STUDY OF THE ELECTRIC STRENGTH OF A CABLE WITH POLYETHYLENE INSULATION UNDER IMPULSE VOLTAGE

G. E. KLEIN, V. S. KOROLEV, N. M. TORBIN

(Presented by the scientific seminar of the Scientific Research Institute of High Voltages).

Recently, considerable attention has been devoted to the study of polyethylene insulation both in the domestic and in the foreign literature [1—5]. However, at the present time the question of the regularities of electrical breakdown of polyethylene insulation under impulse voltages has still not been sufficiently developed and, in particular, there are very few works on experimental studies of the long-term permissible impulse voltage of cables with polyethylene insulation.

In works [3—5], investigations are carried out of the dependence of the electric strength of polyethylene insulation under impulse voltages on the insulation thickness, the conductor diameter, and temperature conditions.

In [3—4] there are data on the investigation of the electric strength of polyethylene cables under impulse voltages; however, the measurement of \(E_{\text{pr}}\) is carried out by a stepwise method, i.e., when the voltage, beginning from the initial value corresponding to the maximum operating electric-field intensity, was increased with each impulse by a certain amount [3].

This method cannot be used to determine the electric strength of insulation as a function of the number of impulses.

Delektorskii [1—2] proposed an analytical method for expressing the life curve of a cable with polyethylene insulation as a function of time for alternating and impulse oscillatory voltages. This method and the regularities obtained can be used for cables employing semiconducting shields and are suitable only for certain cable designs. We carried out an experimental study of the electric strength of the insulation of an IK-2 cable under impulse voltages of various amplitudes as a function of the number of impulses to breakdown, from the maximum value \(U_{\text{pr}}\) to exposure to \(10^6\) impulses, with stepwise reduction of the amplitude of the voltage impulses. The life curve \(U_{\text{pr}} = f(n)\) under impulse voltage gives the most accurate results on the permissible operating voltage of the cable insulation, but requires a large expenditure of time and material.

To carry out the work, we designed and built a 150 kV impulse-voltage generator (Fig. 1), using capacitors of type PKGI-50/0.01, with an impulse repetition frequency of up to 20 cycles per second. The generator parameters were: wave length \(\tau_{\text{v}} = 10\ \mu\text{s}\); minimum impulse-front duration \(\tau_{\text{f}} = 0.35\ \mu\text{s}\); the impulse shape was close to rectangular. The generator capacitance in the discharge was \(C_{\text{ud}} = 6660\ \text{pF}\).

To trigger the GIN, a rotating spark gap was used, consisting of an insulating disk with a double-sided sector of aluminum pressed into it. To rotate the spark gap, an electric motor with an adjustable speed was installed.

Fig. 1. Schematic circuit diagram of the installation.

Such a triggering circuit provides smoother regulation of the generator firing frequency and has greater operating stability than when the GIN is operated in self-triggered mode.

Fig. 2. Oscillogram of the voltage-generator impulse.

In order to reduce the dimensions of the isolating inductances, the coils were designed with steel open cores. This design makes it possible to increase the value of the inductance, which reduces the loss of impulse energy in the \(L - C\) circuits switching the spark gaps during generator operation.

The switching spark gaps are installed directly at the capacitor terminals.

Wave cutting was carried out by a trigatron using a delay line \(LC\) with lumped parameters, assembled according to a T-shaped circuit (Fig. 3).

The dependence \(U_{\mathrm{br}} = f(n)\) was recorded by applying a negative voltage wave with \(\tau_{\mathrm{f}} = 0.8\ \mu\mathrm{s}\), \(\tau_{\mathrm{fr}} = 3\ \mu\mathrm{s}\).

In studies [5] on polyethylene cables, when plotting the life curve by points, a single specimen 40–50 m long is used, which makes it possible to detect only one defective location.

To determine a larger number of defects over such a cable length, we prepared cable specimens 10.6 m long, with 3–5 specimens per point.

From both ends of the cable specimen the copper braid was removed over a length of 30 cm, and the braid was prepared for the grading electrode. Then grading electrodes were fitted onto the prepared ends of the specimen; a copper screen was pulled over their tail section and secured with a bandage. Elimination of the edge effect at the point where the specimens were prepared caused serious difficulties. In the tests we tried several designs of grading electrodes (Fig. 4), of which the best results were shown by conical electrodes (Fig. 4d).

Fig. 3. Electromagnetic delay line with lumped parameters.

Fig. 4. Grading electrodes.

