Systematic Predictive Modeling of Moisture Dynamics for Condition Assessment of Paper-Liquid Insulation System of Transformer

In this study, we present a computationally efficient model for predicting moisture dynamics in paper-liquid insulation systems of transformers, aimed at enhancing real-time condition assessment and operational reliability. Moisture migration between cellulose insulation and dielectric liquids (mineral oil, natural ester, synthetic ester) is governed by temperature-dependent diffusion processes, which significantly influence insulation aging and dielectric performance. The proposed model applies Fick's second law with a diffusion coefficient dependent on moisture content, temperature, and insulation thickness. The governing equations are solved using an implicit finite difference scheme combined with the Tridiagonal Matrix Algorithm (TDMA), ensuring numerical stability and computational speed. Moisture equilibrium at the cellulose-liquid interface is modeled thermodynamically, assuming a constant total water content and instantaneous equilibrium. Three dynamic loading scenarios-a heat run test, IEEE C57.91-2011 Annex G, and IEC 60076-7 profiles-demonstrate the model's accuracy and adaptability. Results show that synthetic ester exhibits superior moisture absorption, leading to enhanced drying of cellulose insulation. Due to its simplicity and computational efficiency, the model is well-suited for integration into online monitoring systems and digital twin frameworks for transformer asset management.