Part 2 · Chapter 7

Corona and Breakdown in Non-Uniform Fields

Every gap so far has been uniform — the same field everywhere, breaking down all at once. Real apparatus is rarely so polite: a thin conductor, a sharp edge, a bolt head concentrates the field at one spot while the rest of the gap stays calm. There, and only there, the gas can break down on its own — a glowing, hissing, self-limiting discharge called corona. This chapter is the physics of that partial breakdown, the law that predicts its onset, and the very real costs it imposes on a transmission line.

High-Voltage Engineering Prof. Mithun Mondal Reading time ≈ 45 min
i What you'll learn
  • Why a non-uniform field breaks down locally first, and how the enhancement factor \(f=E_{\max}/E_{\text{mean}}\) from Chapter 2 governs where.
  • Corona as a self-limiting partial breakdown confined to the over-stressed region while the bulk gap holds.
  • Corona inception and Peek's law: the disruptive and visual critical voltages, the surface-irregularity factor, and the air-density factor \(\delta\).
  • The polarity effect — why a positive point sparks over at a lower voltage than a negative one, despite igniting corona later.
  • The forms corona takes: Trichel pulses, streamers, and the steady glow.
  • What corona costs — power loss, radio interference, audible noise, ozone — and Peek's loss formula for line design.
Section 7-1

The Non-Uniform Gap

In a uniform gap the field is the same everywhere, so the criterion for breakdown is met at every point simultaneously and the whole gap goes at once. Non-uniform geometries — a transmission conductor above ground, a point facing a plane, a sharp protrusion on an otherwise smooth electrode — behave quite differently. As Chapter 2 showed, the field crowds wherever the radius of curvature is small, reaching a peak \(E_{\max}\) at the sharp electrode that may be many times the average \(E_{\text{mean}} = V/d\). The ratio is the field-enhancement factor, and its reciprocal the utilization factor:

Enhancement and utilization (from Chapter 2)
\[ f = \frac{E_{\max}}{E_{\text{mean}}}, \qquad \eta = \frac{1}{f} = \frac{E_{\text{mean}}}{E_{\max}} \]
point plane field crowds at tip E distance from tip → E_max
In a point–plane gap the field peaks sharply at the tip; only there can the breakdown criterion be met while the rest of the gap remains under-stressed

The consequence is decisive. When the voltage is raised, the breakdown criterion is reached at the sharp electrode long before the average field is anywhere near the breakdown value. The gas there ionizes and glows — but the avalanche, advancing into the gap, soon runs into a region where the field has fallen below what it needs, and it stalls. The result is not a spark across the gap but a localized, persistent discharge clinging to the stressed electrode. That is corona.

Section 7-2

Corona: Breakdown in Part

Corona is a self-sustained discharge confined to the high-field region around a sharp or thin electrode, where the gas is fully ionized, while the remainder of the gap stays an insulator. It announces itself unmistakably: a faint bluish-violet luminous sheath, a hissing or crackling sound, and the sharp smell of ozone. Electrically it draws a small current and dissipates power, yet it does not bridge the gap — it is breakdown arrested partway.

Corona is self-limiting precisely because the field falls away from the electrode. The ions it creates form a space-charge cloud that pushes back on the field at the electrode surface, throttling the discharge; the corona settles into a steady or pulsing equilibrium rather than running away. Raise the voltage further, however, and the ionized region extends deeper into the gap, the space charge reorganises, and eventually a streamer (Chapter 6) escapes the corona envelope and crosses to the far electrode — at which point the partial discharge becomes a full spark-over. Corona is thus both a nuisance in its own right and the herald of breakdown in any non-uniform gap.

ground corona glow fully ionized here only bulk gap stays an insulator
Corona is a partial breakdown: the gas ionizes only in the thin high-field sheath at the conductor, while the rest of the gap holds off
Section 7-3

Corona Inception and Peek's Law

The voltage at which corona just begins is its inception (or critical disruptive) voltage. From his classic experiments on parallel conductors, F. W. Peek gave an empirical law that remains the engineer's working tool. Corona starts when the surface gradient reaches a critical disruptive value \(g_0\) — about \(30~\mathrm{kV/cm}\) peak, or \(21.1~\mathrm{kV/cm}\) rms, for air at standard density. For two parallel conductors of radius \(r\) spaced \(d\) apart, the disruptive critical voltage to neutral is:

