EFFECT OF BARRIERS ON THE DIELECTRIC STRENGTH OF OIL-IMPREGNATED PAPER INSULATION

D. D. RUMYANTSEV

Presented by the scientific seminar of the Department of High-Voltage Engineering

To increase the dielectric strength of insulating structures, barriers are widely used. The presence of barriers in a discharge gap in liquid dielectrics and in air leads to a significant increase in breakdown voltages [1—3]. It has been established that the barrier effect is also observed in solid dielectrics [4—6].

It has been shown [5] that barriers made of getinax and textolite in pure paraffin and in cable compound MK-45 raise the breakdown voltage to 160—170%. Barriers made of polyethylene and cellulose triacetate films in celluloid and rock salt [6] increase the breakdown voltage of specimens by 10—70% relative to the breakdown voltage without a barrier under the action of direct, alternating, and impulse voltages.

The purpose of the present work is to establish the existence of a barrier effect in a complex dielectric such as oil-impregnated paper insulation, and to obtain qualitative dependences of the change in \(U_{\text{br}}\) of oil-impregnated paper insulation in the presence of thin high-strength barriers.

PART I

The influence of barriers on the short-time breakdown voltage of oil-impregnated paper insulation was investigated in a sharply nonuniform field of “point–plane” electrodes.

The investigations were carried out with impulse voltage at positive and negative impulse polarity and with alternating voltage of industrial frequency. The source of impulse voltage was a GIN rated at 1000 kV, with a discharge capacitance of 0.022 μF. The impulse front duration was varied from \(0.5 \cdot 10^{-6}\) sec to \(10^{-5}\) sec. Breakdown always occurred on the impulse front. The voltage on the specimen was measured with an OK-19M oscillograph.

The source of alternating voltage was a 150 kV test transformer. The voltage was raised smoothly up to breakdown. The rate of voltage rise was 5 kV/sec.

The specimens were assembled from sheets of cable paper 0.12 mm thick. The paper was impregnated with pure transformer oil without vacuum treatment or drying. Sheets measuring \(20 \times 20\) cm were placed one on another, forming a “package.” The package consisted of 20 sheets in the experiments with impulse voltage, and of 15 sheets with alternating voltage. The specimen was clamped in an insulating frame, which ensured the same thickness of the oil interlayers. To reduce the probability of air inclusions between the sheets, the operation of assembling the specimen was carried out under oil.

The barrier was placed between the sheets. Films of polyethylene, polystyrene, and fluoroplastic-4 were used as barriers. The film thickness was 25–30 \(\mu\).

For each point, 15–25 breakdowns were made. Specimens that were punctured through or partially punctured through were rejected. The increase in breakdown voltage was calculated by the formula:

\[ \Delta U = \frac{U_b - U_0}{U_0}, \]

where

\[ U_b \]

— breakdown voltage of the specimen with a barrier,

\[ U_0 \]

— breakdown voltage of the specimen without a barrier, calculated as the arithmetic mean of a series of measurements.

Some characteristic dependences obtained in the experiments are given below.

Figure 1 shows the dependence of the increase in \(U_{\mathrm{br}}\) of a package of cable paper on the position of a fluoroplastic-4 barrier. As follows from the figure, the greatest increase in \(U_{\mathrm{br}}\) is observed when

Fig. 1. Effect of barrier position on the breakdown voltage of a package of cable paper. Positive-polarity impulses.
1 — \(\tau_{\mathrm{fr}} = 0.5 \cdot 10^{-6}\) sec, 2 — \(\tau_{\mathrm{fr}} = 10^{-5}\) sec,
3 — alternating voltage, 50 cps.

when the barrier is placed at \(0.2\)—\(0.4\) of the specimen thickness from the point under impulse voltage; under alternating voltage, close to the middle of the specimen. The influence of the barrier is more substantial for impulses of longer duration. A dependence analogous to curves 1 and 2 (Fig. 1), obtained with an impulse front duration of \(3 \cdot 10^{-6}\) sec, occupies an intermediate position between curves 1 and 2.

Investigation of the influence of the barrier on \(U_{\mathrm{br}}\) for positive and negative impulses made it possible to establish that the impulse polarity, at times of \(10^{-5}\) sec and less, is not of substantial significance. However, when the barrier is located directly adjacent to the point, a small strengthening is observed for positive impulse polarity and is not observed for negative polarity.

The change in \(U_{\mathrm{br}}\) of the stack as a function of the position of a barrier made of various materials is presented in Fig. 2. The maximum increase in \(U_{\mathrm{br}}\) (by 16%) is observed when a polystyrene barrier (3) is placed at \(0.1\) of the specimen thickness. Barriers of fluoroplastic (2) and polyethylene (1) give increases in \(U_{\mathrm{br}}\) of 12% and 10%, respectively.

