AIT explores new tools for quality, inspection and process knowledge in die-casting

Die-casting foundries are under increasing pressure. Components are becoming more complex, quality requirements are rising, and customers expect stable processes and short lead times. At the same time, component dimensions, functional integration and material requirements continue to increase. Added to this are volatile energy prices, cost pressure and a labour market in which experienced personnel are becoming increasingly difficult to find. To address these challenges, the AIT Austrian Institute of Technology pursues a comprehensive approach under the banner of “Zero-Defect Digital Manufacturing”.

Written by Editors EUROGUSS 365

AIT exhibition stand at EUROGUSS 2026 — AIT presented its solutions under the banner of “Zero-Defect Digital Manufacturing” at EUROGUSS 2026.

At the heart of this approach are technologies for producing complex components using semi-solid casting, quality inspection and assisted troubleshooting through human-centred assistance systems. AIT presented these solutions at EUROGUSS 2026, where they attracted considerable interest from trade visitors. The technologies are designed specifically for industrial realities: existing die-casting cells, cycle-time-driven production environments and the need for economically viable implementation.

Identifying the causes of defects at an early stage

Many casting defects are only discovered after a component has been ejected and machined. Porosity, oxide inclusions, misruns and dimensional deviations, however, typically originate much earlier in the process: in the condition of the melt, during mould filling, through temperature control, venting, lubrication or parameter drift caused, for example, by die wear.

“Scrap is not created at the end of the production line but within the process itself,” says Florian Rötzer, research engineer at AIT. “The key is to identify the relevant influences at an early stage, draw the correct conclusions and translate them into concrete actions. If, however, I do not know the influencing factors, I cannot monitor them. And if I draw the wrong conclusions, I cannot eliminate defects effectively.”

Flexible semi-solid casting for components with high quality requirements

One of AIT’s focal areas is “MELcon Semi-Solid Casting”. The process was developed at the LKR Light Metals Technologies Ranshofen, an AIT subsidiary, and is currently being implemented in an industrial environment by MELTEC. The process transforms the melt directly within the dosing chamber into a semi-solid state containing both liquid and solid phases. This conditioned material exhibits different rheological properties from a fully liquid melt, resulting in improved behaviour during die filling.

“The MELcon semi-solid system enables the production of components with the highest quality requirements through a simple adaptation of an existing die-casting cell while maintaining maximum flexibility,” says Elias Riegler, junior research engineer. “Unlike other rheocasting technologies, no modification of the die geometry is required.”

Elias Riegler: “The MELcon semi-solid system enables the production of components with the highest quality requirements through a simple adaptation of an existing die-casting cell.”

Reduced turbulence, improved component integrity

According to AIT, the system operates with a nominal solid fraction of five to ten per cent. This is sufficient to reduce turbulence during die filling without fundamentally changing the process window. The technology is designed for integration into existing HPDC installations and enables both conventional die-casting and semi-solid operation.

This is technically significant because turbulence during filling is a major source of gas porosity and oxide inclusions. A calmer, more laminar flow improves the metallurgical integrity of the component. AIT identifies three direct benefits: lower porosity, fewer oxide inclusions and longer tool life due to reduced thermal fatigue.

The technology is particularly relevant for components that undergo machining, require high leak-tightness or must later be heat-treated or welded, where internal defects quickly become critical.

Surface inspection with more information than a camera image

Visual inspection of castings remains labour-intensive. Complex geometries, reflective surfaces and varying defect patterns often push conventional 2D inspection systems to their limits. To address this, AIT has developed “Photodex”, a handheld 3D surface inspection system that converts image data into surface information.

The system generates high-resolution synthetic greyscale images (albedo maps) as well as surface normal maps. In simple terms, it captures not only the visual appearance but also the local orientation of the surface. This makes it possible to detect fine scratches, depressions, surface irregularities and machining marks far more reliably than with standard illumination. The system also compares results with CAD data to precisely locate defects. AI-based defect classification is another integrated function.

For manufacturers, this offers two key advantages: more objective inspection results and improved traceability. When defects can be clearly located and classified, root-cause analysis in upstream process stages becomes significantly more effective.

Interactive defect annotation linking people and digitalisation

In parallel, AIT is developing interactive assistance systems for defect annotation and classification directly within the inspection process. These include both a dashboard-based interface and a pen-based system that enables precise annotation directly on the physical component. Defects are marked, classified and stored digitally together with positional information.

