Corona Discharge

What Is Corona

During the past several years, corona and radio influence (RI) problems associated with extra-high- voltage transmission have been studied extensively in the Ohio Brass Voltage Laboratory. The word “corona”, as well as the definitive expressions “plume discharge”, “brush discharge” and “glow discharge” have been used extensively in connection with this work, under the tacit assumption that everyone is familiar with their meanings. It is the purpose of this article to explain the meanings of these expressions as used in such

studies and to describe some of the physical and electrical properties associated with them.

Definition of Corona

Perhaps the most general definition of corona possible is as follows: “Corona is a discharge caused by electrical overstress.” While this definition is very general, and applies to practically all kinds of corona, it is still unsatisfactory since it introduces two new expressions, “electrical overstress” and “discharge,” which, although they describe the “cause” and “effect”, must in turn be defined for various cases.

Actually, corona can appear in solid, liquid or gaseous insulating materials, and its occurrence therein is usually associated with the initial phases of electrical failure of the insulation. In solids, the occurrence of corona generally results in deterioration of the material, while in liquids and gases, removal of the electrical overstress eliminates the discharge, and the material generally recovers its original insulating properties.

For transmission line studies, the insulating material in which the discharge occurs is the air adjacent to conductor or insulator surfaces, when the electrical stress at these surfaces exceeds the critical value.

Limiting corona to transmission line conductors leads to the American Standards Association definition of corona, which reads as follows: “Corona is a luminous discharge due to ionization of the air surrounding a conductor around which exists a voltage gradient exceeding a certain critical value.”

This definition obviously contains the same basic idea given in the above general definition, but has a rather limited scope. For the purpose of this article the scope of this definition will be enlarged to include corona on line hardware and insulators at high voltages.

Nature of Corona

The corona discharges observed at the surface of a conductor are due to the formation of electron avalanches which occur when the intensity of the electric field at the conductor surface exceeds a certain critical value.

There are always a few free electrons in the air as a result of traces of radioactive materials in the earth’s crust and cosmic ray bombardment of the earth from outer space. As the conductor becomes energized on each half cycle of the AC voltage wave, the electrons in the air near its surface are accelerated by the electrostatic field. These electrons, having an inherent negative charge, are accelerated toward the conductor on its positive half cycle and away from the conductor on its negative half cycle.

The velocity attained by a free electron is dependent upon the intensity of the electric field. If the intensity of the electric field is not too great, the collision between an electron and an air molecule such as oxygen (O2) or nitrogen (N2) is elastic; that is, the electron bounces off the air molecule with no transfer of energy to it. On the other hand, if the intensity of the electric field exceeds a certain critical value, any free electron in this field will acquire a sufficient velocity so that its collision with the air molecule is inelastic; that is, the electron has sufficient energy to knock one of the outer orbit electrons clear out of one of the two atoms of the air molecule. This is the phenomenon known as ionization, and the molecule with the missing electron becomes a positive ion.

The initial electron, which lost most of its velocity in the collision, and the electron knocked out of the air molecule, which also has a low velocity, are both accelerated by the electric field, and at the next collision, each electron is capable of ionizing an air molecule. After the second collision, there are now four electrons to proceed, and so on, the number of electrons doubling after each collision. All this time, the electrons are all advancing to-ward the positive electrode and after many collisions, their number has grown enormously. This is the process by which the so-called electron avalanche is built up, each avalanche being initiated by a single free electron which finds itself in an intense electrostatic field.

The intensity of the electrostatic field around a conductor is non-uniform. It has its maximum strength at the surface of the conductor and its intensity diminishes inversely as the distance from the center of the conductor. Hence, as the voltage level in the conductor is raised, the critical field strength is approached and the initial discharges occur only at or very near to the conductor surface. For the positive half cycle, the electron avalanches move toward the conductor and continue to grow until they hit the surface. For the negative half cycle, the electron avalanches move away from the conductor surfaces toward a weaker field and cease to advance when the field becomes too weak to accelerate the electrons to ionizing velocity.

