Özet:

This study has shown that the real and virtual components (e' and e'') of the complex dielectrial function of low-intensity polyethylene (DYPE) can be highly accurately predicted by the method of generalized regression nerve networks (GRSA) both to the contribution of polyanylene (PANI) and temperature-related changes. For this first, pure DYPE and mass 0.7, 1% and 3% PANI added DYPE/PANI compounds are prepared and the frequency-related changes of the components of e' and e' at 20 °C, 50 °C and 80 °C of the corresponding samples are experimentally determined by the method of dielektric spectroscopy. Then, with the help of graphs drawn according to the actual values of the predictive values of the dielectrial parameters, the success performance in the determination of the relevant parameters of the GRSA model was identified as Re' = 0,9998 and Re' = 0,9365. Moving from this point, the GRSA model was previously used to predict the changes of the components e' and e' depending on the frequency of the existing samples at 35 °C, 65 °C and 95 °C. Then, for the two fake composites that have never been experimentally produced (1,5% and 6% PANI contributed to DYPE) on 20 °C, 35 °C, 50 °C, 65 °C, 80 °C and 95 °C, the frequency-related changes of the components e and e' were recommended by the GRSA method. Thus, the dielectric parameters of these samples which have never been experimentally produced can be determined depending on temperature and frequency.

Anahtar Kelimeler: