Periodic Reporting for period 3 - OPENZDM (OPEN PLATFORM FOR REALIZING ZERO DEFECTS IN CYBER PHYSICAL MANUFACTURING)
Periodo di rendicontazione: 2024-12-01 al 2025-11-30
Zero Defect Manufacturing (ZDM) aims to go beyond traditional lean management and Six Sigma by using artificial intelligence, IoT, big data, and digital twin simulation to advance quality in Industry 4.0 cyber-physical production systems and, as we move forward, Industry 5.0. ZDM approaches target zero defects in production. To date, manufacturing in Europe and ZDM strategies have focused on reducing defects by integrating in-line and off-line quality monitoring solutions at shop-floor level. In recent years, research on quality control and AI-driven tools for smart manufacturing under the Industry 4.0 paradigm has provided a solid framework for minimising rework and promoting first-time-right schemes. However, these strategies often lack a clear focus on sustainability alongside product quality, limiting the EU industry’s ability to adapt to the challenges reflected in the UN 2030 Agenda and the Sustainable Development Goals (SDGs).
openZDM addresses this gap by delivering an open digital platform that integrates existing R&D solutions into a state-of-the-art system supporting zero-defect processes across production networks, with a strong focus on sustainable manufacturing. The platform builds on the Reference Architecture Model for Industry 4.0 (RAMI 4.0) and the Asset Administration Shell (AAS), which are essential for interoperability and intra-factory quality management. Beyond RAMI 4.0/AAS-based interoperability, the project incorporates non-destructive inspection (NDI) methods and data-driven quality assessment techniques for online defect detection and quality evaluation. Digital twins are a key enabling technology, supporting online process adaptation, simulation, and waste reduction through “as-designed” and “as-implemented” data services. The openZDM concept combines (i) the open platform for continuous quality assessment and control, (ii) digital twins for online process adaptation, (iii) NDI systems for defect identification using different technologies, (iv) data analytics for real-time monitoring and prediction, and (v) a decision-making module that evaluates alternative process reconfigurations via the digital twin. Its novelty lies in treating quality assessment and control not as external add-ons, but as integral functions of the cyber-physical production system. Finally, openZDM advances the state of the art in manufacturing reconfiguration by using explainable AI for real-time, in-line root-cause analysis, supporting trustworthy intelligent systems aligned with both Industry 4.0 and Industry 5.0. Overall, the methodology promotes sustainable manufacturing by preventing waste through increased flexibility and adaptive control, contributing to circular-economy objectives as well as pollution prevention and control.
The openZDM framework helps manufacturers increase their business value in the following key aspects:
• Increase organisational flexibility and the capacity to continuously control quality and innovate.
• Increase cost savings through waste reduction.
• Promote sustainability through online process control and adaptation.
• Improve quality through data analytics and on-demand decision support, increasing efficiency.
• Increase productivity and competitiveness.
Beyond its core research and innovation objectives, openZDM also targets cross-cutting outcomes, including: (a) contributions to ZDM standardisation through participation in the release of CWA 18230:2025, and (b) skills development via online training tutorials and technical articles on enabling technologies.
openZDM addresses this gap by delivering an open digital platform that integrates existing R&D solutions into a state-of-the-art system supporting zero-defect processes across production networks, with a strong focus on sustainable manufacturing. The platform builds on the Reference Architecture Model for Industry 4.0 (RAMI 4.0) and the Asset Administration Shell (AAS), which are essential for interoperability and intra-factory quality management. Beyond RAMI 4.0/AAS-based interoperability, the project incorporates non-destructive inspection (NDI) methods and data-driven quality assessment techniques for online defect detection and quality evaluation. Digital twins are a key enabling technology, supporting online process adaptation, simulation, and waste reduction through “as-designed” and “as-implemented” data services. The openZDM concept combines (i) the open platform for continuous quality assessment and control, (ii) digital twins for online process adaptation, (iii) NDI systems for defect identification using different technologies, (iv) data analytics for real-time monitoring and prediction, and (v) a decision-making module that evaluates alternative process reconfigurations via the digital twin. Its novelty lies in treating quality assessment and control not as external add-ons, but as integral functions of the cyber-physical production system. Finally, openZDM advances the state of the art in manufacturing reconfiguration by using explainable AI for real-time, in-line root-cause analysis, supporting trustworthy intelligent systems aligned with both Industry 4.0 and Industry 5.0. Overall, the methodology promotes sustainable manufacturing by preventing waste through increased flexibility and adaptive control, contributing to circular-economy objectives as well as pollution prevention and control.
