PREVENTION OF MOISTURE DEPOSITION AND ICE FORMATION ON WIRES BY MEANS OF CORONA DISCHARGE

A. A. VOROB'EV and N. B. BOGDANOVA

Introduction

It is known that solid and liquid particles suspended near a corona wire adsorb ions and, having become charged with charges of the same sign as the wire, are repelled from its surface. The velocity of motion of such charged particles toward the opposite electrode is determined by the magnitude of the maximum charge \(q_{max}\) (a particle of radius \(\rho\) with dielectric constant \(\varepsilon\) in a field of intensity \(E\)) [4, 5].

\[ q_{max}=\left(1+2\frac{\varepsilon-1}{\varepsilon+2}\right)E\rho^2. \]

If corona of the wire occurs in an atmosphere of fog, rain, or snow, then it is possible that water droplets, fog particles, and ice crystals will become charged and interact with the corona wire. In this case a fog particle (water, snow) will move in the direction of the opposite electrode and will no longer reach the corona wire. As a result, the character of precipitation coverage on corona wires may differ from that on non-corona wires or on wires not under voltage at all: the corona on the wire may impede the deposition of moisture on it and the formation of glaze ice.

Experimental installations

The study of moisture deposition on corona wires was carried out by us on installations: 1) a corona wire inside a cylinder coaxial with it (cylinder length 30 cm, diameter 8 cm, wire diameter \(2r_0\) varied from 0.1 to 4.95 mm; at the ends, rings 4 cm long adjoined the cylinder and served as guard rings);

2) a corona wire located opposite a grounded plane parallel to it (plane length 110 cm; the distance to the wire varied from 10 to 20 cm; the wire diameter varied from 0.1 to 2.5 mm; the edges of the plane were rounded);

3) two parallel corona wires (wire length 170 cm; the distance between them varied from 10 to 20 cm; wire diameter—from 0.1 to 2.5 mm; at the ends the wires were embedded in hemispheres).

The study of icing of corona wires was carried out on the same installations, but at an air temperature below \(0^\circ\)C, and also with artificial cooling in special refrigerating chambers. The sche-

A schematic representation of one of such chambers is shown in Fig. 1, where $d$ and $c$ are tin boxes soldered one into the other; the space between them was filled with a cooling mixture. A glass tube $f$ wrapped in foil was inserted into the chamber (its diameter was 35 mm); the corona electrode—a cylindrical rod with a small ball at the end—was inserted on a cork into the tube (the rod diameter was 4.95 mm). Opening $a$ was for the thermometer; a stable temperature inside the chamber was maintained at $-5^\circ$C.

The experiments were carried out both with direct and with alternating current. The critical corona voltage in the direct-current experiments was determined from the reduced characteristics. The voltage was measured by means of spherical spark gaps, and the current by a galvanometer connected into the grounding wire. The electrical circuit of the apparatus for the direct-current experiments is shown in Fig. 2, where $T$ is an 80 kv transformer, $K$ is a kenotron, the corona wires (or wire), $G$ is a galvanometer with a sensitivity of

Fig. 1. Refrigerating chamber.                                     Fig. 2. Electrical circuit of the apparatus.

1.7 $\mu$A/div., recording the corona current. A system of shunts $R_{\mathrm{sh}}$ (10 Ω, 5 Ω, 3 Ω) was connected in parallel with the galvanometer, making it possible to vary the sensitivity of the galvanometer over wide limits; a neon lamp Н. Л., connected in parallel with the resistance $R_6$ and the galvanometer $G$, served to protect the galvanometer from a large current in the event of possible breakdowns between the electrodes; $R_1 = R_2 = 10^5$ Ω are protective resistances, $F$ is a spherical spark gap. Some of the experiments were carried out with a capacitance $C = 0.01\,\mu$F connected in parallel with the electrodes.

In the experiments with corona on alternating current, the same transformer $T$ served as the voltage source. The current was measured by means of galvanometer $G$, using the rectifying action of the corona itself [6, 1].

Artificial rain was produced by means of sprayers filled with distilled water and giving a maximum rain intensity of 1 mm/min. During spraying, a uniformly falling vertical rain was produced, enveloping the corona wires. In experiments with coaxial cylinders, the jet of sprayed water was directed along the wire.

Description of the experiments and their results

The surface of the corona wires was observed through the observation tube; the number of droplets depositing on the section of wire located in the field of view of the tube was recorded, and the dependence of this number on the magnitude of the applied voltage was plotted.

