Academic Journals Database
Disseminating quality controlled scientific knowledge

Efficient Characterization of Microwave Applicator Loaded with Multilayer Dielectric based on Neural Networks

Author(s): Zoran Stanković | Bratislav Milovanović, | Nebojša Dončov | Aleksandar Marinčić

Journal: Facta Universitatis Series : Electronics and Energetics
ISSN 0353-3670

Volume: 23;
Issue: 2;
Start page: 217;
Date: 2010;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

Keywords: Microwave cavity | multilayer dielectric | neural network modeling | resonant frequency

In this paper, new neural network approach for characterization of cylindrical metallic cavity loaded with multilayer dielectric is suggested. Such cavity configuration has a great importance in realization of microwave resonant applicators, widely used for thermal processing of multilayer materials. Proposed neural model is based on representation of multilayer cavity load via several planparallel homogeneous dielectric slabs and superposition of separate influence of each dielectric slab parameters on the cavity resonant frequency. Model is incorporated into MLP neural network enabling its efficient implementation on modest hardware platform. Suggested approach is verified on the example of circular cylindrical metallic cavity loaded with two-layer dielectric.

Tango Jona
Tangokurs Rapperswil-Jona

     Save time & money - Smart Internet Solutions