A Modular Python Framework for Multiphysics Simulation of Cookie Baking

This thesis presents the development of a modular Python framework for the multiphysics simulation of cookie baking, with the aim of reducing the dependence on repeated experimental trials and supporting process optimization. The model was implemented using FiPy and the Finite Volume Method, considering coupled phenomena such as heat transfer, moisture migration, evaporation, and surface browning. The simulation was developed on a two-dimensional cookie domain and organized into modules dedicated to input definition, the physics engine, the BCU quality model, and result validation. A mesh sensitivity analysis showed that approximately 5500 cells represent a good compromise between accuracy and computational cost. The model was validated using experimental data obtained by thermal-camera measurements at baking temperatures between 130 °C and 180 °C. The surface temperature prediction achieved R² values ranging from 0.7772 to 0.9761, with RMSE values between 3.5428 °C and 7.4644 °C. The best result was obtained at 155 °C, with R² = 0.9761 and an overall error of 2.22%. The BCU model, used to describe browning, showed R² values between 0.7662 and 0.9945 and overall errors between 0.07% and 1.08%. Overall, the developed framework demonstrated a good capability to simulate cookie baking and to predict both thermal evolution and quality progression, providing a useful basis for future applications to bakery products and for food process optimization.