Although it's one of the most important aspects of electrical design and installation, grounding continues to be one of the least understood and most misinterpreted concepts in the industry. It's also one of the most expensive when errors are made. The dollar value of equipment in-operation and/or loss — not to mention the potential liability — associated with ground-fault arcing can be staggering. So if grounding implementation problems leave you dazed and confused, don't worry. You're not alone.
Electrical contractors, plant/facility electrical maintenance personnel, and electrical engineers continue to demand more complete and concise information on grounding-related issues. That's why interest in grounding and ground faults has not diminished throughout the various Code cycles. Let's take a look at some of the most common questions that come up when you're contemplating the selection of grounding systems that address ground-fault currents. [Editor's Note: These are real questions submitted to “Ask the Experts” forum on the EC&M Web site (www.ecmweb.com) and answered by I-Gard's Tony Locker.]
Question No. 1: What are the advantages and disadvantages of the various grounding methods for medium-voltage systems in power plants? Also, what practices are adopted by electric utilities both nationally and internationally?
Answer: You can broadly classify medium-voltage (MV) grounding systems into four categories: solidly grounded, low-resistance grounded (LRG), high-resistance grounded (HRG), and insulated neutral (ungrounded) systems. A good reference is ANSI/IEEE Std. 242 (Buff Book), “Protection and Coordination of Industrial and Commercial Power Systems.”
With the solidly grounded system, as shown in Fig. 1, there is no intentional impedance in the neutral-to-earth path. Instead, the neutral is solidly connected to earth. This is why the term “earthing” is sometimes used in place of “grounding.” The phase-to-ground voltage remains constant during a ground fault, and there are very high fault current flows, which can result in extensive damage. The protective device closest to the fault must trip and isolate the circuit as fast as possible. If the fault is in a rotating machine, then there is a high possibility of core damage and replacement costs. The cost associated with the downtime also can be significant.
Fig. 2. Low-resistance grounding system showing phase-to-ground fault and ground-fault current path. Ground resistor limits the magnitude of the ground-fault current.
With LRG systems, as shown in Fig. 2, the ground-fault current is controlled and normally limited to between 25A and 1,000A. The voltage to ground on the un-faulted phases can increase up to the phase-to-phase voltage level, so you must use adequately rated insulation systems and surge suppression devices. You also must detect and isolate the ground fault. Since the ground-fault current is smaller and controlled, ground-fault relaying still has the requirement of fast tripping. However, you can achieve better time current coordination with this type of grounding system. Damage at the fault point is also reduced; therefore, maintenance and repair costs are reduced. The neutral grounding resistor needs to be short time rated (usually 10 seconds), as the fault will be cleared by the protective relay closest to the fault.
With an HRG system, as shown in Fig. 3, the ground-fault current is in the 10A range. The intention here is to allow the system to operate without tripping, even with a phase-to-ground fault on one phase. When a ground fault does occur, only an alarm is raised. This permits time to locate the fault while power continuity is maintained. This also allows repairs to be done at a scheduled shutdown of the faulty equipment. Maintenance costs should be less than that for a low-resistance grounding system. Damage at the fault location also should be small.
If the fault is in a rotating machine, there usually is no iron damage in the stator. The system, with one phase faulted to earth, operates with the un-faulted phases now raised from earth to the full phase-to-phase voltage for an extended period. As such, the insulation system needs to be rated for phase-to-phase voltage. For phase-to-ground voltages to remain at the phase-to-phase level — and not increased beyond that level — the net capacitive charging current at the fault must be less than that from the controlled resistive ground-fault current fed from the neutral grounding. The grounding resistor also needs to be continuously rated because it will carry the let-through current in the event of a fault for an extended period of time.
With the insulated neutral (ungrounded) system, as shown in Fig. 4 on page 36, there is no intentional connection of the system to ground. In effect, the three phases of the system float. When a ground fault occurs, the fault current is contributed by the system capacitance to earth on the un-faulted phases. This is usually small, and the system can be operated without tripping. Because the system is floating, if the ground fault is of the arcing or intermittent type, then there is the possibility of substantial transient overvoltages, which can be six to eight times the phase voltage. These transients often cause a subsequent failure elsewhere, thus raising the possibility of a phase-to-earth-to-phase fault, as shown in Fig. 5 on page 39, and leading to high fault current and extensive damage. Coordinated tripping is often difficult, and extensive damage is seen at the two faulted locations. Maintenance costs are typically the highest among the four types of grounding systems — now at least two pieces of equipment need repair.
