Over-Voltage: Sources and Protection | Electrical Engineering

In this article we will discuss about the sources of over-voltage and its protection.

Sources of Over-Voltage:

Transients are disturbances that occur for a very short duration (less than a cycle) and the electrical circuit is quickly restored to original operation provided no damage has occurred due to the transient. An electrical transient is a cause-and-effect phenomenon. For transients to occur, there must be a cause.  

1. Atmospheric Phenomena:

Over potential surge due to lightning discharge is the most common natural cause of electrical equipment failure. A negative charge builds up on a cloud. A corresponding positive charge can build up on the surface of the earth. A voltage difference of hundreds of millions of volts can exist between the cloud and the earth due to the opposing charges. When the voltage exceeds the breakdown potential of air (about 3 x 10⁶ V/m or 75 kV/inch), a lightning flash occurs. It is sufficient to know that a lightning strike can typically produce a voltage rise in about 1 or 2 µsec that can decline to a value of 50% of the peak voltage in approximately 50 to 100 psec.

A common misconception is that a direct lightning strike is needed to produce destructive over voltages. In fact, it is rare that a failure in an electrical system is due to a direct lightning strike. More often, the electrical and magnetic fields caused by indirect lightning discharge induce voltages in the power lines that result in device failures. Also, lightning discharge current flowing through the earth creates a potential difference between the power lines and ground and in extreme cases causes equipment failure.

Isolation transformers provide limited protection from lightning strikes. Because lightning is a short-duration, high-frequency phenomenon, a portion of the lightning energy will couple directly from the primary winding to the secondary winding of the transformer through the interwinding capacitance. This is why equipment supplied from the low voltage winding of a transformer that is exposed to lightning energy is also at risk.

The amount of voltage that will be coupled through the transformer will depend on the transformer interwinding capacitance itself. The higher the capacitance, the higher the transient energy coupled to the secondary. Transformers provided with a grounded shield between the primary and the secondary windings provide better protection against lightning energy present at the transformer primary winding.

Lightning arresters, when properly applied, can provide protection against lightning-induced low voltages. Arresters have a well-defined conduction voltage below which they are ineffective. This voltage depends on the rating of the arrester itself. For optimum protection, the arrester voltage should be matched to the lightning impulse withstand of the equipment being protected.

Lightning:

Lightning is a potent source of impulsive transients. We will concentrate on how lightning causes transient over-voltages to appear on power systems. Figure 3.3 illustrates some of the places where lightning can strike that results in lightning currents being conducted from the power system into loads. The most obvious conduction path occurs during a direct strike to a phase wire, either on the primary or the secondary side of the transformer. This can generate very high over-voltages, but some analysts question whether this is the most common way that lightning surges enter load facilities and cause damage.

Very similar transient over-voltages can be generated by lightning currents flowing along ground conductor paths. Note that there can be numerous paths for lightning currents to enter the grounding system. Common ones, indicated by the dotted lines in Fig. 3.3, include the primary ground, the secondary ground, and the structure of the load facilities. Note also that strikes to the primary phase are conducted to the ground circuits through the arresters on the service transformer.

Thus, many more lightning impulses may be observed at loads than one might think. Keep in mind that grounds are never perfect conductors, especially for impulses. While most of the surge current may eventually be dissipated into the ground connection closest to the strike, there will be substantial surge currents flowing in other connected ground conductors in the first few microseconds of the strike.

A direct strike to a phase conductor generally causes line flash over near the strike point. Not only does this generate an impulsive transient, but it causes a fault with the accompanying voltage sags and interruptions. The lightning surge can be conducted a considerable distance along utility lines and cause multiple flashovers at pole and tower structures as it passes.

The interception of the impulse from the phase wire is fairly straightforward if properly installed surge arresters are used. If the line flashes over at the location of the Strike, the tail of the impulse is generally truncated. Depending on the effectiveness of the grounds along the surge current path, some of the current may find its way into load apparatus.

Arresters near the strike may not survive because of the severe duty (most lightning strokes are actually many strokes in rapid-fire sequence). Lightning does not have to actually strike a conductor to inject impulses into the power system. Lightning may simply strike near the line and induce an impulse by the collapse of the electric field. Lightning may also simply strike the ground near a facility causing the local ground reference to rise considerably. This may force currents along grounded conductors into a remote ground, possibly passing near sensitive load apparatus. Many investigators in this field postulate that lightning surges enter loads from the utility system through the interwinding capacitance of the service transformer.

