Some Physical Problems of Electric Power Engineering1

J. Slepian, East Pittsburgh, Pennsylvania

The electric-power industry was born in Faraday’s laboratory of physical research, and all its achievements since its inception represent, to a greater or lesser degree, confirmation of Faraday’s law of electromagnetic induction.

Unlike mechanics and civil engineering, electrical engineering did not exist before the development of the corresponding branch of physics, and the first electrical engineers were great physicists—Weber, Kelvin, and others. The generally accepted electrical units—the volt, ampere, etc.—are named in honor of famous physicists. The inventor of the dynamo, Hopkinson, turned to Kelvin for explanations concerning the forces acting on conductors in a space between poles. The mathematician and theoretical physicist Poincaré[^2] analyzed the question of the magnetic energy of alternating current.

I shall have to begin with a definition and delimitation of the concepts I use, “physics” and “electric power engineering.”

By the term “electric power engineering” I shall understand only the production and distribution of electrical energy, excluding its consideration of use; otherwise I would have to deal with an excessive abundance of examples of solved and unsolved physical problems. I shall confine myself to the question of the distribution of low-frequency electric power, so as not to encroach on the field of communications, which deals with electric power of higher frequencies. In addition, I shall consider only the mechanical method of obtaining electric power, i.e., I exclude from consideration batteries, photocells, and other sources of low-frequency electrical energy.

It will be more difficult to define the term “physics.” Beginning with Faraday, everything, in essence, was physics.

At the basis of electrical engineering lies a macroscopic, phenomenological consideration of electromagnetic phenomena in stationary and moving media. The principal role in electrical engineering is played by the field equations, which describe the relations among the various electromagnetic vectors and scalars; the equation of ponderomotive forces arising in an electromagnetic field,

\[ \mathbf{F}=\varepsilon \mathbf{E}+\frac{1}{c}[\mathbf{I},\mathbf{B}], \]

and the supplementary Maxwell equations concerning the various electric vectors in matter, characterized by certain constants,

\[ \mathbf{D}=\varepsilon \mathbf{E},\quad \mathbf{B}=\mu \mathbf{H},\quad \mathbf{I}=\sigma \mathbf{E}, \]

and so forth. Therefore I shall assign to electrical engineering all those cases in which the properties of matter are characterized by such material constants as permeability, conductivity, etc.; and only those cases where a deeper penetration into the nature of the electrical structure of matter is required I shall assign to the field of physics.

One may object that, for example, in studying the magnetic properties of iron

engineer cannot be satisfied with a simple descriptive constant—permeability—but must also consider saturation and hysteresis. However, I use the term “constants of matter” in a more general sense. The family of hysteresis curves, from this point of view, expresses the magnetic properties of a substance in as superficial (although sufficiently accurate for most technical purposes) a manner as, for example, the equation \(\mathbf{B}=\mu\mathbf{H}\). Only questions touching on such phenomena as the structure of the crystal lattice, the spin of the electron, and the like, will I assign to “physics.”

I hoped that I could assess the role of physics in the electric-power industry by examining the contents of the leading electrical-engineering journals of the United States and other countries in recent years. However, I found that this would be too great a task. In fact, in volumes 53 and 54 alone of Transactions of the American Institute of Electrical Engineers for 1934 and 1935, I found 77 articles in the field of physics. Of these, 13, written by physicists, were educational in character and did not describe any new experimental work. It is noteworthy that so much space in an electrical-engineering journal is devoted to keeping readers abreast of contemporary ideas and achievements in theoretical physics.

The contents of the remaining 64 articles compel one to state that a large part of the research being conducted in electrical engineering is physical research. Of these 64 articles, 8 are devoted to the electrical properties of vacuum, electron motion, and electronic devices; 31 investigate the electrical properties of gases used in electrical engineering; 8 consider electrical phenomena in dielectrics (conductivity, breakdown conditions in insulating liquids and solids); 3 investigate magnetic materials; 14 study various questions: the formation of moisture films on the surface of a dielectric, sound absorption, and processes occurring during the heating of bodies.

Although these 2 volumes of the Transactions may not be typical, one may nevertheless say with confidence that the greater part of physical research in present-day electrical engineering concerns discharges in gases. It is surprising that the electrical properties of conductors are not represented in these studies. The new achievements of the theory of metals and semiconductors have evidently not yet reached the electrical engineer.

