Neural Networks Approaches to Determine Electrical and Rheological Properties of Polymer High Voltage Insulators

Abstract

Low-density polyethylene and magnesia composite is famous and has been under scientific investigation recently where it proved that it could provide premium material enhancement to the cable’s insulation. This study presents an attempt to examine the morphology and key electro-mechanical features for LDPE doped with MgO in a marketable vision that links material grade and processing expenses to the quality of the final product. Due to the shortage of the empirical outcomes resulting from the cost impact of the materials consumed and equipment employed during the experiments; also, the difficulties confronted in the laboratories, an approach of developing an artificial neural network that is capable of forecasting the dielectric strength of composites of any filler doping value or submerged in salty medium was found to be fruitful. 7.2. Conclusions Electro-mechanical and rheological properties of different weight loading for the nano and modified nano magnesia compounded to low density polyethylene that had been tested under dry, wet, and wet with different salinity levels media impacts conclusions are summed up as follows: a) Neat LDPE shows the lowest dielectric strength in all concentrations, which clearly indicates that doping LDPE with Nano-MgO, resulting in significant modification of its insulation properties. b) Dielectric strength behavior in all tested cases is similar. Below filler concentrations of about 1.5%, dielectric strength is increasing, while after this value, it decays exponentially till it saturates. c) Salinity value impacting in an inverse proportional manner to the dielectric strength of the LDPE/MgO nanocomposite. d) volume resistivity increased by almost one order of magnitude with the addition of filler up to 1.5% in nano scale filler concentrations.