Fig. 5. Ungrounded system showing phase-to-ground-to-phase fault and ground-fault current path. If the ground fault is an arcing or intermittent type, there is possibility of substantial transient overvoltages, which can be six to eight times the phase voltage. These transients often cause a subsequent failure elsewhere, thus raising the possibility of a phase-to-earth-to-phase fault.
The standards and best practices in various countries generally follow ANSI or IEC standards. The technical literature supports these practices. In power plant applications, MV systems occur in two places: generation and station service. In practice, both station service and generators are low- or high-resistance grounded.
For station service at distribution voltages of less than 15kV, power continuity is very important. Here, you would size the neutral grounding resistor so that the let-through ground-fault current is higher than the net current from the distributed capacitance. If the let-through ground-fault current is less than 10A, then this would be high-resistance grounding. If this current were more than 10A, then it would be low-resistance grounding.
Although it's rare to have station service voltage that is higher than 15kV, if the voltage is higher than this, then the same rule as noted above would apply, except the fault should be detected and isolated by tripping the faulted feeder at the closest protective device.
For generators, the ground-fault current is almost always controlled, and you can employ resistance grounding. The resistor let-through current will be dependent on the size of the generator and generation voltage. Typically, 5A to 400A let-through current grounding is used.
More recently, hybrid grounding has been proposed. Here, two resistors in parallel are used: one of low resistance; the other of high resistance (5A). In the event of an internal earth fault in the stator winding of the generator, a fast-acting generator ground differential relay opens the low-resistance grounding path, thus allowing the high resistance (5A let-through resistor) to control and lower the fault current and reducing the stator damage caused by the internal ground fault after the generator has been isolated (while it is slowing down). Without this reduction of current, the generator would continue to feed energy into the fault while it is coming to a stop. The result would be extensive stator iron damage at the ground-fault location.
Question No. 2: What are the advantages and disadvantages of LRG systems compared to HRG systems? Also, what ratings of resistance (in ohms) are considered low and high resistance?
Answer: Here's a brief summary of the differences between HRG and LRG. The first parameter is voltage. By choosing HRG on systems 600V or below, you reduce ground-fault currents to less than 25A. For systems between 600V and 5kV, you can use either HRG or LRG arrangements. In either case, the resistance chosen must be such that the desired let-through ground-fault current is above the system capacitive charging current. On systems above 5kV, LRG is the typical choice in that it reduces ground-fault currents to between 25A and 1,200A. (Most people use a value of 200A or less.)
The next parameter is system capacitive charging current. Every system has a capacitance value, mostly due to the system's cables and surge arresters/capacitors. The general rule of thumb for estimating this charging current is to use 1A per 1,000kVA. If necessary, you can perform a more detailed calculation based on system components. During commissioning, you should determine this value by taking an actual measurement. Once you know the system capacitive charging current, you can select the resistor let-through current or desired current. As previously stated, the desired resistor current (IR) must be greater than the system capacitive charging current to avoid transient overvoltages. You determine the neutral grounding resistance by using the following equation:
R = VL-N ÷ IR
Where R is neutral grounding resistance, VL-N is line-to-neutral voltage, and IR is ground-fault current
For systems less than 600V, the system capacitive charging current is typically 1A to 3A. Therefore, most people use 5A neutral current as a standard. Because this is less than 10A, all 600V systems are high-resistance grounded.
For systems between 600V and 5kV, you can use either HRG or LRG. The decision is typically based on system capacitive charging current, which can vary from 1A to 10A. This value then determines whether to use HRG or LRG. For systems above 5kV, the system capacitive charging current may be greater than 25A, so LRG is almost always used.
Another factor is system continuity. An HRG system allows a distribution system to continue to operate with one fault, without the faulted feeder being tripped. When the possible total earth fault current is such that an earth fault can be sustained continuously without risk of further fault damage, then the system is considered HRG. Electrical installation codes in various jurisdictions have rules governing this.
