IMPULSE STRENGTH OF DIELECTRICS UNDER SIMULTANEOUS STATIC VOLTAGE EXPOSURE

E. K. ZAVADOVSKAYA

Associate Professor, Candidate of Physico-Mathematical Sciences

Introduction

Under operating conditions, insulation is under the influence of the working voltage at fairly high potential gradients and, at the same time, may be subjected to the action of traveling overvoltage waves.

There are indications in the literature that the simultaneous action on a dielectric of static and impulse voltages may facilitate breakdown [1].

In practice, until recently, in electrical-machine construction it was not necessary to take into account the possibility of a reduction in the electric strength of insulation due to the combined action of the working voltage and traveling overvoltage waves, since the insulation level in machines was determined by the corona-inception voltage; therefore it operated at low potential gradients. At present, when machines are being developed for working voltages of 20–60 kV, the insulation in such machines will be quite heavily loaded electrically, and in its calculation it will be necessary to take into account the combined action of voltages mentioned above.

In the present work the question of the impulse strength of certain dielectrics is examined under the simultaneous action upon them of static voltage. Under such conditions of simultaneous action of two types of voltage there may be line, machine, and apparatus insulation.

In our experiments the following dielectrics were investigated: air, mica, textolite, presspan, rock salt, and transformer oil. Air was chosen as the dielectric under investigation, first, because it is used as insulation in electrical machines, apparatus, and electric-transmission lines; second, under our conditions of testing dielectrics, under comparatively complex conditions, air served as a kind of indicator.

Experimental Method

The investigation of the behavior of insulation under the simultaneous action upon it of two types of voltage was carried out under conditions of a uniform and a nonuniform field.

Five different experimental installations were assembled, representing various combinations of direct-, alternating-, and impulse-voltage installations.

Of practical interest is the simultaneous action on insulation of alternating and impulse voltage. In view of the short duration of the impulse voltages used by us (impulse duration \(10^{-5}\)—\(10^{-4}\) sec), the alternating voltage, in relation to the impulse, rapidly varying voltage, may be regarded as static. Bearing in mind that the results of insulation tests under direct and alternating voltage are identical, in order to avoid the uncertainty, under simultaneous action on the specimen of alternating and impulse voltages, associated with the phase angle, we carried out tests under simultaneous action of direct and impulse voltages.

To obtain the simultaneous action of direct and impulse voltage on a dielectric, an installation was assembled according to circuit I, shown in Fig. 1. This installation consists of an impulse generator; at the same time a direct voltage can be applied to the specimen from the generator power supply. The generator power supply was a rectifier installation assembled according to a voltage-doubler circuit. In Fig. 1 the following are denoted: \(T_2\)—transformer, \(K\)—kenotrons, \(C_1 = 0.02 \mu F\)—paper-oil capacitors. Connected to the capacitances \(C_1\), in addition to the charging resistances \(R_1\) (\(R_1 = 10000 \Omega\)), are two resistances, designated in the figure as \(R_3\)

Fig. 1. Diagram of the combined electrical installation for obtaining direct and impulse voltage.

(\(R_3 = 1000 \Omega\)), through which the direct voltage is applied to the object under investigation. The place where the specimen to be punctured is connected is designated in the figure by the letter \(D\). The impulse-generator circuit is separated from the direct-voltage circuit by isolating spheres \(F_2\), which broke down during the impulses. At this moment the impulse voltage was applied to the specimen under investigation. Such a circuit made it possible to carry out a study of insulation under the simultaneous action upon it of direct and impulse voltage, both of coinciding and of different polarities.

Fig. 2. Diagram of the combined electrical installation for obtaining direct and impulse voltages with an independent source of direct voltage.

For carrying out certain investigations undertaken with the aim of determining the impulse strength of insulation under the action upon it of direct voltage of different magnitude, an installation was assembled according to circuit III, shown in Fig. 2.