In the icing experiments, the ice forming on the wire at a given voltage was knocked off and weighed; the change in this quantity was determined as a function of the magnitude of the applied voltage. In interesting cases, sketches and photographs were made. In parallel with all these observations, volt–ampere characteristics were taken for the corona discharge on wires under normal conditions \((I_0=f(U))\) and in rain or under icing conditions \((I=f(U))\). The volt–ampere characteristics of the corona under these conditions are of independent interest to us [2]. Their analysis also makes it possible to conclude whether the corona prevents moisture from depositing on the wires. Experiments showed that the corona current \(I\), measured when precipitation is present on the wires, is greater than the corona current \(I_0\), measured under normal conditions; if the corona prevents the wires from being covered by precipitation, then as the voltage \(U\) increases above the critical corona voltage \(U_{\mathrm{cr}}\), the current \(I\) should tend toward the values of the current \(I_0\).

As a result of our investigations, it could be established that, as the voltage is increased above the critical corona voltage \(U_{\mathrm{cr}}\), moisture is deposited on the wires in ever smaller quantities, and at a certain voltage \(U_c\) the corona-producing wires remain dry while in rain and do not ice up at low temperature.

We shall present the results of visual observations of the surface of a corona-producing wire for the case in which the wire is placed opposite a plane parallel to it. A viewing tube was set up opposite the wire, and some section of it was strongly illuminated. Every drop landing on this section was clearly visible in the tube. As the applied voltage was increased, the number of drops in the field of view of the tube decreased. Fig. 3 shows this dependence. Here the abscissa axis gives the voltage \(U\) in \(kv\), and the ordinate axis gives the number of drops \(n\) in the field of view of the tube \((l=5\ \text{cm})\); the dimensions of the wire and its distance to the vertical plane are indicated in the figure; the voltage was constant, and the wire was at negative polarity.

Fig. 3. Dependence of the number of drops deposited on the wire on voltage.

At a given voltage and a constant rain intensity, the number of drops on the wire is fairly constant; likewise, this number changes little with time: during the first 1–2 minutes after spraying begins it increases, and then remains unchanged. This equilibrium is dynamic: some drops fly off, others deposit again. The drops that fly off do not leave the wire completely, but, gradually spraying upward like a fountain, dry out. When one drop flies off, another lands. As the voltage is increased, a new equilibrium regime is established with another, smaller number of drops. Finally, at a certain voltage, drops do not deposit on the wire at all. On a plane arranged horizontally beneath the wire parallel to it, at a voltage \(U>U_{\mathrm{cr}}\), a dry track is clearly visible, onto which no drops fall. This track extended along the entire length of the wire, and on it the curtain of raindrops rising above the wire was, as it were, closed (inside which only individual drops fell). In order to make this track clearly visible, we dyed the sprayed water and placed paper on the plane. An imprint was obtained, a photograph of which is shown in Fig. 4.

(wire diameter \(2r_0 = 0.2\) mm, distance to the plane \(H = 9.0\) cm, \(U = 22\) kV, track width 3 cm. Approximately the same pattern of moisture deposition was also observed in other setups; as the applied voltage \(U\) was increased, droplets were deposited on the wires in ever smaller quantities.

Observation of the surface of corona wires has established that increasing the voltage above the critical value corresponds to a decrease in the amount of ice formed. Figures 5 and 6 show photographs of a wire (\(2r_0 = 0.75\) mm) over which water was sprayed for half an hour at an air temperature \(t = -8^\circ\)C. The wire was inside a coaxial cylinder (the latter was open for the convenience of photography). In Fig. 5 it is seen that the wire, in the absence of voltage (\(U = 0\)), became covered with ice along its entire length; the thickness of the ice layer decreases with increasing distance from the sprayer (the latter is on the right); the greatest thickness of the ice layer reaches 2–2.5 mm. The results of repeating this same experiment when the wire was under a constant voltage \(U = 18\) kV are shown in Fig. 6. On the wire passing inside the cylinder and the ring there is no ice at all; ice covered only that part of the wire which protruded from the ring, i.e., on which there was no corona. The critical corona voltage \(U_{\mathrm{cr}}\) for this case is 14 kV; the wire was at negative polarity.

Fig. 4. “Track” under a corona wire

The dependence of the amount of ice \(Q\) deposited on the wire on the voltage \(U\) was plotted. For the case of a wire corona-dis charging against a plane parallel to it under constant voltage, the results are given in Fig. 7. Here the abscissa axis gives the magnitude of the applied voltage \(U\), and the ordinate axis gives the amount of ice \(Q\), in mg, formed during a 10-minute exposure. It is interesting to note the circumstance that os-

… the main decline of the curve begins at a voltage lower than the critical value, as determined from the reduced discharge characteristics for a dry wire.