The concept is that the lightning impulse is so fast that the inductance of the transformer windings blocks the first part of the wave from passing through by the turns ratio. However, the interwinding capacitance may offer a ready path for the high- frequency surge. This can permit the existence of a voltage on the secondary terminals that is much higher than what the turns ratio of the windings would suggest.

The degree to which capacitive coupling occurs is greatly dependent on the design of the transformer. Not all transformers have a straightforward high-to-low capacitance because of the way the windings are constructed. The winding-to-ground capacitance may be greater than the winding-to-winding capacitance, and more of the impulse may actually be coupled to ground than to the secondary winding. In any case, the resulting transient is a very short single impulse, or train of impulses, because the interwinding capacitance charges quickly.

Arresters on the secondary winding should have no difficulty dissipating the energy in such a surge, but the rates of rise can be high. Thus, lead length becomes very important to the success of an arrester in keeping this impulse out of load equipment. Many times, a longer impulse, which is sometimes oscillatory, is observed on the secondary when there is a strike to a utility’s primary distribution system. This is likely not due to capacitive coupling through the service transformer but due to conduction around the transformer through the grounding systems.

This is a particular problem if the load system offers a better ground and much of the surge current flows through conductors in the load facility on its way to ground.

The chief power quality problems with lightning stroke currents entering the ground system are:

1. They raise the potential of the local ground above other grounds in the vicinity by several kilovolts. Sensitive electronic equipment that is connected between two ground references, such as a computer connected to the telephone system through a modem, can fail when subjected to the lightning surge voltages.

2. They induce high voltages in phase conductors as they pass through cables on the way to a better ground.

Ideas about lightning are changing with recent research. Lightning causes more flash overs of utility lines than previously thought. Evidence is also mounting that lightning stroke current wave-fronts are faster than previously thought and that multiple strikes appear to be the norm rather than the exception. Durations of some strokes may also be longer than reported by earlier researchers. These findings may help explain failures of lightning arresters that were thought to have adequate capacity to handle large lightning strokes.

2. Switching Loads ON or OFF:

Switching normal loads in a facility can produce transients. The majority of plant loads draw large amounts of current when initially turned on. Transformers draw inrush currents that range between 10 and 15 times their normal full-load current. This current lasts between 5 and 10 cycles. Alternating current motors draw starting currents that vary between 500 and 600% of the normal full-load running current. Fluorescent lights draw inrush currents when first turned on. Large current drawn through the impedance of the power system sets up transient voltages that affect electrical components sensitive to sags, sub-cycle oscillations, or voltage notch.

There are instances when conditions are such that harmonic frequency currents in the inrush current interact with the power system inductance and capacitance and cause resonance conditions to develop. During resonance, substantial over-voltages and over-currents might develop. In the strict sense, these are not sub-cycle events and therefore may not be classified as transients, but their effects are nonetheless very detrimental. Large inrush currents drawn by certain loads produce other negative effects.

Consider a conductor carrying a large current. The magnetic field due to the surge current could induce large potentials in adjacent signal or data cables by inductive coupling. This is why it is preferable to keep signal or data cables physically distant from power cables. Data and signal wires that run near power cables should be contained in metal conduits made of steel. Steel, due to its magnetic properties, is a better shield at low frequencies than nonferrous metals such as aluminium or copper. Non-ferrous metals make better shields at high frequencies.

When discussing inductive coupling due to transient current, the loop area of the susceptible circuit should not be overlooked, as the larger the area of the loop, the higher the noise voltage induced in the susceptible circuit.

In Figure 3.6, the voltage induced in circuit 2 depends on the magnetic field linking the circuit; the larger the loop area, the larger the flux linkage and, therefore, the higher the noise voltage induced in circuit 2. Twisted pairs of wires minimize the loop area and reduce noise voltage pickup, thus signal and data circuits for sensitive, low-level signal applications installed in close proximity to power wires should use twisted sets of wires to reduce noise coupling.

So far we have examined the effects of switching power to loads during normal operation. Switching power off also generates transients due to the fact that all devices carrying electrical current have inductances (L) associated with them. The current flowing in an inductive device cannot abruptly change when the circuit is opened.