The combined installation, assembled according to circuit III, consists of two main parts: an impulse generator and a direct-voltage installation, which are separated from one another by isolating spheres \(F_2\). With the aid of the installation assembled according to circuit III, it is possible to model the operation of insulation under service conditions, when the insulation (discharger \(D\) in Fig. 2), on one side, is subjected to the prolonged action of static voltage (applied by the right-hand part of the circuit), and on the other, to the action of short-duration impulses (generated by the left-hand part of the circuit).

In addition, some experiments were performed on the breakdown of air gaps in a uniform and a nonuniform field under the simultaneous action of direct and alternating voltage. For these experiments a combined electrical installation was assembled according to circuit V, shown in Fig. 3, which made it possible simultaneously to apply direct and alternating voltage to the arrester. Fig. 3 gives two operating cases for such an installation. The upper circuit 1 represents the case in which the point is positive; when connected according to circuit B the point will be negative.

Fig. 3. Circuit of the combined electrical installation for obtaining direct and alternating voltage.

To measure the current flowing in the given direction, two milliammeters—mA₁ and mA₂—were included in the alternating-voltage installation, with kenotrons K₁ and K₂ connected in series with them; the latter were connected in opposition.

Breakdown of air under impulses with simultaneous superposition of direct voltage

In our experiments the following were studied: first, the impulse strength of various air gaps under the simultaneous action of a direct voltage whose magnitude was close to that of the impulse voltage; and second, the impulse strength of a given constant air gap under the simultaneous application of direct voltage of various magnitudes. The results of the breakdown of various spark gaps in air under impulse voltage with the simultaneous action of a direct voltage close in magnitude to the impulse voltage are given in Fig. 4, where the ordinate gives the impulse breakdown voltage in kilovolts, and the abscissa gives the length of the air gap in mm. Curve I represents the dependence of the magnitude of the impulse breakdown voltage on the distance between the measuring spheres. From this curve it is seen that, under the given experimental conditions, the electric strength of air under impulses was \(E_{\mathrm{br}} = 32\ \mathrm{kV/cm}\).

Fig. 4. Dependence of the impulse breakdown voltage of air on the magnitude of the simultaneously acting direct voltage. Uniform field.

Curve II represents the dependence of the impulse breakdown voltage on the distance between the spheres in the case of superposition of a direct voltage of polarity opposite to the polarity of the impulse voltage, and curve III shows the dependence of the impulse breakdown vol-

voltages on the distance between the measuring spheres when the polarities of the direct and impulse voltages coincide. It follows from Fig. 4 that the impulse breakdown voltage, when a direct voltage is simultaneously superimposed, decreases; moreover, the greater decrease is observed when a voltage of different polarities is applied to the measuring air gap. In particular, it follows from Fig. 4 that under impulses an air gap of 10 mm breaks down at 32 kV, while with the simultaneous application of a direct voltage of opposite polarity the impulse breakdown voltage is reduced by 11 kV. The superposition of a direct voltage of the same polarity causes a reduction of the breakdown voltage by 4 kV. Thus, as a result of the experiments carried out, it may be concluded that the direct voltage has an influence on the impulse strength of air. According to our data, the polarity of the direct voltage simultaneously applied to the air gap plays a role in this case.

The obtained dependence of the impulse discharge voltage on the polarity of the applied direct voltage is presented in Fig. 5, which gives the dependence of the impulse breakdown voltage on the magnitude of the simultaneously applied direct voltage for an air gap 15 mm long in the case of a uniform field. Along the ordinate axis are plotted the values of the impulse breakdown voltage in kilovolts, and along the abscissa axis—the magnitude of the direct voltage in kilovolts. The curves designated by the numeral I show the change in the impulse breakdown voltage for the given gap as a function of the magnitude of the direct voltage when the polarities coincide. It follows from Fig. 5 that, in the case of different polarities, with increasing direct voltage a decrease in the impulse breakdown voltage is observed. When the polarities of the direct and impulse voltages coincide, an increase in the impulse breakdown voltage is observed,

Fig. 5. Dependence of the impulse breakdown voltage of air on the magnitude of the simultaneously acting direct voltage. Uniform field. Test according to circuit III.

I — impulse and direct voltage of different polarities.
II — impulse and direct voltage of the same polarities.
* — breakdown with capacitance \(C = 100\ \text{cm}\).
• — without capacitance \(C\).

and the increase in the impulse breakdown voltage is given.

The results of experiments investigating the influence of the magnitude and polarity of the direct voltage on the magnitude of the impulse breakdown voltage in the case of a nonuniform field have the same form as in the case of a uniform field. In a nonuniform field this influence is expressed more strongly.

Thus, under different experimental conditions a decrease in the impulse breakdown voltage of air is always observed when the direct and impulse voltages have different polarities, and an increase—in the case of coinciding polarities of both voltages. Moreover, as the direct voltage increases, both the increase and the decrease of the impulse breakdown voltage increase; and at a direct-voltage magnitude of 35 kV the maximum difference between the impulse breakdown voltage with superposition of a direct voltage in a nonuniform field, in the case of coinciding and different polarities, reaches 45 kV.

The results of our experiments show that, under the simultaneous action on a spark gap of static and impulse voltages, im-

the impulse strength changes more strongly in the case of a nonuniform field. This circumstance is very significant, since in practice insulation operates under conditions of a nonuniform field. We were not able to establish a quantitative law, and this is hardly possible, since the observed effect must depend strongly on the parameters of the circuit on which the impulse is incident, as we observed in our experiments. We have established the general regularity of the phenomenon. It consists in the fact that, with an increase in the additionally acting dc voltage, the rate of change of the impulse voltage increases. Thus, in moving to installations of higher voltage, the danger to air insulation from the action of combined static and impulse voltages increases.

The observed phenomenon—the change in the impulse strength of insulation under the combined action of dc and impulse voltages—is explained by the addition and subtraction of the voltages supplied by the two sources.

When a spark gap is acted upon by impulse and dc voltages of different polarities, the static voltage and the impulse voltage are added, which may be represented by the diagram shown in Fig. 6, where voltage is plotted along the ordinate axis and time along the abscissa axis. As can be seen, in this case, neglecting losses, the insulation under test is subjected to a voltage equal to the sum of the impulse and static voltages. As a result of such superposition, the impulse breakdown voltage of the gap will be reduced; the degree of reduction will depend on the magnitude of the acting static voltage.

Fig. 6. Diagram of the addition of voltages of different polarities.

We may have such a case when a surge wave falls on an energized transformer. The probability that the action of the surge wave will coincide with the corresponding maximum of the operating voltage is considerable, and under these conditions the air insulation is subjected to a severe test. This circumstance must be taken into account when selecting the insulation level of a transformer. An example of such a breakdown of air insulation under the combined action of the operating voltage and an impulse surge wave may also be found in the operating practice of power-transmission lines. For example, V. N. Zaleskii [1], in an abstract of a paper by Sporn and Gross to the American Institute of Electrical Engineers, gives the following data on lightning damage to power-transmission lines: “Oscillographic measurements on certain lines in 1934–1937 showed that, of 300 recorded faults, in 178 cases (59.3%) the polarity of the faulted conductor was positive, in 102 cases (34%) it was negative, and in the remaining 20 cases (6.7%) the voltage was equal to zero.”

Investigations by many authors [2] have established the predominance of negative lightning discharges, constituting 90% of all discharges.

The data cited above, showing that lightning predominantly strikes the positive conductor, are in good agreement with our results, which indicate that, under the combined action of impulse and static voltage on a spark gap, it breaks down more readily when these voltages have unlike polarities.

Our results were obtained for spark gaps of small magnitude (up to 30 mm). The data of Sporn and Gross evidently refer

because these results can also be extended to spark gaps of the order of several meters.

The decrease in the impulse strength of a spark gap under the simultaneous action of a static voltage of opposite sign, and the increase in strength under the action of a voltage of the same sign, suggested the use of this effect for the purpose of rectifying alternating current.

