A data-driven approach to control of batch processes with an application to a gravimetric blender

In this paper, a data-driven controller design approach for batch processes is proposed. In this strategy, based on a suitable transformation of the input and output signals, a mathematical description of the process dynamics is not needed, and the working cycle is guaranteed to adapt to the desired operating conditions. Throughout this paper, all the steps of the above algorithm are described in detail with the help of an experimental case study dedicated to a gravimetric blender, which is a key element in the plastic extrusion process.