Peek's disruptive critical voltage (per phase, to neutral)
\[ V_{d} = g_{0}\,m\,\delta\,r\,\ln\!\left(\frac{d}{r}\right) \]

Two correction factors carry the real-world departures from the ideal. The surface-irregularity factor \(m\) accounts for the conductor not being a perfect smooth cylinder: \(m=1\) for a polished wire, around \(0.85\) for a weathered stranded conductor, lower still for dirty or damaged surfaces. The air-density factor \(\delta\) corrects \(g_0\) for pressure and temperature away from standard, since a thinner atmosphere ionizes more easily:

Air-density factor (b in cm Hg, t in °C)
\[ \delta = \frac{3.92\,b}{273 + t} \qquad (\delta = 1 \text{ at } 76~\mathrm{cm\,Hg},\ 25^{\circ}\mathrm{C}) \]

Corona becomes visible at a somewhat higher voltage than it begins, because a glow bright enough to see needs a little more ionization. Peek captured this with the visual critical voltage, which adds a curvature term and uses a slightly smaller irregularity factor \(m_v\):

Peek's visual critical voltage
\[ V_{v} = g_{0}\,m_{v}\,\delta\,r\left(1 + \frac{0.301}{\sqrt{\delta\,r}}\right)\ln\!\left(\frac{d}{r}\right) \]
The design lever is the radius. Because \(V_d \propto r\,\ln(d/r)\), the single most effective way to raise the corona threshold is to use a fatter conductor — or, where a solid conductor would be wastefully heavy, a bundle of sub-conductors that behaves electrically like one large one. This is exactly why every EHV and UHV line you see strings two, four, or more conductors per phase.
Section 7-4

The Polarity Effect

Non-uniform gaps behave differently depending on the sign of the stressed electrode — a subtlety that has no counterpart in a uniform gap and that matters enormously for insulation design. Consider a point–plane gap with the point at high voltage.

Positive point. Corona ignites at a relatively high voltage, but once breakdown does set in it crosses the gap at a lower voltage than the negative case. The reason is space charge: electrons created near the positive point are swept instantly into it, leaving a slow positive-ion cloud just ahead of the tip. That cloud effectively extends the anode into the gap, pushing the high field outward toward the cathode and helping a streamer march across. The positive point is therefore the dangerous polarity — it sparks over first.

Negative point. Corona begins at a lower voltage (the well-ordered Trichel pulses of the next section appear early), but full breakdown needs a higher voltage. Here the positive ions are drawn back into the negative point while electrons drift away into a region of falling field and attach; the resulting space charge weakens the field ahead of the tip and chokes the streamer's advance.

+ point + cloud extends field → sparks over EARLIER − point space charge chokes field corona earlier, sparks LATER
The polarity effect: the positive point ignites corona later but sparks over earlier; the negative point ignites corona earlier but holds off longer

This is why high-voltage apparatus is tested, and insulation coordinated, against the polarity that breaks down first — usually positive for rod–plane geometries under switching surges. The lesson recurs throughout Parts 6 and 7 of this course.

Section 7-5

The Forms of Corona

Corona is not one discharge but a family, and which member appears depends on polarity and voltage. With a negative point, the discharge organises into a remarkably regular train of current pulses — the Trichel pulses — at frequencies from kilohertz up to megahertz. Each pulse is one small burst of ionization that is promptly snuffed out by the negative-ion space charge it creates; once that charge drifts clear, the next pulse fires, and the rhythm repeats. Raising the voltage quickens the pulses until, near breakdown, the discharge passes into a steady negative glow.

i time → regular Trichel pulses period T
Negative corona fires as a regular train of Trichel pulses — each a self-quenching burst, repeating once its space charge clears

With a positive point, corona instead takes the form of intermittent burst pulses and outward-reaching streamers that flicker into the gap, with no comparably tidy periodicity; as the voltage climbs these streamers lengthen until one bridges the gap. On a power conductor under AC, both polarities occur each cycle, and the corona breathes in step with the voltage — generating activity twice per cycle and stamping its signature on the current, the radio spectrum, and the audible hum.

Section 7-6

The Costs of Corona

On a transmission line corona is rarely a safety threat, but it is a chronic and expensive nuisance with four distinct costs.

