|dc.description.abstract||During the course of a transformer's operational life, it experiences many challenges to its insulation. Chemical stresses, such as moisture in insulating oil, deteriorate insulation chemically. Mechanical stresses and thermal stresses, such as those experienced during a large external fault, also contribute to weakening a transformer's insulation. Once these stresses have weakened the insulation to the point of breakdown, a fault may occur between windings of the transformer. These faults, known as turn-to-turn faults, are difficult to detect electrically at the terminals of the transformer until they have grown to the point of damaging the transformer beyond repair.
Current differential transformer protection is a simple, reliable, and cost effect method of detecting turn-to-turn faults. This method of protection is only able to detect faults involving 10% of the windings or more. The sensitivity of current differential protection is limited as not to cause false tripping due to normal imbalances in current. Such imbalances in current may occur when a tapchanger is used to increase or decrease the voltage on one side of the transformer.
Digital current differential relays, which monitor tap changer position, compensate for current imbalances due to tapchanger operation. Other causes of current imbalance include current transformer saturation, magnetizing inrush current, and over-excitation. A transformer is designed to operate continuously at 10% above its rated voltage. In this overexcited state, a current imbalance appears which causes a differential current to be sensed by the current differential relay. This limits the current differential relay's sensitivity as it must be designed to ignore current imbalances due to the aforementioned causes.
The current differential transformer protection algorithm, which only makes use of the current magnitude, is based on the principles of an electromechanical relay. Digital relays are capable of computing the negative sequence current on both primary and secondary sides of the transformer along with the phase difference between these two negative sequence currents. By using both phase and magnitude information, negative sequence current could be used to detect turn-to-turn faults involving 3% of the transformer's windings or more.
Turn-to-turn faults may still occur even if no current is flowing on one side of the transformer, such as during energization. With no current flowing in the secondary windings of the transformer, negative sequence current based algorithms become insensitive. A transformer is particularly likely to experience a turn-to-turn fault during this time due to the stresses of energization. This thesis introduces a relay prototype, using both negative sequence current and negative sequence voltage, which retains its sensitivity during energization.
This prototype was constructed using a micro-controller and an analog-to-digital conversion board. The transformer protection relay algorithm, including all hardware interface code and signal processing code, was then designed to suit the prototype's hardware. A 3-phase transformer real-time simulation model, capable of simulating turn-to-turn faults as well as the magnetic properties of the transformer's core, was developed. The voltage and current waveforms generated by the transformer model, running in the real-time simulator, were used to test the relay prototype.
The sensitivity and speed of the relay algorithm proposed in this thesis was then tested, for faults involving 1%, 3%, 5%, 10%, 15%, and 25% of the windings along with two commonly encountered transformer winding configurations. The relay's performance for several commonly encountered system scenarios such as over-excitation, current transformer saturation, non-zero fault resistance, transformer energization, and external faults were also examined for several turn-to-turn fault severities. A fault resistance of one Ohm is typical for transformer turn-to-turn faults. These test results were compared to current differential protection with second harmonic restraint. The experimental results presented in this work indicate that the algorithm proposed in this thesis is faster and more sensitive than restrained current differential protection and capable of detecting turn-to-turn faults occurring during transformer energization.||