Design and development of a test platform for a power transformer using biodegradable dielectric liquid cooling

This paper presents the design, construction, and validation of a cost-effective laboratory platform that reproduces the heating behaviour of a 100 MVA power transformer to locate and quantify hot spots in the low-voltage windings. After reviewing existing setups, several concepts were assessed and the selected design, replicating 15% of the LV winding in a closed loop with transformer-type radiators, was modelled with Computational Fluid Dynamics (CFD). The CFD predictions were validated against heat-run data from the reference transformer filled with mineral oil, yielding an average hotspot difference of 2.0 K (maximum 2.7 K). The platform was then built and tested with mineral oil and a natural ester. Platform–transformer comparisons confirmed close agreement: with mineral oil, average hotspot temperatures differed by 2 K relative to the transformer test; with natural ester, platform results differed by 0.4 K relative to the transformer simulation. Overall, the platform provides a reliable and scalable tool for predicting transformer thermal behaviour and identifying critical temperature locations, and it reproduces the higher hotspots and steeper oil temperature gradients observed with natural ester due to its thermo-physical properties.