With more than 390,000 companies and around 3.7 million employees, the metalworking industry represents the largest secondary sector within the EU. In this sector, machining technology represents one of the most important manufacturing technologies. Numerous key German industries generate a large part of their product value creation through machinings, such as tool and mold making, the semiconductor industry, or engine construction. Due to the high importance of machining production, the companies concerned are interested in the continuous optimization of machining processes in terms of quality, productivity, cost-effectiveness, and, increasingly, sustainability.

Through digitization, machining processes are represented by digital twins, which enables continuous planning, manufacturing, and quality assurance. Models and simulations based on digital twins are mostly excluded in industrial applications due to their computational requirements and the expert knowledge needed to operate them. Consequently, relevant physical effects in industrial practice are either neglected or only approximated by rough estimations. As a result, the quality of the digital twin and the findings and decisions derived from it suffer considerably, which in many cases can lead to significant economic disadvantages for companies.

Due to the high-quality requirements and the usually considerable costs for rejects, simulations make it possible to plan optimized machining processes on the basis of digital twins. These technology-specific simulation models are primarily based on models from analytics, numerics, and, increasingly also, the field of machine learning (ML) (e.g., neural networks). Especially the approaches from the categories numerics and ML regularly lead even powerful digital infrastructures to their limits since they are still based on conventional semiconductor computers and their technical functionality. The resulting lengthy computation times and/or erroneous computation results still hamper the transfer of complete Industrie 4.0 framework models to industry today. Initial investigations show that quantum mechanical functional principles promise decisive advantages in the solution of numerous algorithmic problems, such as, above all, significant accelerations in numerical procedures and improvements in results through “quantum machine learning”-based approaches.

The project addresses the performance of near-term quantum computers and, building on this, develops scenarios for the medium-term goal of quantum advantage for practical applications. For this purpose, the link to hardware development and analysis, among others, through the participation of Forschungszentrum Jülich (FZJ), is central, as this allows hardware capabilities and limitations to be taken into account and the potential of quantum computers to be exploited at an early stage. QUASIM bundles competence from industry and academia. Innovative services are pushed here, which also enable the early use of quantum computers due to their high value. The implementation approach is based on existing interfaces to guarantee transferability and hardware agnosticism. Longer-term goals of the project are the development and testing of quantum computing algorithms and technologies for critical simulation issues in manufacturing, the methodical embedding in Industry 4.0 frameworks as “Quantum-as-a-Service” (QaaS), and the practice-oriented knowledge transfer for production-oriented simulation based on QC.

Besides DFKI, which acts as coordinator, the Fraunhofer Institute for Production Technology (IPT), the technology company Trumpf, the Research Center Jülich, and the software component provider ModuleWorks are involved in QUASIM. In addition, associated partners Ford and MTU Aero Engines are supporting the project.