Power loss. Every corona discharge dissipates energy, drawn continuously from the line — a real \(I^2R\)-like loss that, on a long EHV line in foul weather, can run to tens of kilowatts per kilometre per phase and noticeably dent transmission efficiency.

Radio and television interference (RI/TVI). The fast current pulses of corona radiate across a broad band of radio frequencies, polluting reception for kilometres around the line. Limiting RI is often the binding constraint on conductor surface gradient, stricter even than the loss limit.

Audible noise. The crackle and hum — especially the characteristic 100/120 Hz tone and broadband hiss in wet weather — is an environmental factor in line routing near habitation.

Chemical action. Corona converts atmospheric oxygen to ozone and produces nitrogen oxides; in moist air these form corrosive nitric acid that attacks fittings, and the ozone itself degrades nearby organic insulation.

Section 7-7

Corona on Transmission Lines

Once the line voltage exceeds the disruptive critical voltage, corona loss sets in and grows steeply. Peek's empirical formula gives the fair-weather loss per phase per kilometre:

Peek's corona-loss formula (kW / km / phase)
\[ P_{c} = \frac{241}{\delta}\,(f+25)\sqrt{\frac{r}{d}}\,\bigl(V_{ph} - V_{d}\bigr)^{2}\times 10^{-5} \]

with \(f\) the frequency in Hz, \(r\) and \(d\) in cm, and the phase voltage \(V_{ph}\) and disruptive voltage \(V_d\) in kV to neutral. Two features guide design. First, the loss is zero until \(V_{ph}\) exceeds \(V_d\), then climbs as the square of the excess — so keeping the operating gradient just below the threshold is hugely rewarding. Second, foul weather (rain, snow, fog) collapses \(V_d\) and can multiply the loss many-fold, which is why lines are rated for both fair- and foul-weather corona.

Pc V_ph → V_d fair weather V_d' foul weather
Corona loss is zero below \(V_d\) then rises as \((V_{ph}-V_d)^2\); foul weather lowers \(V_d\) and steepens the climb

The engineer's levers all trace back to the surface gradient: increase the conductor radius, bundle the sub-conductors, polish or grade the surface to raise \(m\), and keep \(V_d\) comfortably above the operating voltage. These same measures, taken to suppress corona, also tame the radio interference and audible noise — so corona control quietly shapes the whole mechanical design of an EHV line.

Section 7-8

Worked Examples

1 Air-density factor

Problem. Find the air-density factor at a hill station where the barometric pressure is \(74~\mathrm{cm\,Hg}\) and the temperature is \(40^{\circ}\mathrm{C}\), and say how it shifts the corona threshold.

Solution. Apply \(\delta = 3.92\,b/(273+t)\):

Working
\[ \delta = \frac{3.92\times 74}{273+40} = \frac{290.1}{313} = 0.927 \]

Since \(V_d \propto \delta\), the disruptive critical voltage falls by about 7%: thinner, warmer air ionizes more easily, so corona starts at a lower voltage — a real concern for high-altitude lines.

2 Disruptive critical voltage (Peek)

Problem. A conductor of radius \(r = 1~\mathrm{cm}\) has spacing \(d = 200~\mathrm{cm}\), \(\delta = 1\), surface factor \(m = 0.85\), and \(g_0 = 21.1~\mathrm{kV/cm}\) (rms). Find the disruptive critical voltage to neutral.

Solution. Use \(V_d = g_0\,m\,\delta\,r\,\ln(d/r)\):

Working
\[ V_d = 21.1\times 0.85\times 1\times 1\times \ln(200) = 17.94 \times 5.298 \approx 95~\mathrm{kV} \]

Corona begins at about 95 kV to neutral. A 220 kV line (\(127~\mathrm{kV}\) to neutral) would be well into corona on this conductor — a clear case for a fatter conductor or a bundle.

3 Visual critical voltage

Problem. For the same line, with visual factor \(m_v = 0.72\), find the voltage at which corona becomes visible.

Solution. Use \(V_v = g_0 m_v \delta r\,(1 + 0.301/\sqrt{\delta r})\ln(d/r)\), with \(\sqrt{\delta r}=1\):

Working
\[ V_v = 21.1\times 0.72\times 1\times(1+0.301)\times 5.298 = 15.19\times 1.301\times 5.298 \approx 105~\mathrm{kV} \]

Visible corona appears at about 105 kV — some 10 kV above the disruptive onset, confirming that corona starts (and starts dissipating power) before the eye can see it.