The physicist working in the electric-power industry uses electron tubes and low-current gas-discharge tubes for measuring the large currents consumed in power installations and for controlling them. Here one may mention electronic regulators, controllers, and other devices that have become widely used. Photocells, known as the “electric eye,” are used in industry in all kinds of counting, sorting, and control operations. I shall not, however, dwell on these examples, since I wish, in accordance with what was stated earlier, to remain within the bounds of the field of the production and distribution of electrical energy.

Let us consider that field of physics to which the largest number of the articles I have reviewed belongs, i.e., the electrical properties of gases, and let us examine where and how the power engineer encounters them.

Electrical discharges in gases appear in all parts of the power system, beginning with the generator and ending with the user’s network. In some cases they act as a destructive factor, causing damage to electrical apparatus; in others, on the contrary, they are used as means for attaining useful purposes. Most often, both of these roles—the positive and the negative—accompany one another, for example, lightning and the protective discharge in the spark gap of a lightning arrester.

Figs. 1–6 show various examples of the use of electrical discharges in gases in the power system. In Fig. 1 a diagram is given

of a large power system. A generator driven by a steam or water turbine produces alternating current at a voltage usually of about 15 kV. This voltage is transformed to 22–287 kV for the transmission of energy over long distances. Then comes a distribution system

Fig. 1

Labels in the figure: Generator; 15 kV; high-voltage transmission line; 22–287 kV; transformers; medium-voltage distribution network; 11–66 kV; low-voltage distribution network; 110–220; rectifier; 600–3000 V; railway, electrochemistry, etc.

at a somewhat lower voltage (11–66 kV) for shorter distances; then follows the local distribution system at 2–4 kV; finally, there is the domestic or industrial network at 110–600 V.

Let us begin with the generator (Fig. 2). Here we have air in contact with the commutator and brushes. As we shall see below, during switching there occurs a discharge between the brush and the receding segment of the commutator. For the operation of the generator, the electrical properties of gases are essential because in the insulation of the generator there are always air gaps—

Fig. 2

Labels in the figure: Commutation; oil air circuit breaker; 15,000 V; generator; corona discharge in the winding.

between them. A corona discharge in a small air gap between an insulated winding and the surface of the metal is one of the most difficult problems arising in generator design. Then the drawing shows a powerful switch functioning by means of an arc between separating electrodes. In the switch shown in the drawing a series of short arcs with cold cathodes is formed, the possibility of which was discussed and established by K. T. Compton,¹ H. Stoll and others.

The resulting current is transformed and enters a high-voltage transmission system (Fig. 3), in which the insulating properties of air in its normal state are used.

Fig. 3

Labels in the figure: “22–287 kV”; “dividing grids”; “high-voltage oil circuit breaker”; “insulating gap”; “ground”; “transformer.”

The construction of a transformer bushing insulator must be designed so that the air surrounding it under normal operating conditions is an insulator; under overvoltage, however, breakdown through the air must protect against breakdown through the transformer insulation, and the discharge path that occurs in this case must not pass too close to the porcelain insulator, so as not to be dangerous with respect to heating. These complex problems of spark discharge at insulators have so far been solved purely empirically and require a good physical theory.

In an oil circuit breaker the arc burns in a bubble of gas formed by the decomposition of the oil. In operation with high voltage, the conditions for forming the bubble must be such that even a short arc is capable of performing the task of interrupting the current at its nearest passage through zero after the contacts have separated.

The transmission line passes through atmospheric air, where lightning sometimes occurs—a powerful high-voltage discharge. By a direct or inductive path lightning causes the appearance of strong currents in the transmission system, if a sufficiently perfect protective device is not provided in the latter. Existing protective devices, unfortunately, are still imperfect, and lightning is still the main cause of accidents occurring on transmission lines.

The protective discharge in the spark gap of a lightning arrester must protect the line from the harmful action of lightning. When the lightning has ended, the protective discharge must cease. The principal physical instrument used by the engineer in studying such problems is the cathode tube, and the principal concept is the physical

the theory of ionization in gases and, in particular, Townsend’s theory of breakdowns in gas gaps. The measurements needed for work with high pressures and high breakdown voltages in Townsend’s theory, the necessity of which had been discovered for the first time by the engineer Rogowski,^4 were developed by him in a number of brilliant and elegant investigations and were subsequently elaborated by various physicists.

Although the state of the theory of a simple spark gap under laboratory conditions may be regarded as satisfactory, the volt-ampere characteristic of a spark gap in a lightning arrester still leaves much to be desired. The deviation from the ideal spark-gap characteristic is due to the nonconstancy of the discharge voltage under atmospheric conditions. The reason for this nonconstancy lies in changes in the insulating properties of porcelain under the influence of moisture.

Fig. 4 shows a new type of lightning arrester: a protective tube placed parallel to a section of the suspended line wires. Because of its low cost, this tube should find wide application. Here the spark gap is arranged as follows. The discharge occurs in a fibrous tube. The heat developed during the discharge decomposes part of the fiber with the liberation of gas. The deionizing action of the latter, owing to the high pressure and velocity, is sufficient to extinguish the arc at the next passage of the current through zero. The principal difficulty here is that the tube must withstand the high pressures developed owing to the abundant generation of gas at large currents and, at the same time, at weak currents must furnish an amount of gas sufficient to extinguish the arc.

Fig. 4

The energy flux arising owing to the magnetic and electric fields created around the transmission line must not be too large, so that the air does not lose its insulating properties. This requirement fundamentally determines the cross section of the conductor, while, for economy of material, the conductor is made hollow. In terms of economic requirements, the operating conditions of the lines are close to the limit at which a weak corona discharge arises. The corona discharge at conductors has been studied theoretically and experimentally by engineers and physicists; in particular, a complete theory of it was given by Holm.^5 However, in the case of conductors coated with oxide and weathered, we still remain at the stage of empiricism. A more detailed study of discharge under these conditions would be highly desirable.

An exhaustive enumeration of the various examples of the application of gas discharges in the power system would be excessively tedious.

Fig. 4 shows a cable used in a medium-voltage distribution network. In order to eliminate the harmful action of ionization in cracks in cables, they are subjected to thorough impregnation and treat—

processing in oil under pressure. We have powerful fuses in which the action of the arc is improved thanks to the generation of gas from fiber walls or vapor in boric-acid cartridge fuses. We also have other arcs used to rectify alternating current into direct current for electrochemical needs, for railway needs, etc. In the low-voltage distribution system in Fig. 5 we see a transformer with a built-in fuse tube, in principle similar to the fuse tube in the high-voltage line. In a low-voltage street lighting network (Fig. 5) we have such examples as the use of an arc in a vast number of bell push-buttons and the use of ordinary switches of the lighting network. By means of electric discharges there are also carried out switchings that are very widespread in automobiles.

Fig. 5

Examining examples of the use of discharges in gases in an energy system, we easily find problems common to them. In all these cases we see the problem of the inception and the problem of the extinction of a useful discharge. For an alternating-current network, extinction of the discharge is usually understood as preventing the re-ignition of the discharge after the current has passed through zero. Let us discuss the significance of physical theory for solving such problems and also try to point out cases where it is insufficiently perfected or not developed at all.

The inception of a discharge during a high-voltage breakdown has already been discussed when considering the spark gap and the arrester. Here the discharge arises along its entire length. In other applications, the cathode part of the discharge more often arises initially, and then a discharge in the remaining space follows spontaneously. Thus, in vacuum gas-discharge tubes, for example in the thyratron, thanks to the presence of specially heated cathodes, when a sufficiently high voltage is applied to the anode and the grid there are electron formations in the device. Langmuir’s theory[^6] of plasma formation satisfactorily explains the behavior of the discharge in this case, but quantitatively it has been developed only for the simplest form and arrangement of electrodes. In ordinary mercury-arc rectifiers the cathode is constantly maintained in an excited state with the aid of an auxiliary anode.

In most of the devices I describe, it is usually not difficult to maintain the discharge after it has arisen, but in

in discharge tubes with low gas pressure there exists a limit that the current strength cannot overstep. When this limiting value of the current strength, whose magnitude depends on the pressure, is reached, an unstable state sets in, in which the discharge is continuously interrupted and renewed, and a high voltage arises in the inductive circuit. This phenomenon has great practical significance, since it limits the power of mercury-arc rectifiers. Langmuir and Mott-Smith7, Hull and Braun8 connected this phenomenon with the degree of ionization of the gas. I find it possible to explain it by the removal of gas from the tube through the one-sided motion of positive ions.

