Data-Driven Control of Thermoforming Machines

Thermoforming processes are widely used in plastics manufacturing and involve heating a plastic sheet to give it a specific shape. Changing the plastic material and the operating points in which the thermoforming machine operates requires an adjustment of its parameters so that the desired plastic piece is obtained and no failures in form occur. In particular, the heating phase parameters are the most critical, as the power and heating time have a significant impact on the final quality of the piece. Parameter tuning is usually done by an operator and often results in a waste of time, energy, and resources that should be minimized. This work investigates two sequential methods based on input-output data to optimize machine parameters that lead to the desired plastic form. Specifically, we propose two data-driven control stages for the thermoforming tuning problem. Both approaches have been validated in simulation, showing their ability to optimize the heating phase parameters to obtain the desired plastic shape.