It should be expected that, when direct and alternating voltages are superimposed on an air gap in one half-period, when the polarities coincide, conditions will be created for strengthening the air gap and breakdown will be hindered. In the other half-period, when the polarities are different, conditions will be created for weakening the air gap and breakdown will occur.

On the basis of these assumptions, a series of experiments was carried out on an installation specially assembled for this purpose, allowing direct and alternating voltages to be applied simultaneously to the air gap under investigation; the circuit of this installation is shown in Fig. 3. To extinguish the arc ignited during breakdown between the electrodes, artificial blowing of the gap with a jet of air under pressure was used.

The rectifying property of such an installation was observed under conditions of a uniform and a nonuniform field.

For certain combinations of the voltage applied to the electrodes and the strength of the blowing, it was possible to obtain complete rectification. In these cases, in the ignition direction the milliammeter indicated a current equal to zero. It seems to us that the briefly described above valve property of air gaps, revealed under the simultaneous action on them of two voltages—direct and alternating—may find practical application for the purposes of rectifying alternating current, and also in control and signaling circuits.

Breakdown of Solid and Liquid Dielectrics by Impulses under Simultaneous Superposition of Direct Voltage

The investigation of the impulse breakdown voltage of rock salt under the simultaneous superposition of direct voltage was carried out by us under conditions of a uniform field on the installations shown in Figs. 1 and 2.

The results of determinations of the impulse breakdown voltage of rock salt under the simultaneous superposition of direct voltage, made on the installation assembled according to circuit 1, are presented in Fig. 7. On the ordinate axis are plotted the values of the impulse breakdown voltages of rock salt; on the abscissa axis, the thickness of the samples being broken down, in mm. Curve I represents the dependence of the impulse breakdown voltage on the thickness of the samples; curve II, the dependence of the impulse breakdown voltage on the thickness of the samples when a direct voltage is superimposed in the case of different polarities of the direct and impulse voltages; curve III, the dependence of the impulse breakdown voltage on the thickness of the samples in the case of coincident polarities; and curve IV, the dependence of the breakdown voltage on the thickness of the samples at direct voltage.

The results obtained are basically the same as in the breakdown of air. The impulse breakdown voltage of rock salt, under the simultaneous action of a direct voltage of opposite polarity, decreases. This result coincides with what Grünwald [3] observed for Madras mica.

When a direct voltage of the same polarity is superimposed, the impulse breakdown voltage of the dielectric increases. The same result is observed, for example, in testing rock salt on an apparatus assembled according to circuit III. The result of these tests is presented in Fig. 8.

Breakdown of rock salt under the simultaneous superposition of a direct voltage, close in magnitude to the impulse voltage, showed that applying a direct voltage of polarity opposite to that of the impulse leads to a lowering of the breakdown voltage of rock salt.

Breakdown of mica in a nonuniform field under impulses with simultaneous action of a direct voltage showed that the breakdown voltage of mica, under impulses with simultaneous action of a direct voltage of opposite polarity, is noticeably reduced.

Fig. 7. Dependence of the impulse breakdown voltage of rock salt on the magnitude of the simultaneously acting direct voltage. Uniform field. Test according to circuit I.

The dielectrics considered above—air, rock salt, and mica—possess an electrical form of breakdown. The porcelain, ebonite, and glass tested by us, owing to the poor quality of the materials available to us, did not in thin layers give such clear results as mica and the other dielectrics; despite a large scatter of points, the general character of the dependence obtained for porcelain, ebonite, and glass was the same as for air, mica, and rock salt.

Fig. 8. Dependence of the impulse breakdown voltage of rock salt on the magnitude of the simultaneously acting direct voltage. Uniform field. Test according to circuit III.

Then dielectrics were investigated which at room temperature have a thermal form of breakdown. This class of dielectrics includes pressboard, getinax, textolite, and transformer oil. Measurements by V. A. Vorob’ev [4] showed that for them there is observed the linear dependence of the logarithm of the breakdown voltage on temperature which is characteristic of the thermal form of breakdown.

