Understanding Ageing Condition of Epoxy Micro Nanocomposites Using Space Charge Characteristics and by Adopting PCA and ANN Analysis

Epoxy micro-nanocomposites are subjected to gamma irradiation for a dosage of 8 kGy and water ageing at
$\mathbf{90^{\circ}\mathrm{C}}$
for a period of 240 hours respectively, to study the characteristic variation in their space charge behavior after ageing. Exposing to gamma irradiation causes structural changes in the material through radiation induced degradation reactions. Water ageing at high temperatures helps in understanding the effect of long-term water intake phenomenon. Homo-charge accumulation near the electrode-specimen interface of all the test specimens, is noticed. A significant increment in the magnitude of space charge density, is noticed in aged specimens compared to unaged specimen. The gamma irradiated and water aged specimens are classified from unaged specimens by adopting principal component analysis (PCA) and artificial neural network (ANN) analysis to space charge characteristics. The number of hidden layer neurons in the neural network architecture are optimized based on classification accuracy and number of epochs required to converge. Artificial neural network analysis as well as principal component analysis reflected good classification accuracy in classifying the unaged, gamma irradiated and water aged epoxy micro-nanocomposites.