Reduced-Order Model Based Temperature Control of Induction Brazing Process

Accurate temperature control is an important part of an induction brazing process, because of a small difference between melting points of aluminum and soldering material. However, direct measurement of the local temperature at the spot of brazing is not possible in a production line. This temperature can be calculated from the results of indirect measurements and the physical model of the system. This paper shows a novel application of model order reduction in case of online temperature measurement. Three different controlling strategies are examined, the results proved that the more sophisticated, program controlled or mixed strategies need higher-order MOR models to accurately solve the problem. The presented, fourth degree meta-model provides sufficiently accurate results, which very fast and may be implemented into a micro-controller.