Preparation of Papers for IFAC Conferences & Symposia: Soft sensor control for twin roll casting processes of magnesium alloys

Increasing lightweight constructions ensures a growing interest in economically produced magnesium sheets. The process of magnesium twin roll casting makes this possible by combining several different process steps into one process. However, due to the harsh process conditions, it is not possible to directly record the relevant process parameters to control product properties. Therefore, in this article will be discussed, how twin roll casting process can be transformed into a property-controlled process with a soft sensor. The digital image of the process is created by usage of an extended layer model, developed as a digital twin for offline computation and whose calculation times are many times shorter than in the FEM method. This makes it possible to train the digital twin and use it for online controlling in the TRC process. As a result, from the sensor data on the roller surface, which in turn forms the computational core of the required soft sensor, conclusions can be drawn about the state of the material in the rolling gap.