REGAIN Shows the Path from ROI to Return to Future

After three years of intensive research work, the REGAIN collaborative project has clearly shown the innovative strength of the German foundry industry. Together with 20 project partners, around 1,000 person-months of research effort and funding of approximately €7.7 million, digital technologies, AI-based assistance systems, and new approaches to resource efficiency were developed and tested directly under real production conditions.

At the closing event during the German Foundry Congress 2026 in Göttingen, Dr. Kai Kerber (insightfabrik solutions GmbH), Prof. Dr. Dierk Hartmann (Steinbeis-Transferzentrum), Jan Nordmeyer (TU Braunschweig), Marija Lindner (TU Braunschweig) and Dr. Felix Ganser-Lentin (Blackforge AG), together with industry representatives Joshua Bissels (PINTER GUSS GmbH), Albert Miller (ABP Induction Systems AG) and Dr. Sebastian Tewes (BDG), presented the project’s key results and outlined perspectives for the future of data-driven foundry processes.

From the very beginning, Dr. Kai Kerber stressed that technological progress alone is not sufficient. What is crucial is the industry’s capability to successfully initiate and implement research projects. The challenges are growing in complexity: funding programs are becoming more competitive, application processes more demanding, and technological developments, especially in the field of artificial intelligence, are advancing at a pace that increasingly outstrips traditional research cycles. At the same time, REGAIN showed that the foundry industry has a highly capable innovation ecosystem. Universities, research institutions, industry associations, and companies have jointly demonstrated that topics such as AI, robotics, digital twins, CO₂ accounting, and Catena-X can be effectively transferred into industrial application.

One of the central insights of the project is that the vision of an intelligent foundry cannot be achieved through a single “all-knowing” AI system. Instead, a range of specialized assistance systems and enabling technologies were developed to close individual knowledge and data gaps, gradually paving the way toward comprehensive optimization solutions. In total, the project delivered ten technological core modules and 18 AI services.

Cast Components Become Entirely Traceable

Prof. Dr. Dierk Hartmann demonstrated how the vision of full traceability can be realized in practice through the work of the sand casting cluster. In collaboration with project partners, a “Digital Casting Pass” was developed, functioning as a digital identity card for cast components. Its core principle is the seamless, end-to-end traceability of every individual casting.

This is achieved by embedding unique codes directly into the mold, which are later retrieved from the finished casting. In addition, cores are given a unique identity through an innovative fingerprinting process, in which the natural grain structure of the sand core serves as an inherent identifier, comparable to a human fingerprint, but embedded in the material itself.

Together, these methods enable for the first time a complete and consistent linkage of process data, core manufacturing data, and quality information at the level of individual components. Hartmann emphasized that this represents a decisive step in overcoming one of the central challenges in sand casting: the direct and reliable connection between production processes and resulting component quality. The developed systems have already been successfully validated in ongoing production at a medium-sized aluminum foundry, unlocking entirely new possibilities for data-driven quality analysis and AI-supported process optimization.

An Assistance System for Die Casting Operators

The die-casting cluster concentrated on increasing transparency in highly dynamic production processes. Jan Nordmeyer from TU Braunschweig presented two particularly successful developments from this area.

The first application is an assistance system for monitoring the quality of release agent application. Using temperature sensors integrated into the mold, artificial intelligence can detect process faults such as failed nozzles, incorrect spray distances, or misalignment of the spray unit. The system is based on analyzing heat extraction effects caused by the evaporation of the release agent, enabling a highly sensitive and indirect form of process monitoring.

A second key focus was melt-front detection during mold filling. For this purpose, researchers developed a sensor system based on piezoelectric force measurements behind ejector pins. The resulting signals are then compared with simulation data from MAGMASOFT, allowing deviations between simulation and real process behavior to be identified at an early stage. This supports faster tool setup and enables long-term optimization of the casting process.

A particularly important aspect is that both solutions can be retrofitted into existing production equipment, significantly simplifying their transfer into industrial practice.

Quality and Energy Efficiency Drive Research in Permanent Mold Casting

Marija Lindner presented the results of the permanent mold casting cluster, which placed a strong focus on quality assurance, energy efficiency, and CO₂ reduction. Across the entire process chain, from melting operations and melt treatment through to final quality inspection, multiple digital assistance systems were developed and tested.

A central highlight was an AI-based assistant for predicting casting quality. The system evaluates process data such as mold temperatures, melt temperatures, and cycle times, and provides operators with a real-time forecast of whether a casting will meet defined quality requirements. In addition, the AI delivers a confidence score that transparently reflects the reliability of each prediction, increasing trust and usability in industrial environments. The impact is clearly measurable: one project partner was able to reduce its scrap rate from seven percent to five percent within just six months, equivalent to a 44% improvement.