On systems with grounding resistor let-through currents higher than 10A — or for systems of voltage ratings greater than 5kV — the faulted feeder should be tripped and the fault isolated. These are generally called LRG systems.
Question No. 3: Why do we ground the wye windings of transformers and generators?
Answer: In section 1.4.2, IEEE Std 142-1991 (Green Book) states: “Numerous advantages are attributed to grounded systems, including greater safety, freedom from excessive system overvoltages that can occur on ungrounded systems during arcing, resonant or near-resonant ground faults, and easier detection and location of ground faults when they do occur.”
Now that we've established why you need to ground the neutral, let's discuss how to ground it. If you effectively ground the neutral, you have just replaced the hazards associated with ungrounded systems with new hazards in the form of arc flash / blast hazards, which are associated with solidly grounded systems.
In section 7.2.4, IEEE Std 141-1993 (Red Book) states: “A safety hazard exists for solidly grounded systems from the severe flash, arc burning, and blast hazard from any phase-to-ground fault.” For this reason, IEEE Std 142-1991 (Green Book), in section 1.4.3, states the benefits of resistance grounding: “The reasons for limiting the current by resistance grounding may be one or more of the following:
“1) To reduce burning and melting effects in faulted electric equipment, such as switchgear, transformers, cables, and rotating machines.
“2) To reduce mechanical stresses in circuits and apparatus carrying fault currents.
“3) To reduce electric-shock hazards to personnel caused by stray ground-fault currents in the ground return path.
“4) To reduce the arc blast or flash hazard to personnel who may have accidentally caused or who happen to be in close proximity to the ground fault.
“5) To reduce the momentary line-voltage dip occasioned by the clearing of a ground fault.
“6) To secure control of transient overvoltages while at the same time avoiding the shutdown of a faulty circuit on the occurrence of the first ground fault (high-resistance grounding).”
As you can see, it's best to not only ground the neutral, but also to ground through high-resistance (typically 5A) for all systems less than 600V and most systems from 600V to 5kV. For systems more than 5kV, low-resistance grounding (typically 200A or 400A) is used.
Question No. 4: What are the results of not grounding a 480/277V, 3-phase, 4-wire diesel engine generator set? Take into consideration the two options of switching or not switching the neutral. Also, what's the effect of bonding the neutral to the chassis of the engine generator set?
Answer: By “not grounding,” let's assume you're referring to an ungrounded system. Ungrounded systems can be extremely unsafe, as per our previously cited excerpts from the IEEE Buff and Red Books. In bonding your neutral to the grounded chassis, you are effectively solidly grounding your generator. The consequence of doing this is that you have now replaced hazards. By limiting the ground fault to 5A, you have avoided the hazards with solidly grounded systems.
In addition, several generator set manufacturers require resistance grounding, since the generators are not rated for ground faults. In fact, these faults are often significantly higher than 3-phase faults. In section 1.8.1, the IEEE Green Book states: “Unlike a transformer … a generator will usually have higher initial ground-fault current than 3-phase fault current if the generator has a solidly grounded neutral. According to NEMA, the generator is required to withstand only the 3-phase current level unless it is otherwise specified …”
This is due to very low zero-sequence impedance within the generator, which causes very high earth fault currents. For generators 600V or below, this may not be an issue. However, it is almost always an issue as the voltage class increases.
The resistor also significantly reduces any circulating currents, which are typically triplen harmonics, leading to reduced overheating in the generator windings. Circulating currents are caused by different pitch windings in generators.
