In metalwork industry, machinery vibration (eg. milling process) can lead to semifinished products that must be discarded in many cases. The quality of the manufactured products may drop or cannot be guaranteed. dynaVR aims at providing a IIoT cloud-based data storage, processing and analysis application for process vibrations and a Virtual Reality (VR) visualization application to better understand the effects of vibrations, anticipate potential machinery malfunctions and take the preventive decisions to deliver Zero-Defect quality products, significantly increasing product quality and lowering costs at the plant. Pilot are foreseen (aerospace sector and ZDMP Experimentation facility).

We will develop a Zero Defect Factory Layout Planner that will allow manufacturers to quicker adapt to market changes with modified factory layouts. This leads to better transparency in the production and less defects. The project is led by Ainak who has built innovative technologies leveraging augmented reality in factory planning and the work is complemented by PTW from Technical University of Darmstadt with the great experience in the production of the future and factory planning specifically with its EU wide partners with VR learning factories. The results will aim becoming a commercial product in ZDMP Marketplace.

SYLENT project aims to develop a novel “system-level quality and productivity optimizer tool” (ZPolicyManager). The tool interacts with the ZDMP components to exchange information about process and product quality. It enables the evaluation of the system level impacts of the defect avoidance policies that can be adopted within single processes. The tool will provide the ZDMP platform users with additional decision-making capabilities allowing to choose the best defect prevention policies at a system level, by avoiding local-optimal solutions, jointly optimizing quality and productivity performances of the entire production system.

The Enterprise Integration for Zero Defect Manufacturing (EZD) project proposes to develop an intuitive, efficient, and bi-directional integration component between Enterprise Information Systems (EIS), such as Enterprise Resource Planning (ERP) systems, with the ZDMP ecosystem. The sub-project will focus on the development of interoperability with the leading brand of ERP systems, SAP S/4HANA, and validate the developed assets in the context of their integration with the existing ZDMP infrastructure.

The hot forging of plastic material is widespread in the production chain of automotive and large household appliances. The cooling and shrinkage phases make it particularly challenging to correctly measure the molded parts and identify any deviations and defects. Our application allows on the one hand to collect measurements quickly and reliably, on the other hand to have a series of specific indicators available to be able to intervene as promptly as possible in case of deviations from the required quality specifications.

Specialization makes quality control a non-deterministic task; due to the high variance among different orders, it is almost a lottery to reach a performance and quality level that does not put the sustainability and profitability of a manufacturing SME at risk. This ZDMP platform validation will foster a zero-defects manufacturing process based on agile and automatized quality control with in-line data inspection of the item calibration and test (zAnomalyDetector), built over a digital-driven adaptative and dynamic metrology, verification of orders compliance (zMaterialID), end-to-end annotations (zImageAnnotator) and stock optimization (zAutomaticMaterialOrdering); to boost effectiveness.

In the food and beverage industry many foods, beers and soft drinks need to get pasteurized, a process that holds a significant role in the quality and taste of the final product but is difficult to monitor due to the process nature. We will develop an AI-enabled zApp that upgrades the typical tunnel pasteurizers giving the ability to monitor the key process parameters and estimate live the pasteurization units of final products, through AI models being compatible with various pasteurizers, via easy setup from shop floor operators. The zPasteurAIzer will be a tool for pasteurization process monitoring, early detecting quality issues, and preventing batches of defective products.

PIQ-ME-NOW delivers production executions system using inline quality control after every manufacturing operation to reduce the number of defective products and decrease waste while also improving the efficiency of the production. We will integrate existing components of ZDMP with an enhanced version of our scheduler, develop new components related to inline quality control, and demonstrate the complete system in a new domain of custom metal fabrication. To improve interoperability of the system we will develop Asset Administration Shell models describing the quality requirements of production operations.

MX3D proposes a solution to address consistency problems in robotic additve manufacturing. Using the scanning results of the UCRAM system (ZDMP open call 1), Smart Manufacturing by Adaptive Robotic Toolpath GENeration (SMARTGen) will adaptively control robotic AM processes. This is done by generating an optimal toolpath layer by layer correcting for layer height variations in previous layers. To enable SMARTGen a brand agnostic robotic control interface will be developed. MX3D's generic 3D slicer will be made available in the ZDMP market place and expanded to generate the toolpaths adaptively in a layer by layer manner.

