Robust Digitalization of Manufacturing Applications

Purpose and goal:

The RoDi projects aim at improving the utilization and robustness of existing data streams by including the physical domain knowledge into the manufacturing data stream, i.e, combining traditional physics-based modelling and simulation with datadriven algorithms. That way, a robust and physically valid database is generated which enhances the decision-making process on the operation and maintenance of machines and factories.

The project goal is a generalized methodology for the inclusion of physical domain knowledge into manufacturing data streams. Expected results and effects:

The primary effect of the RoDi project is increased recourse utilisation of industrial systems and machinery/robots and its related data. RoDi narrows the gap between theoretical advances in AI and machine learning and its practical applications to physical systems. Approach and implementation:

The project runs over three years and brings together small, medium and large companies, and covers key competencies in several complementary areas of the life cycle of production systems, including supplier of data collection and information systems. The consortium is divided into three groups: academic partner KTH (integrated analytics and physics-based modeling); System integrator Nytt AB (machine learning and production follow-up); and the industrial partners Scania, LEAX and ABB are involved through strategically selected case studies (vehicle and industrial robots).