In our experiments, breakdown of pressboard was produced between point-to-plane electrodes with the electrodes immersed in oil. The pressboard specimens, plates measuring \(30 \times 30 \times 0.5\) mm, were previously boiled for a long time (6–8 hours) in transformer oil at a temperature of \(90^\circ\text{C}\). As a result of boiling in oil, the specimens received a uniform impregnation and became fairly homogeneous. Breakdown of such specimens gave a small scatter of points. Before breakdown the specimens were kept in a vessel with oil. In our experiments two grades of pressboard were used, differing in thickness (0.5 and 0.8 mm). For both grades of press—

the porcelain tests yielded identical results. In practice it may be considered that the value of the breakdown voltage of porcelain under impulses, with simultaneous exposure to a static voltage of a polarity coinciding with and opposite to the polarity of the impulses, changes hardly at all. One can note only a certain tendency toward the formation of a fork of the type that we had for air and rock salt.

Fig. 9. Dependence of the breakdown voltage of textolite on the thickness of the specimens under impulses and constant voltage. Uniform field.

Other results were obtained in the breakdown of textolite and transformer oil. The breakdown of textolite was carried out under conditions of a uniform field.

The investigations carried out showed that the application of a constant voltage during testing according to scheme I does not change the impulse breakdown voltage of textolite. The breakdown voltage on direct current for textolite lies considerably below the breakdown voltage under impulses (Fig. 9). Hence it follows that, over the entire range of thicknesses investigated, the impulse strength of textolite is approximately three times greater than the strength under constant voltage.

In testing according to scheme I, when the constant voltage increases together with the increase in the impulse voltage, breakdown naturally occurs at the constant voltage before the impulse voltage rises to the destructive value. The same results were obtained in testing transformer oil by means of an apparatus assembled according to scheme I. Breakdown of the transformer oil was carried out between standard electrodes with distances between them equal to 1 and 2 mm. Investigation of the impulse strength of transformer oil with simultaneous application of a constant voltage, close in magnitude to the impulse voltage, showed the absence of any influence of the constant voltage. Investigation of breakdown under these conditions led to the conclusion that breakdown, as also for textolite, occurred on account of the acting constant voltage.

Conclusion

An experimental investigation of the impulse strength of a number of insulating materials shows that, under simultaneous action of impulse and static voltages of different polarities, a decrease in their impulse strength is observed, and an increase in the case of coinciding polarities. The observed phenomenon—the change in the impulse strength of insulation under the combined action of static and impulse voltages—is the result of the addition or subtraction of these voltages.

In the case when the impulse breakdown voltage of a dielectric considerably exceeds the static voltage, which occurs in the thermal form of breakdown, the effect of the combined action of the voltages is absent. The impulse strength of dielectrics possessing an electrical form of breakdown, for example, air, mica, and others, with simultaneous application of a static voltage of opposite polarity decreases, in the worst case, by the value of the applied constant voltage.

The observed change in the breakdown voltage of an air spark gap under the simultaneous action of two voltages upon it was used by us for the purpose of rectifying alternating current.

The effect of lowering the breakdown voltage of insulation under the simultaneous action of two voltages upon it must be taken into account when selecting the insulation level for apparatus and machines, and also when testing it.

The observed phenomenon can be applied to explain disturbances of the electrical strength of insulation when overvoltage waves fall upon electrical installations that are energized, and it can be used for signaling, control and protection purposes, as well as for rectifying alternating current.

LITERATURE

1. V. N. Zalesskii. Elektricheskie stantsii, 4–5, 93, 1939.
2. I. S. Stekol’nikov. Molniia. Publishing House of the Academy of Sciences of the USSR, 1940.
3. F. Grunwald. Arch. f. El., 12, 1, 1923.
4. V. A. Vorob’ev. Elektrichestvo, 53, 1940.