes reduction. The openZDM aims to enable digital twin or reduced order model services, such as simulation and co-simulation, virtual commissioning/reconfiguration, and as-designed and as-implemented data services, aiming to significantly improve the production sustainability of cyber-physical production systems. The openZDM innovative concept for providing an integrated platform for continuous quality assessment and control towards a ZDM and enabling sustainable manufacturing in Europe, includes the openZDM platform, the digital twin for process online adaptation, NDI systems for defects identification using different technologies and data analytics for quality assessment. Finally, a decision-making module will evaluate alternative process reconfigurations via the digital twin. The novelty of the proposed system is that rather than treating quality assessment and control as unknowns of a cyber-physical production system, they are implemented as an integral part of these. In addition, to enable classic CPS functions such as real-time monitoring, deviation control, process optimization, and life-cycle transparency, this enables supporting a sustainable production’s needs and requirements through the main digital infrastructure of CPPSs as well as holistic and continuous quality control and online adaptation of the manufacturing processes. Regarding the EU regulation 2020/852, the proposed openZDM methodology encourages investment for sustainable manufacturing, preventing wastes via increased production flexibility and adaptive control, thus, encouraging objectives 4) Transition to a circular economy, waste prevention and recycling, addressing primarily the prevention of wastes through its innovative solutions for ZDM, and 5) Pollution prevention and control.
The openZDM framework, is expected to increase the business value in the following key aspects:
• Increase the flexibility of an organisation and its capacity to continuously control its quality and innovate.
• Increase the capacity for cost savings through wastes reduction.
• Promote production sustainability through online process control and adaptation.
• Improve production quality with the capacity to perform data analytics, on-demand decision support and hence increased efficiency.
• Increase productivity and competitive leverage.
Apart from the “hard” research and innovation objectives stated above openZDM targets additional “soft” cross-cutting objectives such as a) contribution to standards for ZDM and b) skills development for ZDM.
Following the above, the main objective of openZDM is to provide an open platform to support production networks’ zero-defect processes, bringing together existing R&D solutions and creating an innovative state-of-the-art integrated solution. openZDM is focusing on the “grand challenge” of Sustainable Manufacturing, with the gravity of enabling zero-defect processes. openZDM is aiming to develop a digital platform that builds on the state-of-the-art Reference Architecture Model for Industry 4.0 (RAMI 4.0) and Asset Administration Shell (AAS) that is basically to implement intra-factory quality management practices. Apart from the RAMI4.0 and AAS based solutions for interoperability, this project considers several non-destructive inspection (NDI) methods and data-driven quality assessment techniques for online defect identification and quality assessment. In addition, the Digital Twin is one of the project’s key enabling technologies for online process adaptation and wastes reduction. The openZDM aims to enable digital twin or reduced order model services, such as simulation and co-simulation, virtual commissioning/reconfiguration, and as-designed and as-implemented data services, aiming to significantly improve the production sustainability of cyber-physical production systems. The openZDM innovative concept for providing an integrated platform for continuous quality assessment and control towards a ZDM and enabling sustainable manufacturing in Europe, includes the openZDM platform, the digital twin for process online adaptation, NDI systems for defects identification using different technologies and data analytics for quality assessment. Finally, a decision-making module will evaluate alternative process reconfigurations via the digital twin. The novelty of the proposed system is that rather than treating quality assessment and control as unknowns of a cyber-physical production system, they are implemented as an integral part of these. In addition, to enable classic CPS functions such as real-time monitoring, deviation control, process optimization, and life-cycle transparency, this enables supporting a sustainable production’s needs and requirements through the main digital infrastructure of CPPSs as well as holistic and continuous quality control and online adaptation of the manufacturing processes. Regarding the EU regulation 2020/852, the proposed openZDM methodology encourages investment for sustainable manufacturing, preventing wastes via increased production flexibility and adaptive control, thus, encouraging objectives 4) Transition to a circular economy, waste prevention and recycling, addressing primarily the prevention of wastes through its innovative solutions for ZDM, and 5) Pollution prevention and control.
The openZDM framework, is expected to increase the business value in the following key aspects:
• Increase the flexibility of an organisation and its capacity to continuously control its quality and innovate.
• Increase the capacity for cost savings through wastes reduction.
• Promote production sustainability through online process control and adaptation.