In mercury-arc lamps of the ignitron type, which are now used in technology, the mathematical theory of the established flow of electricity finds an interesting application. By immersing in mercury a stationary block of material with high resistance, one can produce at the point of contact a concentration of electric field and electric current similar to that which occurs when contacts are parted, and with an analogous result, i.e. with the formation of the cathode spot of an arc. In this type of tube, the limiting value of the current at which instability sets in has not yet been observed, evidently because the anode is located directly on the path of the jet of mercury vapor issuing from the cathode spot.

Fig. 6
1 — deionizing gap, 2 — domestic mains, 3 — industrial mains

It was observed by the engineer Tanberg9 that vapors propagating from the cathode spot at low pressures possess an unusually large mechanical impulse and an energy corresponding to a voltage of over one hundred volts. Tonks10 expressed a number of doubts concerning the reality of the existence of the rapid Tanberg stream, and Meson and Berkey11, apparently under the influence of the opinion of Tanberg, Frisch, and Loeb12, gave a theory based on the assumption of multiple ionization of atoms.

It is quite obvious that the engineer is required to have a good knowledge of the physical theory of the cathode, familiarity with thermionic and cold electron emission, the motion of electrons and ions, etc.

The occurrence of a discharge in switches and circuit breakers takes place so easily with the simple parting of contacts, and the problem of extinguishing the discharge is often so difficult, that attention is not paid to the useful action of the discharge and the arc is regarded as an evil that must be avoided. However, according to Faraday’s law, the sudden reduction of the current to zero at the instant when contacts are parted would lead to the formation of an indefinitely high voltage. Only in the case when the moment of separation of the contacts is so closely synchronized with the instant at which the alternating current passes through zero that the electrostatic capacitance of the system can absorb the instantaneous magnetic energy, does the dangerous high voltage fail to arise. But usually this electrostatic capacitance is so small that unattainable precision of synchronization is required, and only the discharge in the gas frees the engineer from this unpleasantness. Thanks to the discharge, the possibility appears of separating contacts at any value of the current; the arc spontaneously forms

is formed and causes the current to be interrupted at the moment when the current again reaches zero.

Attempts have been made to interrupt current without a discharge. Instead of separating contacts, a resistance was introduced, continuously increasing to a practically infinite value. An example of such a resistance is in fact the brush of a commutator, since the contact surface of the brush and the segment, and consequently also the reciprocal value of the resistance introduced into the circuit, approaches zero continuously during the motion of the segment relative to the brush. Analysis shows, however, that even with a preselected course of variation of the resistance there still remains the necessity of extremely precise synchronization in order to avoid a dangerous overvoltage. Since in practice synchronization is always imperfect, the use of a discharge is nevertheless preferable. With good, although not ideal, synchronization, the discharge current may be very small, so that a quiet discharge occurs, not dangerous and easily capable of passing unnoticed.

The extinction of a useful discharge in an alternating-current network is usually reduced to the problem of preventing the discharge from reigniting after the current has passed through zero. This can be achieved by directly selecting such conditions under which the positive column or the cathode spot, or the cathode glow, cannot form spontaneously again. In high-voltage switches the positive column plays the role. Various means are used so that the positive column has so small a cross-section at zero current that its temperature and degree of ionization fall to a value at which restoration of the column in the next half-period of the current is impossible. Here the leading role is played by K. T. Compton’s theory of thermal ionization of the positive column.^14 The excellent article by Kesselring and Koppelmann, recently published, shows to what extent it is necessary for engineers to understand the complex physics of arcs operating at high current.

The means used to achieve a small cross-section of the positive column are varied. One may mention the barrier that confines the magnetic field in Westinghouse switches with a deionization grid, the motion of oil under pressure in the switches with an oil jet of the General Electric Company, the gas stream in exhaust-type circuit breakers, and switches with a gas jet used in Europe.