The positive ions left in the wake of the electron avalanche move toward the negative electrode. However, they move very slowly because of their mass, which for air molecules is in the order of 50,000 times the mass of the electron. Having a positive charge, these ions attract wandering electrons, and whenever one succeeds in capturing a free electron, it becomes a neutral air molecule again. The energy level of a neutral molecule is less than that of the corresponding positive ion, and hence when a free electron is captured, a quanta or “chunk” of energy is emitted from the molecule. This quanta of energy is exactly equal in magnitude to the energy which initially was required to knock the electron out of the molecule in the first place. It is radiated as an electro-magnetic wave, and for air molecules such as oxygen or nitrogen, this radiation is in the visible light range. Hence, an observer can see this radiation as a soft violet- colored light, which comes principally from the recombination of nitrogen ions with free electrons.

The Manifestations of Corona

The discharges which are produced by electron avalanches may be observed in the laboratory in three different ways. Perhaps the best known manifestation is “visual corona” which appears as a violet colored light coming from the regions of electrical overstress when the test specimen is viewed in the dark. As described above, this light is produced by the re-combination of positive nitrogen ions with free electrons.

The second manifestation of this discharge is “audible corona”, which appears as a hissing or frying sound whenever the specimen is energized above the corona threshold voltage. The sound waves are produced by the disturbances set up in air in the vicinity of the discharge, possibly by the movement of the positive ions as they are suddenly created in an intense electric field.

The third, and perhaps most serious manifesta- tion of this discharge from the point of view of the power company, is the electrical effect which causes radio influence or RI. These avalanches, being electrons in motion, actually constitute electric currents, and as such, produce both magnetic and electrostatic fields in the vicinity. Being formed very suddenly and being of short duration, these magnetic and electrostatic fields can induce high-frequency voltage pulses in nearby radio antennas, and hence may cause RI. These electrical disturbances are usually measured in the laboratory with a radio noise meter which is closely coupled to the test specimen by means of the standard NEMA circuit. This laboratory set-up measures the generated radio noise, and is usually called the “radio

influence voltage” or RIV of the test specimen.

The Aspects of Corona

The three different types or degrees of corona discharges which are recognized on EHV test specimens in the laboratory are called “plume discharge,” “brush discharge” and “glow discharge.” The plume discharge is the most spectacular of the three, and is so named because of its general resemblance to a plume. When viewed in the dark, it has a concentrated stem which may be anywhere from a fraction of an inch long to several inches in length, depending upon the voltage level of the conductor. At its outer end, the stem branches many times and merges into a violet-colored tree- like halo which may range in length from a few inches at lower voltages to a foot or more at very high voltages. The audible manifestation associated with plume discharges is generally a rather intense snapping and hissing sound, readily recognized by the experienced corona observer. The brush discharge is a streamer projecting radially from the conductor surface. These discharges generally occur all around the periphery of the conductor. The length of these discharges may vary from a small fraction of an inch at low voltages to one or two inches at higher voltages. The name is suggested by the resemblance which the discharges have to the bristles of a round bottle brush. The audible manifestation associated with brush discharges is generally a continuous background type of hissing or frying sound. The glow discharge is a very faint, weak light which appears to hug the conductor surface and does not project there from as does a brush discharge. It also may appear on critical regions of insulator surfaces during high humidity conditions. There is generally no sound associated with glow discharges.

Properties of Corona

On a clean, dry, smooth conductor energized a little above its critical voltage, only brush discharges occur, and these are generally limited to the negative half cycle. Fig. 1 shows the corona which occurs on a smooth conductor when photo- graphed under various conditions. Fig. 1-A shows the conductor as viewed by ordinary means. Fig. 1- B shows it on the negative half cycle only, while Fig. 1-C shows how it looks on the positive half cycle only. These last two photographs were taken with a rotating disk located in front of the camera lens. This disk, which had two 90-degree open segments and two 90-degree closed segments, was rotated at 1800 rpm with a synchronous motor. The disk was adjusted on the motor shaft so that the camera could “see” the conductor only one half the time, corresponding to either pre-selected half cycle of the 60-cycle voltage wave applied to the conductor.

As indicated in Fig. 1-B, the corona discharge appears the same on the negative half cycle as it does in Fig. 1-A for both half cycles. On the positive half cycle, the corona is practically absent, except for one lone incipient plume discharge being present, probably due to a tiny defect on the conductor surface.