The insulating bushing placed under the conical electrode must be made of a material with \(\varepsilon\) equal to \(\varepsilon\) of the polyethylene insulation of the cable.

The use of insulating bushings made of materials with \(\varepsilon = 5 \div 6\) with the same conical electrodes inevitably led to breakdown of the specimen in the cable preparation, owing to the nonuniform distribution of the electric field in materials with different \(\varepsilon\). An attempt was made to eliminate the edge effect by placing the specimen with the prepared end in a semiconducting medium (technical water).

However, the flashover voltage along the surface of the exposed cable ends proved to be of the order of \(1—1.2\ \text{kV/cm}\) under impulse voltage, which led to a very great length of specimen going into the prepared portion, taking into account the value of the maximum breakdown voltage \((U_{\max\ \text{imp}} = 420\ \text{kV})\).

Comparison of the breakdown results for specimens using conical electrodes in the prepared section and specimens without grading electrodes showed that the decrease in the maximum breakdown voltage under impulses is 15–25%, and under alternating voltage of 50 cps is 6–8%; moreover, breakdown occurred only in the prepared section of the cable. Specimens prepared in this way were coiled into reels and suspended, while the prepared ends were placed in a transformer tank.

transformer oil. The voltage was applied through a bushing insulator to the cable conductor; the shield was grounded. The moment of breakdown of the specimen was recorded with an OK-19-2M oscilloscope. The cable life curve was obtained from 7 points; for each point, 3–5 specimens were used.

Fig. 5. Diagram of the test setup.

The number of impulses \((n)\) applied to the specimens at one and the same voltage showed considerable scatter. This can probably be explained by the nonuniformity of the electrical insulation along the length of cables structurally made without semiconducting shields, to which the IK-2 cable under study also belongs [4]. Thus, for example, at \(U = 145\ \text{kV}\) one specimen withstood 384 impulses before breakdown, whereas two other specimens were punctured at approximately \(n = 4 \cdot 10^3\) impulses (Table 1).

Table 1

The maximum electric strength of the IK-2 cable, with an insulation thickness of \(5.4\ \text{mm}\) and under the action of a single impulse of negative polarity with \(\tau_{\mathrm{f}} = 0.8\ \mu\text{s}\), in our experiments was \(E_{\max\ \mathrm{br}} = 78\ \text{kV/mm}\), while the short-time electric strength at alternating voltage of 50 Hz was \(20\ \text{kV/mm}\) at a voltage rise rate of \(1\ \text{kV/s}\), which agrees well with the data of [4] for cables with the indicated insulation thickness.

It has been shown [2] that when a specimen is subjected to a very large number of impulses, the breakdown voltages under impulses and under alternating voltage coincide as a function of the number of impulses; therefore, for a number of impulses per point \(n = 10^6\) and higher, the tests of specimens were carried out by us under alternating voltage (Table 2).

Table 2

It follows from Fig. 6 that the long-term electric strength of the IK-2 cable under impulse voltage is 13 kV/mm. The long-term permissible operating voltage of the IK-2 cable under impulse voltage of negative polarity should not exceed 70 kV.

On the basis of the work carried out, the following conclusions may be drawn:

1. The edge effect at the cut of cable specimens can be eliminated by using grading conical electrodes.

2. The maximum electric strength of the IK-2 cable for a single impulse is \(E_{\max \,\mathrm{br}} = 78\) kV/mm for negative impulse polarity.

Fig. 6. Life curve of the IK-2 cable.

1. The minimum electric strength at a long-term permissible impulse voltage of negative polarity is 13 kV/mm.

2. The IK-2 cable has nonuniform electric strength along its length, since there is a large scatter in the number of impulses to breakdown from specimen to specimen at all points of the life curve.

REFERENCES

1. G. P. Delektorskii. On certain regularities of breakdown of polyethylene insulation, Elektrichestvo, No. 11, 1961.

2. G. P. Delektorskii. Regularities of breakdown of high-voltage cables with polyethylene insulation during transmission of voltage impulses, Vestnik elektropromyshlennosti, No. 1, 1963.

3. R. M. Laker and D. L. Sharpe. Polyethylene and its application in cable engineering, Gosenergoizdat, 1958.

4. G. Nuht, M. Konopoulos, P. Ware. Dielectric strength and life curve of polyethylene, Power appratus and sustems, 1958, No. 35.

5. Power engineering abroad, high-voltage cable lines, issue 2, Gosenergoizdat, 1959.