Journal of Microwave Power and Electromagnetic Energy (JMPEE)
TITLE	A Finite Difference Thermal Model of a Cylindrical Microwave Heating Applicator Using Locally Conformal Overlapping Grids: Part II - Numerical Results and Experimental Evaluation [PDF]
AUTHORS	H. M. Al-Rizzo, R. Adada, J. M. Tranquilla, F. Ma and B. C. Ionescu 2005 40 2 78-100
YEAR
VOLUME
ISSUE
PAGES
Abstract In this paper, we present numerical results obtained from a robust, locally conformal 3-D Orthogonal Grid Finite Difference (OGFD) thermal algorithm introduced in Part I of our current investigation [Al-Rizzo et al., 2006] integrated with an Orthogonal Grid Finite-Difference Time Domain (OGFDTD) scheme [Al-Rizzo et al., 2000], which accurately models the volumetric electromagnetic (EM) power deposition pattern. A unified meshing scheme, which utilizes identical overlapping grids in Cartesian and cylindrical coordinates, is employed within the load zone in the OGFDTD and OGFD models. Local temperature profiles excited by the absorbed microwave energy were measured at seven locations within the sample as a function of heating time. In order to benchmark, or validate our model, an alternative analysis of the coupled EM and thermal simulations was performed using state-of-the-art, Finite Element Method-based Ansoft’s High Frequency Structure Simulator (HFSS) and the coupled thermal/stress analysis tool ePHYSICS (http://www.ansoft.com). Additionally, we compare our numerical simulations against measured dynamic temperature profiles induced within a mineral ore sample maintained for exposure period of 28.5 minutes inside a cylindrical multimode heating furnace energized at 915 MHz with a microwave source power of 12.5 kW and accompanied with significant temperature elevation. A combination of convective and radiation thermal boundary conditions are considered at the interfaces between the cavity walls, air, and sample. There is a general agreement between simulated and measured spatial and temporal temperature profiles, which validates the proposed model. Results indicate that inevitable fluctuations in the frequency spectrum and output power of the magnetron, non-uniformity of sample packing, and heat released by uncontrolled exothermic chemical reactions have a significant effect on the comparisons between measured and computed temperature patterns.