Fig. 2. Influence of the barrier position on the breakdown voltage of a stack of cable paper. Positive-polarity impulses, \(\tau_{\mathrm{fr}} = 10^{-5}\) sec.
Barrier material: 1 — polyethylene, 2 — fluoroplastic-4, 3 — polystyrene.

Table 1 gives data on the dielectric strength of barriers in the field of “point–plane” electrodes.

Since the strength of the barrier itself is small in comparison with the strength of the stack (Table 1), and the strength of a specimen with a barrier is considerably higher than the combined strength of the specimen and the barrier taken separately, one may speak of a “barrier effect” in paper-oil insulation. On the basis of the results obtained (Fig. 2, Table 1), it may be concluded that the barrier effect in paper-oil insulation is higher when barriers with a larger \(U_{\mathrm{br}}\) are present.

Table 1

It was of interest to investigate the effect of two barriers on \(U_{\mathrm{br}}\) of a pack of cable paper. The results of the experiments are given in Table 2.

Table 2

As can be seen from Table 2, the maximum increase in \(U_{\mathrm{br}}\) (up to 130%) is observed when the barriers are located at 0.1 and 0.35 of the specimen thickness. Other combinations of barrier locations give a smaller increase in \(U_{\mathrm{br}}\).

PART II.

The experimental results presented in Part 1 show that barriers made of thin high-strength films significantly increase the short-time breakdown voltage of paper-oil insulation.

To determine the possibility of practical application of this phenomenon, experiments were carried out to study the effect of barriers made of polyethylene film on the strength of paper-oil insulation in a nonuniform field of “disk–plane” electrodes. The experiments were carried out at alternating voltage of industrial frequency. The high-voltage source was a 150 kV test transformer.

The high-voltage electrode was a brass disk with sharp edges, 40 mm in diameter. The grounded electrode was aluminum foil cut in the form of a circle 120 mm in diameter. In studying the effect of barriers on the electric strength of paper-oil insulation in a uniform field, the electrodes used were a plane with rounded edges 40 mm in diameter and a sheet of foil of a size exceeding the size of the specimen.

The specimens were prepared by the method described above. The number of sheets of paper in a specimen was 12. Some of the specimens were prepared from cable paper dried and impregnated under vacuum. The specimens were dried and impregnated at a residual pressure of 1 mm Hg. The drying temperature of the specimens was 100–110°C, and the impregnation temperature was 55–65°C [7].

A polyethylene film 40–50 μ thick was used as the barrier. Polyethylene was chosen by us because it has greater thermal stability than polystyrene, which is important during drying and impregnation. Compared with fluoroplast-4, polyethylene is considerably cheaper, less scarce, and can be manufactured in the form of rolls of film of practically unlimited width.

The barrier was placed between the paper sheets, four sheets away from the high-voltage electrode. In experiments with two barriers, the first was placed two sheets away from the high-voltage electrode, the second four sheets away. In experiments with 3 and 4 barriers, they were placed after 2, 4, 8 and after 2, 4, 6, 9 sheets, respectively.

The voltage applied to the specimen was raised at a rate of 6.5–7 kV/sec to 0.75 of the presumed \(U_{\mathrm{br}}\), and then to breakdown at a rate of 1.5–2 kV/sec, which amounted to 2–2.5% \(U_{\mathrm{br}}\) per second.

In studying the influence of stable ionization on the barrier effect in paper-oil insulation, voltage was applied simultaneously to 6 specimens with a barrier and without a barrier. The value of the field strength at which stable ionization arises was calculated from the formula given in [8] and was equal to 14.6 kV/mm.

Fig. 3. Integral distribution functions of specimens by \(U_{\mathrm{br}}\).
1 — without a barrier; 2 — with one barrier; 3 — with two barriers.

In [8] it is indicated that the inception field strength for sustained ionization in paper-oil insulation is considerably lower if impregnation was carried out at a residual pressure of 200 mm Hg and higher. In the present experiments the specimens were impregnated without vacuum and, consequently, at a field strength of 14.6 kV/mm sustained ionization was guaranteed.

After 120 minutes or 480 minutes had elapsed, the voltage was switched off, and then the short-time \(U_{\text{br}}\) was determined by the usual method.

To measure the time to breakdown \((\tau)\), the voltage was rapidly raised to \(0.6\,U_{\text{br}}\) of specimens without a barrier and was kept unchanged until breakdown.

When determining the integral distribution functions of the specimens with respect to \(U_{\text{br}}\) and \(\tau\), 30–60 specimens were broken down. In the remaining experiments, 12–15 specimens were broken down for each point. The error in measuring the voltage was 2%.

The experimental results are given below.