This difference in the appearance of corona on the alternate polarities may be readily explained by the above described discharge mechanism. As the voltage rises from zero in the positive direction, electrons in the vicinity of the conductor surface move towards it into a region of higher field intensity, bumping their way through the molecules of air. As long as each electron’s velocity at the instant of collision remains below the ionizing velocity, no positive ions are formed, and hence there is no discharge. This process continues until each electron in the vicinity finally arrives at and enters the conductor surface. If the maximum field intensity at the voltage crest is insufficient to cause ionization, there will be no discharge.

As the voltage rises from zero in the negative direction, electrons in the vicinity of the conductor surface move away from it into a region of weaker field intensity, bumping their way through the molecules of air. As long as the field intensity is too low to accelerate any electron to its ionizing velocity between successive collisions, there is no discharge. However, unlike the positive half cycle, electrons initially present at the start of the negative half cycle are still present as the 60-cycle voltage wave approaches crest. If the maximum field intensity at the voltage crest is insufficient to cause ionization, there will be no discharge. For a sine wave of voltage, the field intensity is the same for either crest. Hence, the appearance of brush discharges on the negative half cycle and no discharges on the positive half cycles for a smooth conductor indicates that a higher field intensity is required for ionization with the conductor surface positive. This difference is explained by the removal of electrons from the neighborhood of the conductor surface on the positive half cycle as the voltage increases at the 60-cycle rate toward its crest value.

Any defect on the conductor which projects however slightly above the nominal conductor surface, increases the field intensity in its immediate vicinity. Hence, the defect on the conductor shown in Fig. 1 projected far enough so that at a test voltage of 200 kV, it increased the field intensity sufficiently to accelerate to ionizing velocities those free electrons still remaining in its vicinity, and thus produced the incipient plume discharge shown in Fig. 1-C.

A second method of studying corona discharges is to observe which polarity produces the higher readings in the radio noise meter. This is accomplished easily by connecting  a cathode ray oscilloscope to the phone jack of the radio noise meter and superimposing its output upon a 60-cycle reference voltage wave which is synchronized with the 60- cycle test voltage applied to the specimen. A series of such oscillograms is shown in Fig. 2. The oscillogram of Fig. 2-A is for the clean dry smooth conductor described in Fig. 1. As shown the disturbance originates on the negative half cycle only, and always near its crest. The RIV produced by the single incipient plume discharge on the positive half cycle, as shown in Fig. 1-C, is negligible compared to that produced by the negative half cycle brush discharges, and hence is lost in the oscillogram. Introducing an intentional surface defect of some- what greater magnitude caused a full-fledged plume discharge to appear on the conductor surface on the positive half cycle as is shown by the series of

photographs of Fig. 3. The oscillogram of Fig. 2-B shows that now the principal RIV occurs on the positive half cycle, with the negative half cycle discharges causing only relatively slight disturbances. The one full-fledged plume on the positive half cycle produced many times as much disturbance as did all of the brush discharges on the negative half cycle. This difference is shown by the radio noise meter readings which gave an RIV level of 3000 microvolts for the brush discharges of Fig. 1 and 25,000-microvolt level for the single plume discharge of Fig. 3.

As shown in Fig. 3-B, the negative polarity brush discharge from the intentional defect is much larger than are those from the smooth portion of the conductor surface. To the right of the intentional defect there is also a pair of large brush discharges which apparently came from some un-intentional defect on the conductor surface. The brushes from the defects appear much larger and brighter than do those coming from the smooth conductor surface, because they begin to develop at a somewhat lower voltage on the 60 cycle wave and the crest field intensity on the conductor surface at the defects is higher than it is on the smooth portion of its surface. Water drops on the conductor surface provide a multiplicity of projections from which corona discharges can originate. The series of photographs of Fig. 4 show that many small plumes occur on the positive half cycle and likewise many brush discharges occur on the negative half cycle. The oscillogram of Fig. 2-C shows that the plumes on the positive half cycle establish the RIV level with the wet conductor.

A weathered ACSR conductor generally has a multiplicity of tiny surface defects which project above the nominal surface of the conductor. The series of photographs of Fig. 5 show that many small plume discharges occur on the positive half cycle and that many small brush discharges occur on the negative half cycle. The oscillograms of Fig. 2-D show that the discharges on the positive half cycle usually establish the RIV level of the conductor.

