Predictive modelling and computational evaluation of top electrode voltage in 50-ohm radio frequency heating systems for pulse flours

Abstract

Accurate prediction of top electrode voltage is critical for optimizing heating uniformity and energy efficiency in radio frequency (RF) heating systems. This study presents two analytical models and an inverse simulation framework for predicting top electrode voltage in a 50 Ω RF system applied to pea and pinto bean flours. Model 1 includes a transient heat accumulation term (−dT/dt), whereas Model 2 omits it. Comparative analysis showed that Model 2 predicted voltages closer to simulation outputs and improved the electric field strength alignment. MATLAB-based inverse simulations, incorporating finite difference solutions of heat conduction with dielectric heating, were validated using experimental fibre optic temperature data, achieving high correlation (R2 = 0.99, RMSE <5). Pinto flour exhibited higher peak temperatures (∼128 °C) and broader heat dissipation than pea flour (∼114 °C), highlighting the influence of material-specific dielectric and thermal properties. Results underscore the need to exclude transient heat terms in voltage models and emphasize the importance of accurate dielectric profiling for each pulse type. While inverse simulation proved effective for voltage calibration, its broader industrial applicability remains limited by model assumptions, numerical stability, and sensitivity to experimental noise. These findings support the integration of simulation-based prediction tools into RF system design and point toward future improvements through real-time dielectric tracking and data-driven modelling for enhanced process control.