In low-voltage switches and in deionization-type switches for higher voltages, conditions are created under which the formation of a cathode column is impossible. With a change of polarity there is no thermionic activity of the new cathode; the discharge must be a glow discharge, at least at first, and, if the available voltage is insufficient to maintain the cathode glow, the discharge does not form again. However, in stable arcs at large currents the voltage required for reignition of the discharge is considerably lower than the minimum voltage for ignition of the discharge under ordinary conditions, and for interruption it is necessary to resort to the aid of the positive column. The theory of this effect has not yet been developed. In deionization-type switches the arc moves rapidly at the surfaces of the anode and cathode, and even at very large currents no decrease in the restriking voltage takes place.

In mercury-arc rectifiers, ignitrons, thyratrons, tubes with a glow discharge, and, in general, in gas-discharge tubes, in order to extinguish the discharge at the moment when the current passes through zero, the absence of a cathode spot and the presence of a voltage smaller than that needed to sustain the glow discharge are necessary.

At low gas pressure the voltage necessary to sustain the glow discharge increases, and thus operation with relatively high voltage becomes possible. However, in all these devices a phenomenon sometimes occurs for which, in my opinion, there is not yet a satisfactory theory.

Namely, despite the absence of the conditions which the theory considers ordinarily necessary, the cathode spot nevertheless forms again after the current passes through zero and causes a short circuit. The statistical frequency of these cases of “restriking” indicates that certain molecular aggregates, perhaps torn from impurities on the surface of the electrodes, or particles undergoing Brownian motion, pass through the gas and enter collision with the anode. There is, however, no detailed theory of restriking caused by the presence of such aggregates. Such cases of restriking at voltages exceeding 30–40 kV are very frequent; therefore the problem of tubes of this type for such high voltage has not yet been solved. In order to accomplish the often-discussed problem of transmitting direct current at high voltage, it is precisely this problem whose solution must be awaited.

It is possible that the use of a positive column or arc discharge at elevated pressure, proposed by Marx[^15] in his rectifier with a gas jet, will prove to be the best solution.

It is not known, however, whether phenomena analogous to restriking are excluded in such rectifiers. If an arc in the switch with a gas jet accidentally ignites again after the current has passed through zero, we shall have current for one additional half-period, which is quite immaterial and usually escapes the operator’s attention. But in a rectifier, restriking of the arc is equivalent to a short circuit. Such abnormal restriking once in a million cycles is already impermissible.

I hope that I have been able to show clearly the manifestation of the phenomena constituting the subject of one of the chapters of physics—electrical phenomena in gases—in every part of an operating power system. I hope that I have also been able to show how indispensable the physical theory developed for solving related technical problems is, and how engineers who became physicists, in their turn, developed the theory. I also hope that, by citing examples in which the imperfection of the theory or its complete absence hampers the design and characterization of an apparatus, I have shown the timeliness of further work by the physicist and the engineer-physicist in the electric-power industry.

References

1. H. Poincaré, Ecl. Electr., 50, 293, 1907.
2. K. T. Compton, Phys. Rev., 21, 266, 1923.
3. H. Stolt, Z. Physik, 26.2, 95, 1924; Ann. d. Physik, 74, 80, 1924.
4. W. Rogowski, Beg. Arch. Electrotechn., 16, 496, 1926.
5. R. Holm, Wiss. Veröff. Siemens-Konzerns, IV, 1, 14, 1925.
6. I. Langmuir, J. Frankl. Inst., 214, 275, 1932.
7. I. Langmuir and Mott-Smith, Gen. Elec. Rev., 27, 770, 1924.
8. R. W. Hull and H. D. Brown, Trans. A. I. E. E., 50, 744, 1931.
9. R. Tanberg, Phys. Rev., 35, 1080, 1930.
10. I. Tonks, Phys. Rev., 46, 278, 1934.
11. R. C. Mason and W. E. Berkey, Phys. Rev., 38, 943, 1931.
12. R. Risch and F. Ludi, Z. Physik, 75, 812, 1932.
13. J. Slepian, J. Frankl. Inst., 214, 413, 1932.
14. K. T. Compton, Phys. Rev., 21, 286, 1923.
15. E. Marx, E. T. Z., 53, 737, 1932.

1. Journal of Applied Physics, 1937. Translation by A. L. Komarova. ↩