The plume and brush discharges shown in Fig. 6 were obtained during a series of tests on a 1.602- inch diameter ACSR stranded conductor. The plumes occur only at locations where one of the conductor strands was intentionally displaced so that it protruded about 0.10 of an inch above the nominal conductor surface. The brush discharges occurred only on the negative half cycle, while the plume discharges occurred only on the positive half

cycle. Corona discharges on a 4-conductor bundle tested in the laboratory (see page 4) appear to be principally brushes, there being only a few small plumes apparent. The corona discharges, which appear only on the outer surface of each conductor, project radially from the geometric axis of the bundle. This is the direction of maximum surface gradient.

The Formation of Brush and Plume Discharges

The reason why brush discharges occur on the negative half cycle and plume discharges occur on the positive half cycle may be explained in terms of the above described corona mechanism.

a. Brush Discharge

As the negative half cycle voltage increases at the normal 60-cycle rate, the field intensity, which  is always greatest at the conductor surface, in- creases at the same rate. When the field intensity reaches the critical value, somewhat below the voltage crest, free electrons near the conductor surface are accelerated to ionizing velocities; and avalanches start to form, moving in the positive direction of the field, which is away from the conductor surface. The heavy positive ions left behind form a positive space charge between the conductor surface and the tip of the advancing electron avalanche. This tends to weaken the field produced by the energized conductors. On the other hand, the advancing avalanche, being an accumulation of a large number of electrons, produces a

negative space charge just ahead of itself, which tends to reinforce the intensity of the field due to the conductor. This increased field causes free electrons ahead of the advancing tip likewise to initiate new avalanches which advance the negative space charge still further. This process is accumulative, and the brush discharge will continue to grow in length until the net field strength at the outer tip of the advancing discharge is insufficient to cause further ionization. Obviously the length of the streamer will be determined by the conductor voltage and will achieve its maximum length at crest voltage.

The existence of a brush discharge becomes apparent by the visible light which is emitted by each positive ion as it recombines with a free electron to form a neutral molecule. Thus, the location and extent of the brush discharge created by electron avalanches is revealed.

b. Plume Discharge

When the field intensity near the conductor surface reaches the critical value on the positive half cycle, electrons in the vicinity of the surface are accelerated to ionizing velocity and, therefore, initiate electron avalanches, which advance toward the conductor surface and ultimately strike it. The heavy ions left behind form a positive space charge, the tip of which projects outward from the conductor surface. The increased field intensity at the tip promotes the formation of additional electron avalanches which move toward it. As a result, the positive space charge projection continues to grow outward from the conductor.

Near the conductor surface, avalanches are created principally in front of the advancing positive space charge projection, since here the field from the conductor itself is intense and the combined field is more or less radial to the conductor. The further the positive space charge projection advances, the weaker becomes the conductor field and the resultant field at its tip becomes more divergent, approaching a hemispherical radial field.

Hence, electron avalanches now can approach the tip of the advancing positive space charge from different directions, thus initiating positive ion projections which branch out in different directions from the initial stem. Many of these branches continue to grow and again branch out much like the limbs, branches, and twigs of a tree. This growth continues until ultimately the plume reaches such a size that the field intensity at the tips of the numerous branching positive streamers is insufficient to cause further ionization.

The plume shown in Fig. 3-C is an excellent example of a full-fledged plume, while the discharge shown in Fig. 1-C is an example of an incipient plume which had barely reached the branching stage.

Radio Influence Associated with Corona Discharges

Positive polarity plumes are a prolific source of electrical disturbance, producing RIV levels in the order of 10 times those of the negative polarity brush discharges. In general, it has been found in tests on trans- mission line conductors above 0.5-inch diameter that when plumes occur, RIV levels are generously in excess of 10,000 microvolts. The observed range of RIV levels for brush discharges is from about 100 microvolts to 5000

microvolts. The lower figure would apply to just a few audible, as well as visible, brush discharges on the conductor just above the visual corona voltage, while the upper figure applies to higher test volt- ages, or to voltages just below those at the point where they occur. Glow discharges are observed only infrequently on conductor surfaces, and when they do occur, they usually persist at voltages below which audible and visual brush discharges have disappeared. RIV levels associated with glow discharges are usually under 10 microvolts. These discharges appear to be caused by the tiniest of surface defects and can usually be eliminated by cleaning the conductor surface at the point where they occur.