4 Corona loss (Peek)

Problem. The line of Example 2 operates at \(V_{ph} = 120~\mathrm{kV}\) to neutral, \(f = 50~\mathrm{Hz}\). With \(V_d = 95~\mathrm{kV}\), \(r = 1~\mathrm{cm}\), \(d = 200~\mathrm{cm}\), \(\delta = 1\), estimate the fair-weather corona loss per phase per km.

Solution. Apply Peek's formula with \(\sqrt{r/d}=\sqrt{1/200}=0.0707\) and \((V_{ph}-V_d)^2 = 25^2 = 625\):

Working
\[ P_c = \frac{241}{1}(50+25)(0.0707)(625)\times 10^{-5} \approx 8.0~\mathrm{kW/km/phase} \]

About 8 kW per km per phase — and since it grows as \((V_{ph}-V_d)^2\), pushing \(V_d\) up by even a few kV (a slightly larger conductor) would cut this loss sharply.

Review

Chapter Summary

Local first

A non-uniform field peaks at the sharp electrode, so the breakdown criterion is met there long before the average field is critical.

Corona

A self-limiting partial breakdown confined to the high-field sheath; the bulk gap still holds. It is the herald of full spark-over.

Peek's law

\(V_d = g_0 m\delta r\ln(d/r)\), with \(g_0\approx 21.1\,\mathrm{kV/cm}\) rms; the visual onset is a little higher. Radius is the design lever.

Polarity effect

A positive point ignites corona later but sparks over earlier; a negative point ignites earlier but holds off longer. Test the dangerous polarity.

Forms

Negative corona fires as regular Trichel pulses; positive corona as burst pulses and streamers. Both ride the AC cycle on a line.

The costs

Power loss \(\propto (V_{ph}-V_d)^2\), radio interference, audible noise and ozone. Bundling and larger radii tame them all at once.

Practice

Problems

For each item, first name the idea it tests — field enhancement, corona inception, Peek's law, the polarity effect, the forms of corona, or corona loss — then work it through. Difficulty rises down the list.

  1. Explain why a non-uniform gap breaks down at the sharp electrode first, and define the field-enhancement factor in your answer.
  2. Why is corona described as a "self-limiting" or "partial" breakdown? What stops it bridging the gap at inception?
  3. Compute the air-density factor for \(b = 70~\mathrm{cm\,Hg}\), \(t = 35^{\circ}\mathrm{C}\), and state its effect on the disruptive voltage.
  4. A smooth conductor has \(r = 1.2~\mathrm{cm}\), \(d = 250~\mathrm{cm}\), \(m = 0.9\), \(\delta = 1\), \(g_0 = 21.1~\mathrm{kV/cm}\). Find the disruptive critical voltage to neutral.
  5. State, with reasons, which polarity of a rod–plane gap sparks over at the lower voltage, and why this matters for impulse testing.
  6. Describe a Trichel pulse: which polarity produces it, what quenches each pulse, and how the repetition rate changes with voltage.
  7. List the four principal costs of corona on a transmission line, and identify which one often sets the binding limit on surface gradient.
  8. For the conductor of Problem 4 operating at \(V_{ph} = 150~\mathrm{kV}\) to neutral, \(f = 50~\mathrm{Hz}\), estimate the fair-weather corona loss per km per phase using Peek's formula.
  9. Using \(V_d \propto r\ln(d/r)\), explain why a bundle of two sub-conductors raises the corona threshold more effectively than simply increasing the spacing \(d\).
  10. A line just below its disruptive voltage in fair weather goes well into corona in rain. Explain in terms of \(V_d\), \(m\) and the \((V_{ph}-V_d)^2\) loss law why foul-weather loss rises so steeply.
Tip: the whole chapter hangs on one inequality being satisfied somewhere but not everywhere. In a uniform gap the breakdown criterion is met everywhere at once, so there is no corona — just a spark. In a non-uniform gap it is met only in the small high-field region, which ionizes and glows while the rest holds; that is corona, and pushing the voltage higher is just a matter of extending the criterion-satisfying region until a streamer finally reaches the far electrode. Peek's law is simply that local criterion written for the geometry of a power line. Carry that picture into Part 3, where the medium changes from gas to liquid and solid but the contest — local stress versus local strength — stays exactly the same.