Data-Intensive Capacity Assessment and Overload Risk Mitigation of Distribution Transformers Under High Renewable Integration

Aksöz A., Şeker M., Şafak İ.

IEEE TRANSACTIONS ON POWER DELIVERY, ss.1-10, 2026 (SCI-Expanded)

Özet

Distributed generation reverses and reshapes the loading of distribution transformers, yet capacity planning still compares an annual peak with a static rating. This paper fuses the 275-substation inventory of a German distribution system operator with one year of 10-min telemetry from seven transformer nodes and reanalysis ambient temperature. Loading is decomposed by seasonal-trend decomposition after a Box Cox transformation, a moving block bootstrap generates 1000 realizations per node, and each realization is propagated through the thermal models of IEEE Std C57.91-2025 and IEC 60076-7 with explicit reverse-flow terms. A dual criterion, a nameplate exceedance probability of at most 2% together with a 120°C hot spot limit and an equivalent aging factor of at most one, defines the admissible growth factor of each transformer. At the reverse flow node, where biomass cogeneration exports up to 478.8kW through 1250kVA, loading peaks at 38.4% and the insulation consumes less than one hour of life per year. The admissible growth factors range from 1.26 to 4.29, and a deterministic thermal assessment on the smoothed profile overstates them by 12% to 114%. The rated hot-spot gradient and top-oil rise dominate the aging sensitivity, while the reverse-mode bounds raise aging by at most 65%.