Title :
Neural networks for estimating electrical characteristics of the surface of the Earth
Author :
Custódio, A. C S ; Cavalcante, G.P.S.
Author_Institution :
Dept. de Engenharia Eletrica, Univ. Fed. do Para, Brazil
Abstract :
The objective of this work is to utilize neural network resources to determine the electrical characteristics of the ground. The data processing of the input and output is done by calculating the field strength of the surface wave, using the Bremmer and Van der Pol equations, to train a neural network simulator. From the results of the measurement of electric field strength of the outskirts of Belem and rural area, using radio stations with medium wave as transmitters, the following values of the ground parameters were found: 1.4 mS/m for conductivity and 15 for relative permittivity
Keywords :
electric field measurement; electrical conductivity; feedforward neural nets; geophysics computing; land mobile radio; learning (artificial intelligence); permittivity; radiowave propagation; telecommunication computing; terrestrial electricity; 1.4 mS/m; Bremmer equation; Earth surface; FIELD computer program; Van der Pol equation; conductivity; data processing; electric field strength; electrical characteristics; ground parameters; neural network simulator; neural network training; radio stations; relative permittivity; surface wave attenuation; surface wave field strength; vegetation; Area measurement; Conductivity measurement; Data processing; Electric variables; Electric variables measurement; Equations; Neural networks; Permittivity measurement; Radio transmitters; Surface waves;
Conference_Titel :
Mathematical Methods in Electromagnetic Theory, 1996., 6th International Conference on
Conference_Location :
Lviv
Print_ISBN :
0-7803-3291-1
DOI :
10.1109/MMET.1996.565684