The openZDM framework helps manufacturers increase their business value in the following key aspects:
• Increase organisational flexibility and the capacity to continuously control quality and innovate.
• Increase cost savings through waste reduction.
• Promote sustainability through online process control and adaptation.
• Improve quality through data analytics and on-demand decision support, increasing efficiency.
• Increase productivity and competitiveness.
Beyond its core research and innovation objectives, openZDM also targets cross-cutting outcomes, including: (a) contributions to ZDM standardisation through participation in the release of CWA 18230:2025, and (b) skills development via online training tutorials and technical articles on enabling technologies.
During the reporting period, significant progress has been made in the development, deployment, and validation of key technologies:
- Final versions of NDIs developed, with 11 deployed in production (laser line triangulation for point cloud acquisition deployed between M30 and M36), and 2 remaining offline (X-ray systems).
- All NDIs, except the X-ray systems, are fully integrated with the openZDM platform, with their data being utilised by integrated applications.
- 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.
- Continuous performance-optimising updates of the AAS middleware have been released.
- A drift controller based on generative and explainable AI has been implemented to mitigate potential drift in AI-based quality assessment modules. The result has been documented and published in a scientific article.
- Explainable AI has been adopted as an add-on layer to quality assessment modules, such as one in the VDLWEW use case, to enable real-time root-cause analysis.
- A new AAS Submodel targeting machine vision systems has been standardised through ECLASS.
- An analysis was conducted on the applicability of AI model transfer for cross-product applications in the APTIV pilot, with the findings being documented in a published research article.
- An analysis was performed to identify the potential of CNN for 3D defect localisation in the inspection of EV battery modules, with the findings being documented in a published research article.
- An IPR agreement has been established between partners and has been included in the private version of the final WP6 deliverable.
- The consortium actively engaged in the Horizon Results Booster services, focusing on the openZDM integrated platform and the G3F NDI as Key Exploitable Results to refine exploitation intentions, value propositions, and facilitate the identification of early adopters.
- Consortium partners have contributed to the publication of CWA 18230:2025 “Zero Defects Manufacturing – Basic Principles and Requirements”.
- The industrial testbeds have updated and tested the updated version of NDIs from WP3 and applications from WP4, providing critical feedback for validation.
- Final versions of NDIs developed, with 11 deployed in production (laser line triangulation for point cloud acquisition deployed between M30 and M36), and 2 remaining offline (X-ray systems).
- All NDIs, except the X-ray systems, are fully integrated with the openZDM platform, with their data being utilised by integrated applications.
- 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.
- Continuous performance-optimising updates of the AAS middleware have been released.
- A drift controller based on generative and explainable AI has been implemented to mitigate potential drift in AI-based quality assessment modules. The result has been documented and published in a scientific article.
- Explainable AI has been adopted as an add-on layer to quality assessment modules, such as one in the VDLWEW use case, to enable real-time root-cause analysis.
- A new AAS Submodel targeting machine vision systems has been standardised through ECLASS.
- An analysis was conducted on the applicability of AI model transfer for cross-product applications in the APTIV pilot, with the findings being documented in a published research article.
- An analysis was performed to identify the potential of CNN for 3D defect localisation in the inspection of EV battery modules, with the findings being documented in a published research article.
- An IPR agreement has been established between partners and has been included in the private version of the final WP6 deliverable.
- The consortium actively engaged in the Horizon Results Booster services, focusing on the openZDM integrated platform and the G3F NDI as Key Exploitable Results to refine exploitation intentions, value propositions, and facilitate the identification of early adopters.
- Consortium partners have contributed to the publication of CWA 18230:2025 “Zero Defects Manufacturing – Basic Principles and Requirements”.
- The industrial testbeds have updated and tested the 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 11 deployed in line and integrated with the openZDM platform. Two NDIs remain offline due to safety and regulation-related concerns (X-rays).
2. Agent-based decision-making approach for alternative scenario authoring and evaluation.
3. Combination of generative AI with explainable AI for the quality assessment module’s drift detection and mitigation.
4. Improved performance in the openZDM AAS middleware, making the platform more appealing in time-critical applications.
1. Final versions of NDIs developed, with 11 deployed in line and integrated with the openZDM platform. Two NDIs remain offline due to safety and regulation-related concerns (X-rays).
2. Agent-based decision-making approach for alternative scenario authoring and evaluation.
3. Combination of generative AI with explainable AI for the quality assessment module’s drift detection and mitigation.
4. Improved performance in the openZDM AAS middleware, making the platform more appealing in time-critical applications.