The voltage produced in an electrical device due to a sudden change of current is given by-

E = Ldi/dt

Where, di/dt is the rate change of current and L is the inductance associated with the device.

Total flux linking circuit 2 and noise voltage induced in circuit 2 are proportional to loop area 2.

For example, if 60 A of current flowing through a coil of inductance L = 20 mH drops to zero in 2 msec, then the voltage generated across the coil is given by-

E = (20*10^-3*60)/(2 *10^-3) = 600 V

It is easy to see that substantial voltages can be developed while switching inductive devices off. The transient voltage produced can easily couple to other circuits via stray capacitance between the inductive device and the susceptible circuit. This voltage can appear as noise for the second circuit. The closer the two circuits are spaced, the higher the amount of noise that is coupled. Voltages as high as 2000 to 5000 V are known to be generated when large inductive currents are interrupted. In low voltage power and signal circuits, this can easily cause damage.

3. Interruption of Fault Current:

During fault conditions, large currents are generated in an electrical system. The fault currents are interrupted by overcurrent devices such as circuit breakers or fuses.

Figure 3.7 shows a simplified electrical circuit where an electrical fault is cleared by a circuit breaker. C represents the capacitance of the electrical system up to the point where the over current device is present. Interruption of the fault current generates over-voltage impulse in the electrical system, and the magnitude of the voltage depends on the amount of fault current and the speed with which the fault is interrupted.

Older air circuit breakers with slower speed of interruption produce lower impulse voltages than vacuum or SF6 breakers, which operate at much faster speeds during a fault. While using vacuum or SF6 technology to clear a fault quickly and thereby limit damage to equipment is an important advantage, a price is paid by the generation of higher level voltage transients.

Once the fault is interrupted, the generated voltage impulse can interact with the inductance and capacitance of the electrical system and produce oscillation at a frequency much higher than the fundamental frequency. The oscillations are slowly damped out by the resistance associated with the system. 

The voltage can build up to levels equal to twice the peak value of the voltage waveform. The over-voltage and associated oscillations are harmful to electrical devices.

A very serious case of over-voltages and oscillations occurs when the over current device is supplied from overhead power lines that connect to long lengths of underground cables. Underground cables have substantial capacitance to ground. The combination of the inductance due to overhead lines and capacitance due to underground cables could generate high levels of over-voltage and prolonged oscillations at low frequencies. Such transients are very damaging to transformers, cables, and motors supplied from the lines. In extreme cases, voltages as high as three to four times the AC peak voltage may be generated.

4. Switching of Capacitor Banks:

One of the more common causes of electrical transients is switching of capacitor banks in power systems. Electrical utilities switch capacitor banks during peak load hours to offset the lagging kVAR demand of the load. The leading kVARs drawn by the capacitor banks offset the lagging kVAR demand of the load, reducing the net kVA load on the circuit. Switching of capacitor banks is accompanied by a surge of current which is initially limited by the characteristic impedance of the power system and resistance of the line.

A sharp reduction in the voltage is followed by a voltage rise, which decays by oscillation at a frequency determined by the inductance and capacitance of the circuit. Several cases of power system component failures and malfunctions due to capacitor bank switching operations. Typically, the voltage rise due to capacitor switching operation can attain values 1.5 to 2 times the nominal voltage.

Power equipment can withstand only a limited number of exposures to such rises in voltage magnitude. With time, the insulation systems of such devices weaken, and a point is reached when the devices can fail. In one particular instance, two power distribution transformers failed at the same time; the cause was traced to large capacitor bank switching operations by the utility at a substation located adjacent to the affected facility.

Adjustable speed drives (ASDs) and solid-state motor controllers are quite sensitive to voltage rises resulting from capacitor bank switching operations. The ASD might shut down the motor due to voltage on the system rising beyond the maximum tolerance. In some cases, capacitor switching causes the voltage waveform to undergo oscillations and produce stray crossings of the time axis. This is unacceptable for devices that require the precise number of zero time crossings for proper performance.

