IMPULSE ELECTRIC STRENGTH AND SELECTIVITY OF THE DISCHARGE PATH IN AN INHOMOGENEOUS SOLID DIELECTRIC

V. I. Kurets

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

The electric strength of solid insulation is determined to a considerable extent by the homogeneity of the dielectric structure. The presence in solid insulation of foreign inclusions having specific resistivities and dielectric permittivities different from these same characteristics of the insulation itself leads to a change in the field intensities on the surface of the inclusion or inside it, which must affect the magnitude of the average breakdown gradients and the direction of the discharge path in the solid dielectric.

Inhomogeneities (pores, inclusions) in solid insulation, in high-voltage apparatus and devices, are often encountered as a result of imperfections in the technology and manufacture of the insulating material, and also during operation when ionization processes develop inside the dielectric. Thus, investigations of the influence of a metallic inclusion on the electric strength and on the path of an electric discharge in a solid dielectric are of considerable practical interest.

In connection with the above, in the present work investigations were carried out to evaluate the influence of conducting inclusions in a solid dielectric on the average breakdown gradients and on the direction of the discharge path in it.

Experimental procedure. As the source of high voltage, an impulse-voltage generator (IVG) for 240 kV with a discharge capacitance of 1700 pF was used. The front duration of the impulse-voltage wave was \(0.2 \cdot 10^{-6}\) sec. The electrical circuit of the impulse generator and the voltage-measurement circuit are shown in Fig. 1. The impulse voltage was measured by means of an OK-19M electronic oscillograph and an ohmic divider with a resistance of 6000 ohms. In the investigations, paraffin was used as the solid dielectric; the conducting inclusion in the dielectric was simulated by shot embedded in the specimen. In preparing the specimens, a special temperature regime made it possible to obtain paraffin specimens without large air pores and with an electric strength of the order of 15 kV/mm. Specimens measuring \(40 \times 30 \times 25\) mm were used in the investigation. In the breakdown experiments, in order to prevent flashover of the specimens along the surface, the specimens were immersed in transformer oil with a strength of 30 kV in a standard breakdown tester. The electrodes—

served as electrodes: a brass plane 100 mm in diameter and a pointed rod with an apex angle of \(30^\circ\). The plane was grounded, and voltage pulses of positive polarity were applied to the rod electrode.

Fig. 1. Electrical circuit of the pulse generator.

In studying the influence of a metallic inclusion on the reduction of the breakdown gradients of the dielectric, the point of the rod was set along the axis of the metallic inclusion; in studying changes in the discharge path, the point was displaced from the vertical axis of the metallic inclusion.

Breakdown of the specimens was carried out on the front of a single voltage pulse. Each point in the experimental dependences presented below was constructed from the results of breakdown of 4–6 specimens. The scatter in the values of the breakdown voltages did not exceed \(\pm 20\%\) of the mean value.

Fig. 2. Dependence of the average breakdown gradients on specimen thickness.
\(1\) — specimens do not contain a metallic inclusion; \(2\) — specimens contain a metallic inclusion.

Experimental results. Figure 2 presents the dependence of the average breakdown gradients on the dielectric thickness for specimens containing \(2\) and not containing \(1\) a metallic inclu-

... The calculation of the average breakdown gradients for specimens containing a metallic inclusion was carried out according to the formula

\[ E_{\mathrm{av.\,br}}=\frac{U}{S-d}, \]

where \(U\) is the breakdown voltage of the specimen,
\(S\) is the thickness of the specimen,
\(d\) is the diameter of the metallic inclusion.

Thus, the size of the metallic inclusions was not taken into account in calculating the average breakdown gradients. As can be seen from the graph in Fig. 2, in both cases the average breakdown gradients decrease with increasing thickness; however, their values for specimens without a metallic inclusion are 25–30% higher than in specimens containing an inclusion. Thus, the presence of a local region of increased electric-field strength on the surface of the conducting inclusion led to a relative decrease in the magnitude of the impulse breakdown voltage.

In this work, studies were carried out of the change in the average breakdown gradients in a specimen with a conducting inclusion for various relative increases in the discharge path; the selective direction of the impulse discharge in a solid dielectric was also investigated.

Figure 3 presents the change in the average breakdown gradients for different magnitudes of displacement of the high-voltage electrode from the network of locations of the metallic inclusion. From the graph in Fig. 3 it is seen that

Fig. 3. Dependence of the average breakdown gradients on the magnitude of displacement of the high-voltage electrode from the location of the vertical axis of the metallic inclusion.

the average breakdown gradients are grouped into three regions corresponding to different directions of the discharge path. In the region of small displacements, from 2 to 5 mm, relatively low values of the breakdown gradients (10 kV/mm) are observed, which increase with increasing displacement of the high-voltage electrode. In this region, each impulse discharge passes through the conducting inclusion. The second region, from 5 to 9 mm, corresponds to discharges which, under the action of increased field strengths on the surface of the metallic inclusion, change their path, tending toward the conducting inclusion. The magnitudes of the average breakdown gradients in this case vary from 11 kV/mm to 15 kV/mm. With a further increase in displacement (more than 9 mm), the influence

of the metallic inclusion does not affect the discharge path, and the average breakdown gradients remain constant and correspond to the electric strength of paraffin without inclusions.

In the experiments it was established that, with a considerable displacement of the high-voltage electrode, the selective direction of the discharge through the metallic inclusion is preserved. Figure 4 presents a plot of the change in the percentage of discharges passing through the metallic inclusion as a function of the relative increase in the discharge path. As can be seen

Fig. 4. Dependence of the percentage number of discharges that passed through the metallic inclusion on the relative increase in the discharge path.

from the graph in Fig. 4, a relative increase in the discharge path of up to 30% does not lead to a change in the selectivity of the discharge path in a solid dielectric, and all 100% of the discharges pass through the metallic inclusion; even with a relative increase in the discharge path to 1.5, about 80% of the discharges pass through the conducting inclusion.

Conclusions

1. It has been shown experimentally that the presence of conducting inclusions in a solid dielectric leads to local increases in the electric-field intensity on the surface of the inclusion, which lowers the value of the average breakdown gradients in inhomogeneous dielectrics.

2. The presence of a conducting inclusion in a solid dielectric leads to a selective direction of the discharge path within the dielectric.