The pen-based approach in particular allows direct interaction at the workpiece and reduces the need to switch between the physical component and its digital representation. This makes inspection more intuitive and ergonomic while ensuring structured documentation.

Mehdi El Mnasri — Mehdi El-Mnasri: “Our goal is not to remove people from the inspection process, but to support them with digital tools and relieve them of specific tasks.”

For industry, the resulting quality data provide a foundation for root-cause analysis, traceability and the development of future AI models. Interactive defect annotation therefore supports data-driven quality assurance and creates the basis for scalable automation.
“Our goal is not to remove people from the inspection process but to support them with digital tools and preserve expert knowledge for the long term,” explains research engineer Mehdi El-Mnasri.

Robotic inspection based on CAD data instead of manual teach-in

The commissioning of automated inspection cells often represents a major bottleneck. New components, variants and geometry changes typically require extensive manual programming. AIT is therefore developing robotic inspection planning directly from CAD data.

The aim is a flexible framework that can be transferred across different robotic cells and sensor systems. According to AIT, two operating modes are possible: complete surface acquisition for geometry reconstruction and digital twin generation, or targeted inspection of critical areas with optimised cycle times.

“With CAD-based inspection planning, we replace time-consuming manual teach-in with an automated, model-based workflow. The system derives where and how a component needs to be inspected directly from its geometry while taking sensor characteristics and robot kinematics into account,” says Vanessa Staderini, scientist at AIT.

Vanessa Staderini: „The lower the effort required for reconfiguration and training, the more economically viable automated inspection becomes, even for medium production volumes.”

Economically viable automated inspection concepts

A major advantage of the framework lies in its versatility. It is not tied to a specific component, robot or sensor. As long as a geometric model of the component, a sensor model and a kinematic description of the robotic system are available, the framework can generate optimised inspection poses and paths for a wide range of applications. This enables faster commissioning, reproducible inspection quality and flexible deployment across heterogeneous robotic cells.

Staderini adds: “For serial manufacturers with frequent product changes, this is a significant lever. The lower the effort required for reconfiguration and training, the more economically viable automated inspection becomes, even for medium production volumes.”

Assistance systems for problems that are not covered by the manual

Many die-casting defects cannot be explained by a single parameter. A flash may be linked to tool condition, locking force, temperature distribution or release agents. Porosity may be caused by venting, filling behaviour, melt treatment or residual moisture. Experienced employees often recognise these patterns quickly, whereas less experienced personnel struggle to do so.

To support troubleshooting, AIT is developing a human-centred assistance system for use directly at the die-casting cell. It combines knowledge databases, process data, mathematical models, correlations, artificial intelligence and intuitive interaction concepts. Operators are guided systematically through root-cause analysis and receive recommendations for effective corrective actions.

The practical benefit is particularly evident in shift operations. Even when experienced personnel and specialist knowledge are not immediately available, complex decisions can be made more transparently, reaction times shortened, escalations reduced and downtime minimised.

Research with industrial depth

AIT combines expertise from several research domains. The Center for Vision, Automation & Control develops solutions in image processing, automation, control engineering, digitalisation and artificial intelligence. LKR Light Metals Technologies Ranshofen contributes expertise across the entire light-metal value chain, from alloy development and casting technologies to recycling. The Center for Technology Experience adds knowledge in human-machine interaction.

For die-casting companies, the decisive factors are production metrics: Does the scrap rate decrease? Are cycle times stabilised? Does tool life increase? Can knowledge be transferred more effectively? Can new components be brought into serial production more quickly?

Stephan Strommer: “Given the exponential pace of development in artificial intelligence, one thing is clear: companies that do not actively shape this transformation will be left behind.“

“Digitalisation is no longer merely an enabler; it is the fundamental prerequisite for competitiveness. What matters is the continuous acquisition and component-specific linking of process and metadata, together with the systematic digitalisation of expert knowledge, which is becoming a critical production factor in times of skilled labour shortages,” says senior research engineer and project manager Stephan Strommer.

Only the integration of data and knowledge enables contextualised visualisation, creates genuine transparency throughout the value chain and systematically reveals optimisation potential. Strommer concludes: “Given the exponential pace of development in artificial intelligence, one thing is clear: companies that do not actively shape this transformation will be left behind. At the same time, technologies such as gigacasting and closed-loop recycling systems are defining the next stage of efficient and sustainable production.”

Author

Editors EUROGUSS 365

euroguss365@nuernbergmesse.de