Ground-fault Protection for DC Systems
Direct current (DC) systems have positive and negative buses. If either bus is intentionally grounded, then it is referred to as a grounded system. If neither bus is grounded, then it is referred to as an ungrounded DC system. A ground fault on a DC system may cause damage to the source as well as in the field. If the system is ungrounded, then it is possible to use a ground-fault relay by installing a ground-reference module between the two buses to establish a neutral point (see Figure 1). The ground-fault relay uses this reference to create a path for fault current to flow, which can then be measured. The ground-reference module is selected to limit fault currents to a very low value.
Figure 1.
Ground-fault Protection for AC Systems
Ungrounded AC systems, as shown in Figure 2, were used where continuity of power was critical. For example, chemical plants or refineries involving processes that cannot be interrupted without extensive dollar or product loss may have an ungrounded system. However, experience has proven that these systems are problematic and are being replaced with resistance grounded systems. Two major problems with ungrounded systems are transient overvoltages and difficulty locating ground faults.
Figure 2.
An ungrounded system has no point in the system that is intentionally the non-current-carrying metal parts to ground. Grounding occurs only through system capacitance to ground (as shown in Figure 2). Continuity of operation occurs because the system can operate with one phase faulted to ground. An intermittent or arcing fault can produce high transient overvoltages to ground. These voltages are impressed on the phase conductors throughout the system until the insulation at the weakest breaks down. This breakdown can occur at any point in the electrical system, causing a phase-to-ground-to-phase fault. Although a ground fault can be detected or alarmed on the system, it is difficult to determine the location of the fault.
There are two methods used to detect ground faults on ungrounded systems. One method is to monitor the voltages between the phases and ground. As a ground fault develops, the faulted phase will collapse to ground potential, causing an indicator light to dim. The indicator lights on the unfaulted phases become brighter. A second method to detect a ground fault is to measure the insulation resistance. As the insulation deteriorates, a relay continuously monitoring the insulation resistance can alarm at different levels for predictive maintenance. A visual indicator or meter can also be used.
Solidly Grounded Systems
Due to the problem of ungrounded systems, a shift in philosophy occurred and designs moved from ungrounded to grounded systems. In most cases, the type of grounding system chosen was solidly grounded. A solidly grounded system is a system of conductors in which at least one conductor or point is intentionally grounded (usually the neutral point of transformer or generator windings). The problem with the direct connection is that ground-fault current can be excessive, causing Arc-Flash hazards, extensive equipment damage, and possible injury to personnel. A solidly grounded system cannot continue to operate with a ground fault.
Figure 3.
In a solidly grounded system, the wye point (or neutral) of the power source is connected solidly to ground and offers a very stable system that maintains a fixed phase-to-ground voltage.
The high ground-fault current is easy to detect with fuses, circuit breakers, or protection relays, or a combination thereof, allowing for selective tripping (tripping the faulted feeder and not the main feeder). When a ground fault occurs, high point-of-fault damage can quickly result since the energy available to the ground fault is only limited by the system impedance (which is typically very low). Due to excessive ground-fault current and Arc-Flash Hazards, the faulted feeder must be removed from service. This does not allow for continuous operation during a ground fault.
Figure 4.
Figure 4 illustrates an example of the dangers associated with solidly grounded systems. In this example, a ground fault occurs and the overcurrent protection is set at 600 A.
Assume that this ground-fault is not a bolted fault, but an arcing fault due to an insulation breakdown or a partial reduction of clearances between the line and ground.
Because of the arc resistance, fault current may be as low as 38% of the bolted-fault level. This can be in the range of a normal load or a slight overload. The fault current may be low enough that the overcurrent device (600-A circuit breaker) does not sense a fault, or may pick it up but not trip for a long time. The energy being supplied by the source is concentrated at the arc and could cause severe equipment damage very quickly.
This energy release could cause a fire that in turn, could damage the premises and present an extreme hazard to personnel. Aside from converting this solidly grounded system to resistance grounding, the best way to prevent damage is to detect low-level ground leakage prior to it becoming a ground fault. In order to accomplish this, the protection relay must be able to sense a low-level ground leakage without nuisance tripping.
In modern facilities, equipment often generates noise or harmonics that can interfere with a protection relay's ability to function properly. For example, the noise or harmonics may be higher than the desired ground-fault relay settings, causing the relay to falsely operate when there is no fault on the system. The protection relay must be able to filter out noise or harmonics to provide reliable protection.
Resistance-Grounded Systems
Resistance grounding solves the problems commonly associated with both ungrounded systems and solidly grounded systems. The name is derived from the addition of a resistor between the system neutral and ground (as shown in Figure 5). The specifications of the resistor are user determined to achieve a desired ground-fault current, which must be greater than the system capacitive charging current (explained later in this section).
Figure 5.
Transient overvoltages can be eliminated by correctly sizing the neutral-grounding resistor (NGR) to provide an adequate discharge path for the system capacitance. Continuity of operation with one ground fault is typically allowable when ground-fault current is <− 10 A. The NGR limits the available ground-fault current. This eliminates or minimizes point-of-fault damage (Arc-Flash Hazards) and controls the ground-fault voltage. Pulsing current can be used to locate ground faults when ground-fault current is <−10 A. Pulsing current is created by using a shorting contactor to short out half of the resistance, causing the ground-fault current to double (usually one cycle per second). A hand-held zero-sequence meter is used to detect the fluctuating ground-fault current, and locate the ground fault.
The only disadvantage of resistance grounding is that if the resistor fails, the system will become ungrounded. Resistor monitoring is recommended to protect against this. A protection relay for resistance-grounded systems is used to detect a ground fault and to monitor the neutral-to-ground connection. It can be used to provide alarms or to trip the feeder from service upon the detection of a ground fault.
The relay can provide a pulsing circuit that can be used to locate the ground fault. The relay can also alarm or trip if the neutral-to-ground path fails. For systems 5 kV and less, high resistance grounding can be used. High-resistance grounding typically limits the resistor current to 10 A or less. By doing so, the ground fault can remain on the system, given that the system is rated for the voltage shift. For systems above 5 kV, neutral-grounding resistors are typically rated for 25 A or more, and ground-fault current is cleared within 10 s
System Capacitive Charging Current
Although not physically connected to ground, electrical conductors and the windings of all components are capacitively connected to ground. Consequently, a small current will flow to ground from each phase. This current does not occur at any particular location; rather, it is distributed throughout the system just as the capacitance to ground is distributed throughout the system. For analysis,it is convenient to consider the distributed capacitance as lumped capacitance, as shown in Figures 6 and 7.
Figure 6.
Even if the distributed capacitance is not balanced, the ammeter will read zero because all the current flowing through the CT window must return through the CT window. System charging current is the current that will flow into the grounding connection when one phase of an ungrounded system is faulted to ground (see Figure 9). It can be measured as shown below if appropriate precautions are taken:
▪ If the fault occurs on the supply side of the CT, the sum of the currents in the CT window is not zero.
▪ Ammeter A will read the sum of the capacitive currents in the unfaulted phases. This value is the charging current of all the equipment on the load side of the CT.
Figure 7.
A single-line diagram of a three-feeder, resistance-grounded system with a fault on feeder 3 is shown in Figure 8. A CT (A1 and A2) on unfaulted feeders will detect the charging current of that feeder. A CT (A3) on a faulted feeder will detect the sum of the resistor current (IR) and the charging currents (I1 +I2) of the unfaulted feeders.
Figure 8.