Example:

A 2000-kVAR, 13.8-kV, Y-connected capacitor bank is connected at the end of a 25-mile transmission line with an inductive reactance of 0.5 £2 per mile. Find the natural frequency of the current that would be drawn during turn on-

Total inductive reactance -XL- 25*0.5 = 12.5 Ω

Inductance (L) = 12.5/120n = 0.033 H

Current through the capacitor bank (I_c) = 83.7 A

Capacitive reactance (XC) = 7967/83.7 = 95.18 Ω

Capacitance (C) = 27.9 µF

The current drawn from the source will have a frequency of 166 Hz which will decay as determined by the power system resistance. Due to impedance drops associated with the currents, the voltage waveform would experience similar oscillations prior to settling down to nominal levels. The series resonance circuit formed by the system inductance and the capacitance could produce a voltage rise in the electrical system.

Depending on the severity of the voltage rise and the time to decay, equipment damage can result, especially if the switching operations are frequent. The condition is made worse if a second capacitor bank is brought online. The natural frequencies associated with this action are higher and so is the time to decay.

Considerable energy is exchanged between the two capacitors before steady- state operation is attained.

Principles of Over-Voltage Protection:

The fundamental principles of over-voltage protection of load equipment are:

1. Limit the voltage across sensitive insulation.

2. Divert the surge current away from the load.

3. Block the surge current from entering the load.

4. Bond grounds together at the equipment.

5. Reduce, or prevent, surge current from flowing between grounds.

6. Create a low-pass filter using limiting and blocking principles.

Figure 3.16 illustrates these principles, which are applied to protect from a lightning strike.

The main function of surge arresters and transient voltage surge suppressors (TVSSs) is to limit the voltage that can appear between two points in the circuit. This is an important concept to understand. One of the common misconceptions about varistors and similar devices, is that they somehow are able to absorb the surge or divert it to ground independently of the rest of the system.

That may be a beneficial side effect of the arrester application if there is a suitable path for the surge current to flow into, but the foremost concern in arrester application is to place the arresters directly across the sensitive insulation that is to be protected so that the voltage seen by the insulation is limited to a safe value.

Surge currents, just like power currents, must obey Kirchhoff’s laws. They must flow in a complete circuit, and they cause a voltage drop in every conductor through which they flow. One of the points to which arresters, or surge suppressors, are connected is frequently the local ground, but this need not be the case. Keep in mind that the local ground may not remain at zero potential during transient impulse events.

Surge suppression devices should be located as closely as possible to the critical insulation with a minimum of lead length on all terminals. While it is common to find arresters located at the main panels and subpanels, arresters applied at the point where the power line enters the load equipment are generally the most effective in protecting that particular load. In some cases, the best location is actually inside the load device.

For example, many electronic controls made for service in the power system environment have protectors [metal-oxide varistor (MOV) arresters, gaps, zener diodes, or surge capacitors] on every line that leaves the cabinet.

In Fig. 3.16 the first arrester is connected from the line to the neutral-ground bond at the service entrance. It limits the line voltage V, from rising too high relative to the neutral and ground voltage at the panel. When it performs its voltage limiting action, it provides a low impedance path for the surge current to travel onto the ground lead. Note that the ground lead and the ground connection itself have significant impedance. Therefore, the potential of the whole power system is raised with respect to that of the remote ground by the voltage drop across the ground impedance.

For common values of surge currents and ground impedances, this can be several kilovolts. One hopes, in this situation, that most of the surge energy will be discharged through the first arrester directly into ground. In that sense, the arrester becomes a surge “diverter.” This is another important function related to surge arrester application.

In fact, some prefer to call a surge arrester a surge diverter because its voltage-limiting action offers a low-impedance path around the load being protected. However, it can only be a diverter if there is a suitable path into which the current can be diverted. That is not always easy to achieve, and the surge current is sometimes diverted toward another critical load where it is not wanted.

In this figure, there is another possible path for the surge current—the signal cable indicated by the dotted line and bonded to the safety ground. If this is connected to another device that is referenced to ground elsewhere, there will be some amount of surge current flowing down the safety ground conductor. Damaging voltages can be impressed across the load as a result.

The first arrester at the service entrance is electrically too remote to provide adequate load protection. Therefore, a second arrester is applied at the load—again, directly across the insulation to be protected. It is connected “line to neutral” so that it only protects against normal mode transients. This illustrates the principles without complicating the diagram but should be considered as the minimum protection one would apply to protect the load.