For solidly grounded systems, neutral switching is a viable option. According to the NEC, when the service falls under the requirements of 230.95, you should ground the neutral at each source and switch it where the Code requires ground-fault detection coordination. When the service rating equals or exceeds 1,000A (833kVA), 230.95 requires ground-fault protection on the service disconnect. Along with this, if you have an alternate power supply, you must switch the neutral. If you have a service larger than 1,000A, the NEC requires ground-fault protection at the main service disconnect. If the generator neutral grounding runs via a solid connection to the main service neutral and the generator experiences a ground fault while feeding the load, the main service disconnect will open. This will not disconnect the arc fault from the generator, and coordination will be lost. Furthermore, if the neutrals of the two sources are separately grounded, you must switch the load neutral conductor to the source feeding the load, as per 230.95(C) FPN No. 3. Ground-fault current will return only to the source from which it originates, providing for coordination of the ground-fault protection scheme. It's not always necessary to separately ground the generator neutral conductor. However, if you do, you may need to switch a load neutral along with its phase conductors when transferring loads between power sources, particularly when you use ground-fault protection. The NEC requires ground-fault protection for 480/277V, 3-phase, 4-wire, wye-connected services rated 1,000A or more, but it's optional in other configurations that don't include ground-fault protection. However, where a branch circuit neutral conductor transfers between sources, the switching means should assure the neutral conductor switching contact does not interrupt current.
For high-resistance grounding, there are two options: 1) Install resistance grounding on each source; or 2) Derive a neutral on the paralleling bus via a zig-zag transformer, and then add resistance grounding on the derived neutral (Fig. 6). Be careful here in that you cannot use this neutral for any loads or connect it to anything except the resistor. There are advantages/disadvantages for each option. By having a resistor on each source, the total ground-fault current is dependent upon the total number of sources in operation. However, the system is always grounded. If the resistor is on the paralleling bus, the ground-fault current is always the same value however, the system is only grounded if the paralleling bus in operation. Most people choose option 2 and have a resistor on the paralleling bus.
Question No. 5: Why is a neutral-ground (N-G) bond required at the secondary of a delta-wye transformer? Can't I just use lamps connected from each phase to ground to indicate a ground fault?
Answer: The technique referred to here is used on an ungrounded system (Fig. 4). A lamp dims to show the existence of a ground on the system and identifies the faulted phase. However, the dimmed lamp does not pinpoint the exact location of the ground fault, which could be anywhere on the system. Remember, an ungrounded system does not have a direct neutral-to-ground connection, so there is no direct return path for a ground fault. So, the ground fault consists of only the system capacitive charging current, which is typically 1A to 3A for a 600V or less system. Due to this low ground-fault current, no overcurrent protection device will operate, and the ground fault will remain on the system. The resulting continuous operation is why an ungrounded system is used in many continuous manufacturing processes. However, there are serious safety hazards with ungrounded systems. An intermittent (or arcing) ground fault can cause transient overvoltages due to the charging of the system capacitance.
In section 8.2.5, the IEEE Buff Book states: “If this ground fault is intermittent or allowed to continue, the system could be subjected to possible severe overvoltages to ground, which can be as high as six to eight times phase voltage. Such overvoltages can puncture insulation and result in additional ground faults. These overvoltages are caused by repetitive charging of the system capacitance or by resonance between the system capacitance and the inductance of equipment in the system.”
In section 7.2.1, the IEEE Red Book states: “Accumulated operating experience indicates that, in general-purpose industrial power distribution systems, the overvoltage incidents associated with ungrounded operation reduce the useful life of insulation so that electric current and machine failures occur more frequently than they do on grounded power systems.”
Locker is director of business development for I-Gard, Mississauga, Ontario, Canada.
Sidebar: Grounding Topology Comparison
Let's briefly talk about resonance grounding, which is not preferred in the United States mostly due to economics and complexity. Resistance grounding is a passive method that performs independent of system topology and frequency, whereas resonance grounding must adapt to system capacitance. Resonance grounding uses an inductor to create an impedance to match the system capacitance impedance. In doing so, both components cancel each other out, and the result is a small resistive ground-fault current.
There are disadvantages to using resonance grounding:
- Typically, the inductance is slightly larger to avoid a true resonance condition. If not, an overvoltage condition will occur.
- System capacitance continually changes as feeders are brought on- and off-line. So, you must install a monitoring system, and the inductor must be variable.
- Costs for monitoring and inductor variability are high.
- Physical size of inductor is significantly larger than resistor.
Resistance grounding, on the other hand, offers a fixed ground-fault current independent of system topology. However, the fixed current must be larger than the system capacitive charging current. So, a value of 200A to 400A is usually selected.