Figure 3 presents the integral distribution functions of specimens without a barrier (1), with one (2), and with two (3) barriers, according to the magnitude of the short-time \(U_{\text{br}}\). As is seen from Fig. 3, at a 50% probability of breakdown, one barrier increases \(U_{\text{br}}\) by 9%, and two by 17% relative to \(U_{\text{br}}\) without a barrier.

Fig. 4. Integral distribution functions of specimens with respect to \(\tau\).
1 — without a barrier, 2 — with one barrier, 3 — with two barriers.

Figure 4 shows the integral distribution functions of the specimens with respect to time to breakdown \((\tau)\) at a voltage equal to \(0.6\,U_{\mathrm{br}}\). It follows from the figure that specimens with one barrier (2) have \(\tau\) approximately 2 times greater, and those with two barriers (3) 2.5 times greater, than specimens without a barrier (1). It should be noted that, in the presence of barriers, the scatter of the values of \(\tau\) is considerably greater than that of \(U_{\mathrm{br}}\).

Figure 5 shows the dependences of \(U_{\mathrm{br}}\) on the number of barriers at an unchanged number of sheets (1) and on the number of sheets (2) in the specimen. As can be seen from Fig. 5, two barriers increase \(U_{\mathrm{br}}\) by 17%, whereas when the number of paper sheets in the specimen is increased from 12 to 15, \(U_{\mathrm{br}}\) increases by 13%. Obviously, in order to increase \(U_{\mathrm{br}}\) while limiting the insulation thickness, the use of barriers is expedient.

Table 3 gives data from the investigation of the barrier effect in paper-oil insulation under various conditions.

Fig. 5. Influence of the number of barriers and the number of sheets on \(U_{\mathrm{br}}\) of a stack of cable paper.
1—effect of the number of barriers (12 sheets in the specimen),
2—effect of the number of sheets (specimens without a barrier).

Table 3

As follows from Table 3, an increase in \(U_{\mathrm{br}}\) due to the barrier is observed: in a uniform and a nonuniform field; in the presence of stable ionization for 480 minutes; and in samples impregnated under vacuum.

It is known that polymer films have low corona resistance. However, in [9] it is shown that the short-time \(U_{\mathrm{br}}\) remains unchanged up to 0.75 h. Consequently, barriers made of polymer films should increase \(U_{\mathrm{br}}\) in the case of a short-time existence of ionization in paper-oil insulation. This is confirmed by our experiments.

Since a prolonged existence of ionization is inadmissible in paper-oil insulation, one may speak of the practical use of film barriers in paper-oil insulation. The possibility of practical application of barriers in paper-oil insulation is also indicated by the fact that barriers increase \(U_{\mathrm{br}}\) in uniform and nonuniform fields, in insulation impregnated under vacuum.

CONCLUSIONS

1. In paper-oil insulation, in the presence of barriers made of high-polymer films (polyethylene, fluoroplastic-4, polystyrene), an increase in \(U_{\mathrm{br}}\) by 10–20% is observed under impulse and alternating voltage.

2. The barrier provides the maximum strengthening when it is placed at 0.1–0.4 of the sample thickness from the high-voltage electrode.

3. An increase in \(U_{\mathrm{br}}\) in the presence of a barrier made of polyethylene film is observed in a uniform and a nonuniform field.

4. Barriers made of polyethylene film increase \(U_{\mathrm{br}}\) in the presence of stable ionization in paper-oil insulation.

5. Increasing \(U_{\mathrm{br}}\) of paper-oil insulation by means of film barriers may find application in the design of high-voltage insulation.

LITERATURE

1. Voloshchenko P. F. “Elektrichestvo,” 3, 24, 1946.
2. Komelkov V. S., Lifshits A. M. “Izv. AN SSSR,” 10, 39, 1951.
3. Shilevan A. A. Tr. LPI, 1, 300, 1954.
4. Leont’ev Yu. N., Torbin L. M. “Izv. vuzov ‘Energetika’,” 12, 34, 1961.
5. Mitov B. Godishnik mash.-elektrotekhn. in-t, 6, 2, 27, 1959 (Bolg.).

1. Rumyantsev D. D., Torbin N. M. “Breakdown of dielectrics and semiconductors.” Collection of reports of the IV Interuniversity Conference, 170, 1964.

2. Freisukh M. A., Kuchinskii G. S., et al. Paper-oil insulation in high-voltage structures, GEI, 1963.

3. Kuchinskii G. S., Kaplan D. A., Tikhonova O. V. Izv. vuzov “Energetika”, 8, 30, 1959.

4. Koikov S. P., Platonov M. P., Tsiin A. N. Proceedings of the conference “Ionization aging, corona resistance,” 27, TsINTIEP, 1963.