Frequently, surge suppressors will have suppression on all lines to ground, all lines to neutral, and neutral to ground. While lightning surge currents are seeking a remote ground reference, many transient over-voltages generated by switching will be those of a normal mode and will not seek ground.

In cases where surge currents are diverted into other load circuits, arresters must be applied at each load along the path to ensure protection. Note that the signal cable is bonded to the local ground reference at the load just before the cable enters the cabinet. It might seem that this creates an unwanted ground loop. However, it is essential to achieving protection of the load and the low voltage signal circuits. Otherwise, the power components can rise in potential with respect to the signal circuit reference by several kilovolts.

Many loads have multiple power and signal cables connected to them. Also, a load may be in an environment where it is close to another load and operators or sensitive equipment are routinely in contact with both loads. This raises the possibility that a lightning strike may raise the potential of one ground much higher than the others. This can cause a flash over across the insulation that is between the two ground references or cause physical harm to operators.

Thus, all ground reference conductors (safety grounds, cable shields, cabinets, etc.) should be bonded together at the load equipment. The principle is not to prevent the local ground reference from rising in potential with the surge; with lightning, that is impossible.

Rather, the principle is to tie the references together so that all power and signal cable references in the vicinity rise together. This phenomenon is a common reason for failure of electronic devices. The situation occurs in TV receivers connected to cables, computers connected to modems, computers with widespread peripherals powered from various sources, and in manufacturing facilities with networked machines.

Since a few feet of conductor make a significant difference at lightning surge frequencies, it is sometimes necessary to create a special low inductance, ground reference plane for sensitive electronic equipment such as mainframe computers that occupy large spaces.

Efforts to block the surge current are most effective for high frequency surge currents such as those originating with lightning strokes and capacitor switching events. Since power frequency currents must pass through the surge suppressor with minimal additional impedance, it is difficult and expensive to build filters that are capable of discriminating between low frequency surges and power frequency currents.

Blocking can be done relatively easily for high frequency transients by placing an inductor, or choke, in series with the load. The high surge voltage will drop across the inductor. One must carefully consider that high voltage could damage the insulation of both the inductor and the loads. However, a line choke alone is frequently an effective means to block such high frequency transients as line-notching transients from adjustable speed drives.

The blocking function is frequently combined with the voltage limiting function to form a low-pass filter in which there is a shunt-connected voltage limiting device on either side of the series choke.

Figure 3.16 illustrates how such a circuit naturally occurs when there are arresters on both ends of the line feeding the load. The line provides the blocking function in proportion to its length. Such a circuit has very beneficial over voltage protection characteristics. The inductance forces the bulk of fast-rising surges into the first arrester. The second arrester then simply has to accommodate what little surge energy gets through.

Such circuits are commonly built into outlet strips for computer protection. Many surge protection problems occur because the surge current travels between two, or more, separate connections to ground. This is a particular problem with lightning protection because lightning currents are seeking ground and basically divide according to the ratios of the impedances of the ground paths.

The surge current does not even have to enter the power, or phase, conductors to cause problems. There will be a significant voltage drop along the ground conductors that will frequently appear across critical insulation. The grounds involved may be entirely within the load facility, or some of the grounds may be on the utility system.

Ideally, there would be only one ground path for lightning within a facility, but many facilities have multiple paths. For example, there may be a driven ground at the service entrance or substation transformer and a second ground at a water well that actually creates a better ground.

Thus, when lightning strikes, the bulk of the surge current will tend to flow toward the well. This can impress an excessively high voltage across the pump insulation, even if the electrical system is not intentionally bonded to a second ground.

When lightning strikes, the potentials can become so great that the power system insulation will flash over somewhere. The amount of current flowing between the grounds may be reduced by improving all the intentional grounds at the service entrance and nearby on the utility system. This will normally reduce, but not eliminate entirely, the incidence of equipment failure within the facility due to lightning.

However, some structures also have significant lightning exposure, and the damaging surge currents can flow back into the utility grounds. It doesn’t matter which direction the currents flow; they cause the same problems. Again, the same principle applies, which is to improve the grounds for the structure to minimize the amount of current that might seek another path to ground.

When it is impractical to keep the currents from flowing between two grounds, both ends of any power or signal cables running between the two grounds must be protected with voltage limiting devices to ensure adequate protection. This is common practice for both utility and end user systems where a control cabinet is located quite some distance from the switch, or other device, being controlled.