This sub-project is aimed towards the validation of 6 zApps. The ZDMP call enables Multicursor to test a wood coating curing quality control system using the provided zApps in a new type of setup that includes the principles of zero defects, Industry 4.0 and IoT in a currently outdated sector. This sub-project will result in a tested and validated zApps with the real-world data from the industrial environment, and also the preparation of software for using in other zero defects scenarios.

SmartTwin4ZDM proposes a novel quality control and monitoring system based on smart Digital Twins (DTs) that are able to predict the evolution of the product along the manufacturing line in the most efficient and effective way to ensure product quality and reduce the appearance and propagation of defects11 . The smart DT will predict the values of the critical product KPIs and will be continuously updated with the actual values of the critical parameters of influence, measured in-line, at the same pace as the line. In this way, not only the generation of defects will be prevented but also at system level, it will prevent the propagation of the defects to downstream processes and products.

This proposal paves the way for revolutionizing process&quality control for manufacturing SMEs by taking advantage of the digital transformation not only to reduce costs and increase productivity, but also to lower the environmental impacts. The main challenge is that this transition for must be affordable, i.e. smoothly integrated in the existing automation of processes and quality control. Main novelty is using unsupervised deep learning methods for discovering change points in complex multidimensional spaces and understanding their impact on the instability in the process as a whole. Outcome is an innovative ZDMP-enabled solution, for monitoring zero defect zero waste production, integrated in ZDMP Platform and offered through ZDMP Marketplace.

Robotic cells are designed nowadays for very specific tasks, and their communication with the manufacturing line is usually very limited. However, robotic cells have great potential to improve quality, interoperability and intelligence on factories. CELL-OS aims to enhance robotic cells with AI and analytics new features, to perform predictive maintenance and quality prediction, so the cell performance can be assessed in real-time, obtaining zero-defect processes with maximum reliability for high-demand environments. CELL-OS aims to 1) validate ZDMP platform and components and 2) add new functionalities to robotic cells, providing a new generation of self-controlled and self-assessed units.

Zero-defects concept focuses on minimising the mistakes both due to humans and machines. Mental and physical states such as stress and fatigue as well as human mistakes such as account sharing, forgetting to log out are a few of the sources of human error. CALM will tackle these sources of human error by introducing continuous behavioural analysis into the Zero-defects concept. Continuous behavioural authentication combined with a higher level of behavioural analysis and anomaly detection will automatically and in real-time prevent potential defects in the production line. CALM will predict and reduce defects, increase safety and productivity as well as reduce costs and waste.

SF-Zero will leverage components from ZDMP and integrate, develop, and demonstrate an efficient toxic vapour monitoring and control system for zero-defect in process industries (e.g., cosmetics, pharma). SF-Zero extends the existing ‘smart-Flow’, a novel fume hood based on a patented push-pull high-capturing technology, manufactured by Osmose. SF-Zero will improve the currently commercialised smart-Flow with IoT and AI, to enable defect detection and safety features, hands-free control and analysis of real time and historical information, allowing engineers and operators to detect errors, react to different situations, improve the decision making and enable Industry4.0 services.

Welding is one of the traditional applications of industrial robots. However, its application to flexible automated productions is still limited. The main reasons are the limitations in robot programming and parametrization. The AI4R.WELD pilot combines automatic robot collision free program generation with advance sensing and machine learning for welding parametrization. The approach is human-centric and innovative promoting its usage and standardisation for welding application in a wide range of application areas. Fast and intuitive welding parametrization is the missing link for truly effective robotized welding.

Visual-based Non-Destructive Testing (NDT) represents the biggest share of manufacturing quality inspection methods. Recent advances in AI and Computer Vision (CV) increased the ability to automate these methods to detect defects, but widespread reliance on “black box” Deep Learning solutions makes it impossible to fully audit, trust, and therefore comply with always more demanding regulation around the explainability of industrial AI solutions. In this context, XAINDT will demonstrate a novel explainable AI method for visual-based NDT, based on zMachineAnalytics, applied to radiographic inspection in aerospace foundries to spot 100% of welding, surface and precision constraints defects.

zExplAIn is a new explainable-based zApp to improve Predictive Maintenance AI models in the ZDMP platform. zExplAIn provides rational and insights on the origin and root causes of the anomalies detected within ZDMP, acting as an analytical layer that expands the results already provided by existing AI-based zApps. zExplAIn complements ZDMP black-box models with an additional dedicated surrogate model (hybrid approach) to provide understandability metadata. It will be fully integrated in the ZDMP platform at its three levels (design, use, run-time), and validated in machining processes, plastic injectors and glass-melting processes.