• Improve production quality with the capacity to perform data analytics, on-demand decision support and hence increased efficiency.
• Increase productivity and competitive leverage.
Apart from the “hard” research and innovation objectives stated above openZDM targets additional “soft” cross-cutting objectives such as a) contribution to standards for ZDM and b) skills development for ZDM.
During the reporting period from M19 (December 1st, 2023) to M30 (November 30th, 2024), significant progress has been made in the development, deployment, and validation of key technologies:
- Final versions of NDIs developed, with 10 deployed in production, 1 pending deployment (laser line triangulation for point cloud acquisition), and 2 remaining offline (X-ray systems).
- Ffinal calibration methodologies for NDI systems have been established and documented (D3.5).
- Final AAS implementations, including the AAS creation workflows for Type 2 and Type 3, have been completed and reported in D4.3.
- Due to the lack of specific technical documentation, the Type 3 AAS implementation was developed based on the consortium’s understanding, ensuring functionality within the project scope.
- An analysis was conducted on developing AAS Type 3 using Basyx instead of the openZDM AAS middleware. Results were documented and submitted to an article.
- The platform infrastructure has been expanded from 3 virtual machines to 6virtual machines and a dedicated server, improving scalability and performance.
- Deployment of the openZDM system through Kubernetes has been started and progressed to enhance flexibility and orchestration.
- Updates and enhancements have been made to key software applications, including the digital twin toolset and models, the analytics modules, the decision support tool, and the integrated platform with the AAS middleware.
- The G3F system was updated and mounted on a robot at the VWAE pilot to enhance automation capabilities.
- Visual transformers were implemented for detecting minute aesthetic defects in the APTIV pilot, improving defect detection accuracy.
- XAI (Explainable AI) techniques were incorporated into digital twins and quality assessment modules, adding interpretability to the analytics framework.
- The industrial testbeds have installed and tested updated version of NDIs from WP3 and applications from WP4, providing critical feedback for validation.
- Final versions of NDIs developed, with 10 deployed in production, 1 pending deployment (laser line triangulation for point cloud acquisition), and 2 remaining offline (X-ray systems).
- Ffinal calibration methodologies for NDI systems have been established and documented (D3.5).
- Final AAS implementations, including the AAS creation workflows for Type 2 and Type 3, have been completed and reported in D4.3.
- Due to the lack of specific technical documentation, the Type 3 AAS implementation was developed based on the consortium’s understanding, ensuring functionality within the project scope.
- An analysis was conducted on developing AAS Type 3 using Basyx instead of the openZDM AAS middleware. Results were documented and submitted to an article.
- The platform infrastructure has been expanded from 3 virtual machines to 6virtual machines and a dedicated server, improving scalability and performance.
- Deployment of the openZDM system through Kubernetes has been started and progressed to enhance flexibility and orchestration.
- Updates and enhancements have been made to key software applications, including the digital twin toolset and models, the analytics modules, the decision support tool, and the integrated platform with the AAS middleware.
- The G3F system was updated and mounted on a robot at the VWAE pilot to enhance automation capabilities.
- Visual transformers were implemented for detecting minute aesthetic defects in the APTIV pilot, improving defect detection accuracy.
- XAI (Explainable AI) techniques were incorporated into digital twins and quality assessment modules, adding interpretability to the analytics framework.
- The industrial testbeds have installed and tested updated version of NDIs from WP3 and applications from WP4, providing critical feedback for validation.
The main outputs implemented in the reporting period that could deliver impact are as follows:
1. Final versions of NDIs developed, with 10 already deployed in line, 1 is pending for deployment (laser line triangulation system for point cloud acquisition) and 2 will be offline (X-rays).
2. Methodology to dynamically create a digital twin using data-driven and/or-physics based methods
3. Integration of XAI methods for increased quality assessment outcomes interpretability and for building trust to the users of the openZDM system
4. Updated infrastructure for hosting the openZDM system, targeting increased performance and better data handling
1. Final versions of NDIs developed, with 10 already deployed in line, 1 is pending for deployment (laser line triangulation system for point cloud acquisition) and 2 will be offline (X-rays).
2. Methodology to dynamically create a digital twin using data-driven and/or-physics based methods
3. Integration of XAI methods for increased quality assessment outcomes interpretability and for building trust to the users of the openZDM system
4. Updated infrastructure for hosting the openZDM system, targeting increased performance and better data handling