Conference Agenda

This study presents a modelling framework for analyzing moisture migration and dielectric behavior in cellulose insulation impregnated with mineral oil, natural ester, and synthetic ester liquids. Using the dynamic transformer moisture modelling, transient moisture transport is simulated through moisture and temperature dependent diffusion, sorption isotherms, and oil–cellulose exchange over a thickness range of 0.1–10 mm and moisture conditions typical of service. Heat run temperature profiles reveal optimal thicknesses that maximize drying, with synthetic ester producing the strongest drying, natural ester moderate, and mineral oil the weakest. Dielectric analysis shows permittivity rising with moisture and varying across liquids, shifting electric field distribution and losses. The results emphasize the importance of incorporating both thickness and liquid dependent moisture behavior and dielectric properties in transformer condition assessment

India’s grid is undergoing a major transformation driven by renewable energy integration, rapid urbanization, and increasingly variable operating conditions. Transformers, the backbone of the transmission network, have traditionally relied on static nameplate ratings, a practice that often results in thermal stress during peak demand and underutilization during favorable ambient conditions. This limitation underscores the need for a dynamic, condition-aware approach to transformer loading. Adaptive Transformer Rating (ATR) introduces an intelligent, real-time methodology to assess transformer loading capacity based on actual operating conditions. By continuously analyzing parameters such as load current, ambient temperature, oil and winding temperatures, and cooling performance, ATR applies IEEE and IEC thermal models to estimate hot-spot temperature, aging acceleration factor, and residual life. This enables users to safely operate transformers beyond nameplate limits when conditions permit, improving asset utilization and operational flexibility.

This paper presents the ATR methodology, POWERGRID’s enterprise architecture for implementation, validation strategy, operational governance, and lessons learned from early deployments.

The cooling system design of power transformers significantly influences their service life, as the hot-spot temperature directly affects insulation aging. This work focuses on the Oil Directed (OD) flow design. Improved thermal management (i.e., a reduction in hot-spot temperature) is achieved by using narrowing rings (barriers) that force the oil to flow horizontally through the winding discs in a zig-zag pattern. This paper uncovers that the number of barriers is critical: using too many creates excessive hydraulic impedance, restricts total oil flow, and hinders effective heat dissipation, while using too few reduces cooling efficiency. The position of the barriers is decisive in ensuring cooling in locations where losses are highest.

An optimization analysis is used to determine the optimal geometric configuration of the barriers, minimizing the maximum steady-state hot-spot temperature. To ensure this improvement without increasing material costs, the optimization is performed while maintaining the same number of barriers as in the original transformer design. This demonstrates that the enhancement of cooling efficiency can be achieved solely by varying their positions. The problem is formulated as a Mixed-Integer Non-Linear Optimization (MINLP), with an objective function derived from a coupled multiphysics Finite Element Method (FEM) simulation. The loss distribution under load conditions is determined through an electromagnetic field (EMF) simulation, integrated into a Computational Fluid Dynamics (CFD) model coupled with a Conjugate Heat Transfer (CHT) study. This CFD/CHT model can accurately resolve flow and heat transfer phenomena, including localized effects such as the risk of stagnant or reverse flow, which is critical in OD-type cooling.

Given the high computational demand of CFD-CHT simulations, simplifications are considered to reduce the model size while maintaining a good physical representation. The model is formulated as a 2D axially symmetric model, a necessary simplification due to the significant computational effort required. The optimum is achieved within the design constraints and model simplifications. The relative temperature decrease observed in the 2D model should be maintained in a more complex design (3D), although the final absolute values may diverge. Despite these limitations, Bayesian Optimization (BO) is used in the design process, leveraging surrogate models to efficiently explore the complex search space and reduce the computational effort required to reach the optimal solution.

Natural ester liquids have emerged as a sustainable alternative to mineral oil, offering significant advantages such as a higher fire point and reduced flammability during fault conditions. Derived from renewable seed-based sources rather than fossil fuels, natural ester liquids have gained popularity in transformer applications. However, with just over 30 years of industry use—compared to over a century for mineral oil—questions about their performance persist. This paper addresses common myths and clarifies the facts surrounding the performance characteristics of natural ester liquids.

Mineral transformer oil (MTO) is widely used as an insulating and cooling medium in electrical transformers due to its excellent dielectric properties and thermal stability. However, MTO can undergo combustion under high thermal stress or fault conditions and posing significant fire hazards in electrical infrastructure. This study aims to comprehensively analyze the combustion behavior of MTO with a focus on the evolution of CO₂ and CO gas concentrations, the fire growth index (FGI), and burn duration under different thermal loads. The objective is to gain insights into the fire hazard potential of MTO and essential for enhancing fire safety protocols, guiding material selection, and improving thermal management strategies in transformer applications to reduce the risk of catastrophic fires in electrical systems

Experiments were conducted on transformer oil contaminated with cellulose particles using a sphere–sphere electrode configuration under DC electric stress (2 kV, 7.5 kV, and 15 kV). The results show that particle behavior follows a stage-wise progression consisting of polarization, field-driven migration, electrode accumulation, and pre-bridge formation. Higher electric stress significantly accelerates particle transport and reduces the time required for conductive pathway initiation. A clear correlation between particle accumulation and conduction current was observed. A physics-based simulation model is also presented to describe early-stage particle migration under non-uniform electric fields. The simulation results show good qualitative agreement with experimental observations, particularly in demonstrating field-dependent particle movement toward high-field regions.