Fig. 5. Icing of a wire located inside the cylinder, in the absence of voltage \((U = 0)\).

water. Apparently, here the effect takes place owing to the early corona from the irregularities of the wire. On a wire with diameter \(2r_0 = 1.55\ \text{mm}\), likewise dis—

Fig. 6. Icing of a corona wire \((U = 18\ \text{kV})\).

placed opposite the plane at a distance \(H = 10\) cm, in the absence of voltage 210 mg of ice formed; at a voltage \(\sim U_{\mathrm{cr}} = 30\) kV the amount of ice proved to be only 10 mg, and \(Q = 0\) was reached only at \(U = 55\) kV.

The study of the current-voltage characteristics of the corona discharge showed that the corona current \(I\) on wires exposed to rain at small voltages is always greater than the corona current \(I_0\), measured under normal conditions. As the voltage increases, the current \(I\) tends toward the values of the current \(I_0\) (as the voltage \(U\) increases, fewer drops settle). In the arrangements: a wire placed opposite a plane, and two parallel corona wires, at a voltage \(U > U_c\) (when moisture no longer deposited on the wire) \(I \simeq I_0\); in coaxial cylinders at \(U > U_c\), the current \(I\) was invariably less than \(I_0\).

Fig. 7. Dependence of the amount of ice forming on the wire on the magnitude of the applied voltage.

Fig. 8. Dependence of the ratio \(\dfrac{U_c}{U_{\mathrm{cr}}}\) on the wire diameter \(2r_0\) for the case of a wire corona-discharging inside a cylinder coaxial with it.

From the current-voltage characteristics and from visual observations, the values of the voltage \(U_c\) were determined at which the corona wires are no longer covered with deposits. The ratio \(\dfrac{U_c}{U_{\mathrm{cr}}}\) proved to be different for all three configurations considered and, for each given configuration, changed little with variation of the wire diameter. In Fig. 8 the dependence of \(\dfrac{U}{U_{\mathrm{cr}}}\) on the diameter of the corona wire is given for the case of a wire placed inside a cylinder coaxial with it (the voltage is constant, the wire is at negative polarity—

... As can be seen, \( \frac{U_c}{U_{cr}} \) is equal to 1.3 and does not depend on \(r_0\). Connecting a capacitance that smooths the voltage pulsations reduces this value to 1.1–1.2. With positive polarity of the wire, \( \frac{U_c}{U_{cr}} = 1.7 \) was obtained. Approximately the same value of \( \frac{U_c}{U_{cr}} \), and even a somewhat higher one, is obtained under the action of alternating voltage. For the case of a wire placed opposite a plane, with negative polarity of the wire, \( \frac{U_c}{U_{cr}} = 1.6 \); with positive polarity, \( \frac{U_c}{U_{cr}} = 1.8 \); and, finally, for two parallel wires under the action of direct voltage, \( \frac{U_c}{U_{cr}} \) proved to be somewhat greater than 2.

The number of droplets deposited on the wire at a given voltage depends on the intensity of the rain, although this dependence is not sharp. The value of \(U_c\), however, changed almost not at all with changes in rain intensity, at least within the limits of variation of this quantity available to us—from fractions to \(1\ \mathrm{mm/min}\).

Conclusion

Thus, corona on a wire prevents it from being covered by precipitation. The voltage \(U_c\), at which a corona-producing wire is not covered by precipitation, proved to be smallest in the case of a unipolar corona with negative polarity of the corona-producing wire.

The picture of the behavior of droplets near the surface of a corona-producing wire appears to us as follows. The corona-producing wire is surrounded by a corona sheath, beyond the limits of which there is localized a space charge of the same sign as the sign of the corona-producing wire. A water droplet entering the zone of such a charge behaves in the same way as a dust particle in the ionic current of an electrostatic precipitator: a neutral droplet becomes charged; if the droplet previously had a charge of the same sign as the space charge of the external region, its magnitude increases; if the droplet previously possessed a charge opposite in sign to the space charge of the external region, it is recharged [3]. In addition, an electric wind acts on the droplet.

All the phenomena that we observed—the “curtain” of droplets above the corona-producing wire; the dry path beneath the wire, onto which droplets do not fall; the cleaning of air from droplets in coaxial cylinders; and, finally, the fact that at \(U > U_c\) the wires remain dry in the rain and do not ice over at low temperature—are the result of the combined action of the effects indicated.

References Cited

1. V. F. Pyk — Dielectric Phenomena in High-Voltage Engineering, ONTI, 1934.
2. N. B. Bogdanova — Proceedings of TPI, 1948.
3. N. N. Tunitskii, N. V. Tikhomirov, and V. V. Petryanov — ZhTF, vol. 10, no. 20, 1940.
4. N. Rohman — Z. f. Phys., 17, 253, 1923.
5. Pautenier and M. Hannot — Journ. d. Phys. et le Rad., 3, No. 12, 590, 1932.
6. W. G. Hoover — El. Eng., 55, No. 5, 1936.