Benefits of the Pulse Injection Method of Ground Fault Detection
Modeling a Ground Fault for a Generator or Motor with Brushless Exciter
Figure 1 - Field Ground Fault on a Generator or Motor with Brushless Exciter
A field ground fault can occur when insulation damage creates an electrically conductive path from anywhere on the field winding (or elsewhere on the excitation circuit) to the rotor forging. In this diagram, we represent those faults as resistors. If two faults occur at different locations along the winding, excitation current is diverted through the forging, and severe burning of both insulation and rotor steel can occur rather quickly.
Two Methods of Ground Fault Detection
Figure 2 - Classical DC Voltage Injection Method
Figure 3 - Advanced Pulse Injection Method
Most field ground detectors in service today use the classical DC Voltage Injection Method (figure 2). Notice the excitation circuit with an exciter that supplies current to the field winding. When a ground fault occurs, it is represented by a resistance (RL) from some location on the winding to the rotor ground. The voltage potential at the fault location relative to the negative field terminal is K*VX, where K is called the location factor, and VX is the excitation voltage. K takes a value between 0 and 1 for locations between the negative terminal and the positive terminal. The ground detector circuit makes a connection between the rotor ground and the excitation circuit, usually on the negative terminal as shown here. A small DC voltage is injected across this connection, and the current flow through the circuit is measured. If a ground fault, RL, occurs, a current will flow through the ground detector circuit. When that current exceeds a fixed current threshold, a ground fault alarm is triggered. The injection voltage insures that some current will flow when a ground fault occurs, even if the fault occurs at the negative excitation terminal. Faults occurring closer to the positive terminal along the field winding will occur at higher fault resistance. In other words, the severity of fault that will cause an alarm is dependent on the fault location.
The above plot demonstrates a limitation of classical detection methods. The three curves on this plot show the sensitivity at the alarm point: for a fault at the negative terminal in blue, the midpoint of the field in red, and the positive terminal in green. You see that for a field voltage of around 300 volts, the severity of the fault at the alarm point will vary by more than an order of magnitude.