Iterative learning control for vacuum heat treatment process

Distributed parameter systems constitute an important class of modern industrial processes. However, in many practical applications the engineers still tend to adapt some classical control techniques developed for lumped systems totally neglecting the spatial dynamics of the investigated process. In a view of increasing demands imposed on system accuracy and performance, such conventional control algorithms simply become insufficient and there is a great necessity for novel identification and control methods taking into account both the temporal and spatial dynamics. This work reports a dedicated approach to control design for repetitive thermal processes consisting of the extension of the existing feedback control scheme with an intelligent data-driven component using the iterative learning control technique. Although this is a method which emerged in the context of time-invariant systems, it become adapted to more complex systems due to its flexibility and inherent robustness. The characterization of the resulting control scheme is discussed together with control design and implementation details. In order to compare the quality of the regulation, the approach is illustrated with simulation on the realistic model of wafer heating in an industrial vacuum furnace.