In parallel, Fraunhofer IIS developed an AI assistant for automated X-ray image evaluation. It reliably detects even the smallest defects with high consistency while significantly reducing the effort required for manual inspection. Additional assistance systems were implemented for energy-optimized melting shop planning, hydrogen degassing, grain refinement and modification, as well as for identifying energy flexibility potential and CO₂ footprints. Together, these developments demonstrate how data-driven approaches can simultaneously enhance product quality and deliver substantial reductions in energy consumption and resource usage.

From Research to Industrial Reality: Enabling Scalable AI Deployment in Foundry Operations

While the technical solutions were developed within the individual clusters, the digitalization cluster addressed an equally critical question: how research results can be sustainably transferred into industrial application. Dr. Felix Ganser-Lentin presented the Core Platform developed for this purpose, which forms the technological foundation for the productive deployment of AI systems.

The platform handles data management, visualization, user administration, and the standardized integration of a wide range of AI applications. In addition, interfaces to Catena-X and other data spaces were implemented to enable future cross-company data exchange. The overarching goal is to establish an open, scalable infrastructure in which new applications can be deployed in a manner similar to apps in a digital marketplace.

Ganser-Lentin emphasized that many research projects produce functional AI models that ultimately fail to transition into long-term industrial operation. The Core Platform was therefore deliberately designed as a bridge between research and productive use. The demonstrators developed within the project already illustrate how data from diverse sources can be combined, analyzed, and translated into actionable recommendations.

How Does Industry Benefit? From ROI Thinking to a “Return to the Future” Mindset

The discussion at the closing event went far beyond the presentation of individual technologies. Contributions from Joshua Bissels, Albert Miller, and Dr. Sebastian Tewes made it clear that the central challenge is no longer the development of additional demonstrators, but the transfer of existing solutions onto the shop floor.

The past three years have demonstrated that AI applications can work in foundry environments. The key question now is how these systems can be embedded into day-to-day operations.

The response from industry representatives was unambiguous: the goal is not to build a fully autonomous foundry overnight. Instead, companies must develop the necessary infrastructure and build experience step by step. Data needs to be made accessible, interfaces must be established, and employees actively involved. Only under these conditions can digital assistance systems unfold their full potential.

At the same time, it became clear that digitalization is no longer a temporary initiative. It is increasingly becoming a corporate mindset and, in some cases, an independent business model. Companies that continue to treat digitalization as an isolated IT project risk falling behind. Data-driven processes, AI-supported decision-making, and digital value creation must become integral components of corporate strategy.

A particularly important development is the growing integration of production control and economic optimization. The future lies not only in technically stable production, but in contribution-margin-oriented manufacturing. AI systems will increasingly support decisions on which products should be produced, under which conditions, and with what expected profitability. This shifts the focus from pure efficiency improvements toward the intelligent management of the entire value creation process.

The Time for a “Doers’ Community” Has Come

Participants also pointed out that too little attention is currently given to new organizational models, forms of collaboration, and consolidation strategies within the industry. Technological developments dominate the discussion, whereas topics such as shared platforms, innovative business models, and structured partnerships between companies are often not addressed with the same intensity. However, in an environment marked by increasing investment demands and growing international competition, these aspects are likely to become key success factors.

A broad consensus emerged that the industry must now transition from analysis to execution. The proposal was made to establish an active “doers’ community” that exchanges experience, implements pilot projects, and accelerates the transfer of successful solutions into practice. The necessary technology is available. The expertise is available. What is now required is the determination to apply both consistently. A statement that particularly reflected the atmosphere of the event was: “We are condemned to optimism.”

Even under challenging economic conditions, high energy costs, and structural pressures, the European foundry industry continues to demonstrate considerable technological potential. REGAIN has illustrated that innovation is not limited to start-ups or large technology corporations but is also strongly present in established industries such as foundry manufacturing.

This shifts the focus to a question that goes beyond “if” digitalization and AI create value, toward “how” this value should be understood and evaluated. For many years, companies have primarily relied on traditional ROI, Return on Investment. The discussions at the REGAIN closing event introduced a more comprehensive perspective: “Return to Future.”

This concept describes the ability of organizations to ensure long-term viability through digitalization, AI, and data-driven processes. The value generated is not confined to short-term savings or reduced scrap rates. It also appears in stronger competitiveness, increased responsiveness, better decision-making capabilities, and the emergence of entirely new business models.

This may well represent one of REGAIN’s most significant achievements. The project not only delivered technological developments; it also demonstrated that the foundry industry possesses the knowledge base, partnerships, and innovative capacity needed to actively shape its own future. The AI systems developed do not replace human expertise. Instead, they create transparency, support better decision-making, and help skilled professionals understand and control complex processes more effectively. The path toward the intelligent foundry is not yet complete but thanks to REGAIN, it has become clearly visible.