A methodology for experimentally verifying simulation models for distribution transformer internal faults

Abstract

Internal winding faults comprise 70-80% of modem transformer breakdown. In this era of deregulation, this phenomenon is likely to increase since loading transformers to their optimum capacity is becoming normal practice. These internal faults result from degradation of the transformer winding insulation, which tends to cause a breakdown in the dielectric strength. This breakdown either causes adjacent windings to short or a winding to be shorted to a grounded part of the transformer. Such faults can be very catastrophic and hence expensive. Utilities therefore welcome inexpensive methods employed to detect such faults in the incipient stage. The long-term objective of this research is the development of an inexpensive technique for the detection of transformer incipient winding faults. As part of this research, the thesis presents: 1. Internal winding models of single-phase, distribution transformers. These models are adapted from an earlier work of modeling internal winding faults of three-phase power transformers. They are compatible with the Alternative Transients Program and enable the transformer winding terminal parameters to be monitored. They allow the simulation of faults between any turn and the earth or between any two turns of the transformer windings. 2. Simulation of various internal winding faults of a single-phase distribution transformer using the models. 3. A general methodology to experimentally verify simulation models for distribution transformer internal winding faults including details of the design and layout of a field experimental setup containing a 25kVA, 7200V/240V/120V single-phase, custom-built transformer and a 25kW resistor load bank. 4. A comparison of the simulation and corresponding field experiment results. Although the simulation models neglected factors such as saturation and consequently transformer nonlinearities, the simulation and field results were very similar. As a contribution, the experimental setup presented in this work could generally be used for simulation model verification by following the proposed methodology with appropriate modifications. The validated models can be utilized to generate fault data for all kinds of scenarios including those that would be impossible to stage experimentally due to high levels of fault currents. These data can be used as a basis for a single-phase transformer incipient fault detection system.