Advanced Control System for the Optimization of Thermal Processes in a Coffee Capsule Production Machine

This thesis develops and validates advanced control strategies through an industrial case study conducted at CEC-Coesia Engineering Center. The work begins with an analysis of an existing coffee capsule production machine, followed by product characterization, reverse engineering and system modeling. These activities provide the basis for designing Gain-Scheduled Linear Quadratic Regulator (LQR) and Gain-Scheduled Model Predictive Control (MPC) strategies to improve the performance of the thermoforming and sealing processes. The objective is to enhance temperature regulation and support format changeovers, considering both standard and compostable capsule materials. The controllers are initially developed in Python, and subsequently implemented on a PLC to enable experimental testing. These tests assess the effectiveness and the precision of advanced control strategies, especially dealing with different material and temperature conditions, closing the loop between the model-based design and an industrial deployment.