Impact of Reactive Power Assumptions on Physics-Informed Assessment of Transformer Ageing in Distribution Networks

Abstract

Power distribution networks are undergoing a fundamental shift in their utilisation driven by the rapid increase in the prevalence of residential photovoltaic (PV) systems, coupled with the increasing penetration of inverter-based household loads (such as heat pumps), changing expected reactive power flows. However, many Distribution System Operators (DSOs) have limited reactive power monitoring and instead rely on historical constant power factor assumptions, masking the true transformer current and thermal stress, resulting in biased lifetime estimates. This paper proposes an assessment methodology that integrates a low-voltage distribution network model with the physics-based IEEE C57.91 transformer thermal-ageing model to quantify transformer loss of life. The methodology evaluates different LCT penetration levels using representative load and reactive-power profiles. The methodology is demonstrated using a spatially coherent real-world distribution network model, derived from geographical network data, with PV generation and residential load profiles. We conduct a comparative analysis between the standard DSO assumption (constant power factor) and the actual reactive power profile. The results indicate that the assumption of a constant power factor of 0.95 significantly over/underestimates total current and thermal stress, resulting in biased transformer health assessments. The study demonstrates that a revised reactive power assumption is an important consideration for transformer health assessment and asset-management decisions.