Modeling of Microwave Heating Systems with Octagonal Tube Cavities: A Comparative Study of Fuzzy-Based and ARX Approaches

Abstract

In the quest to design a robust model for microwave heating systems with symmetrical octagonal tube cavities (MWHSO), a fuzzy-based approach, specifically the Takagi Sugeno Fuzzy Model, was explored to capture the dynamics of the heating process. To achieve this, the mathematical model was adaptively adjusted according to varying input conditions through the utilization of fuzzy logic. Input data were sourced from two magnetrons, with the system outputs derived from measurements acquired from five temperature sensors placed on the heated object. For performance evaluation, the Root Mean Square Error (RMSE) was employed. A comparison was drawn with the autoregressive model with exogenous variable (ARX), a traditional approach wherein the system's mathematical model remains static. Simulation studies were conducted, treating every probe measurement across all dataset validations as distinct cases. It was found that the T-S Fuzzy model surpassed the ARX40 in performance in 33 of the total cases, accounting for 92.49%. The most notable performance of the fuzzy-based approach was observed at a 180-Watt power input, recording an average RMSE of 0.00574 across the five sensors. In contrast, the ARX-based model registered an RMSE of 0.01256. These findings suggest that the fuzzy-based modeling approach presents a compelling alternative for representing the dynamic heating processes within MWHSO.