Introduction
Transmission lines and high voltage transformer used in electric power delivery are subject to various constraints such as insulator pollutions, which is of prime interest regarding power quality issues and reliability.
Insulator pollution issues are considered as a continuous or intermittent (deposition) of impurities coming from various sources. During rain, at night, or in the morning when cold fog is there, the polluted layer formed on the insulator surface becomes wet and the surface conductivity increases. This causes increasing leakage current, formation of dry bands, initiation of partial discharge and under certain conditions, which may lead to flashover and possibly power outage [
1,
2].
The insulator is considered as two electrodes whose interval comprises three zones. These areas are constituted of three insulators in parallel with different behaviors, which are the air gap, the dielectric material, and the interface air and dielectric material.
The flashover voltage of a polluted insulator depends mainly on the conductivity of pollutants deposits, the degree of pollution, as well as the distribution of this layer on its surface. The flashover of a polluted insulating surface has been the subject of many researches [
3–
10]. The flashover voltage of a polluted insulator is as follows: ① The leakage current flows through the electrolyte that covers the insulation. It causes a heating of the electrolyte which has the effect of increasing the conductivity of the medium and consequently the current. ② The increase in heating due to the joule effect created by the leakage current causes a drying of the polluting layer. The dry zone formed tends to extend laterally until the complete interruption of the current. ③ The applied voltage is practically completely transferred to the terminals of the dry zone and local arcs are likely to appear. In the vicinity of the head of a local arc, the construction of the current lines leads to an enlargement of the dry zone. ④ From this point on, the evolution of the discharge can take place in different ways: the local arc can be extinguished, or it can move laterally to find a more stable position corresponding to a shorter length of the discharge, arc, or it may elongate until it reaches the electrode and thus cause the flashover. In this case, the elongation of the arc takes place on the surface of the electrolyte without forming a dry zone.
In this paper, the impact of the conductivity as well as the distribution of pollution on the behavior of the 1512L cap and pin insulator (Figs. 1 and 2) and its experimental model proposed (Fig. 3) under AC voltage then a comparison has been examined.
The dimensions of the 1512L experimental model proposed in this paper are obtained from the dimensions of the insulator (Table 1). The proposed model has a shape of a glass disc with a radius (r) of 200 mm (Fig. 3) and thickness (e) of 5 mm, having the property of resisting heat due to electrical discharge. The proposed model is provided with two electrodes in aluminum paper. The first one is circular (radius r1= 86 mm) representing the cap and connected to the ground. The second one is circular, too (radius r2= 13 mm), representing the pin related to high voltage supply. The pollution is supposed to be distributed in the form of circular bands on the surface of the proposed experimental model which reproduces the surface condition of the 1512L insulator. For both the 1512L insulator and its proposed experimental model, the artificial pollution is realized by a salt solution (NaCl+ distilled water), having various conductivities. The artificial pollution is applied by pulverizing the surface of the experimental model (Fig. 3), and filling the 1512L (zones (Fig. 4, Table 2)) insulator.
Experimental study
The test voltage is measured using a capacitive divider, connected to the secondary of a transformer test (220 V/140 kV, 5 kVA), whose primary winding is connected to a regulating transformer in order to adjust the voltage with the desired value. It can deliver a secondary voltage ranging from of 0 to 140 kV. An average of 10 tests is performed in each case. Different conductivities of the pollution are used by varying the salt concentration of the salt solution, eight conductivities (1.823 mS/cm, 3.33 mS/cm, 8.02 mS/cm, 12.61 mS/cm, 16.32 mS/cm, 30.5 mS/cm, 50.4 mS/cm, 93.7 mS/cm) are obtained.
The pollution is applied on the 1512L insulator by filling the different zones (Fig. 4) by the salt solution with different conductivities to obtain different levels of pollution (Fig. 4, Table 2).
Flashover process
To observe the effect of the pollution on the behavior of the insulator (real model), one value of conductivities is prepared, and to measure the flashover voltage correspondingly, the following steps must be complied: ① Cleaning the insulator: cleaning the insulator first with distilled water and drying it with papers. Then, cleaning it again with 70o alcohol. ② Preparing the level L1 (Fig. 5) of pollution by filling the zones (Z1+Z2+Z3+Z4) of the insulator (Fig. 4) as is referred to in Table 2 by salt solution. ③ Applying a high voltage until a flashover is obtained. ④ Taking the measurement of flashover voltage. For each level, all steps are repeated until level L8.
In order to study the effect of the pollution severity, another value of conductivities is chosen with respect to all precedent steps and we have repeated all the precedent steps and we’ve taken the measurement.
In order to study the effect of the pollution on the behavior of the experimental model (Figs. 6 and 7), at first, one value of conductivities is prepared, and to measure the flashover voltage corresponding, the following steps must be complied: ① Clean the experimental model. Cleaning the insulator first with distilled water and drying it with papers. Then, clean it again with 70o alcohol. ② Preparing the level L1 (Figs. 6 and 7) of pollution by pulverizing the surface of the experimental model as is referred to in Table 2 by salt solution. ③ Applying a high voltage until a flashover is obtained. ④ Taking the measurement of flashover voltage.
Laboratory observations
It has been observed that the application of a few kilovolts between the electrodes generates a leakage current (initiation of arcs Fig. 8(a) and Fig. 9(a)). The high current density in the vicinity of the HV (high voltage) electrode causes an evaporation of the salt solution, by the Joule effect, and a dry area appears. The increase in the applied voltage causes the lengthening of the arcs in the direction of the opposite electrode (Fig. 8(b), Fig. 9(b)). By increasing the voltage, a critical state is reached, beyond which further increases in voltage causes a total flashover by development of the random arcs (Fig. 8(c), Fig. 9(c)).
Effect of pollution on the flashover voltage
In this section, the variation of flashover voltage according to the conductivity Fig. 10 (real model) and Fig. 11 (experimental model) is investigated. This phenomenon, characterized by the no generation of partial arcs, is due, at the same time, to the nature of the pollution used and the fact that the overall length of the equivalent clean band exceeds the breaking value from which no stable discharge is propagated [
11–
14].
Figures 10 and 11 present the variation of the flashover voltage according to different levels of pollution for real model and experimental model respectively. An increase in flashover voltage following the reduction in the level of pollution is expected. However, this increase is unimportant and does not exceed in the extreme case 30% of the initial flashover voltage in real model and the value of 12% in experimental model.
Effect of pollution severity on the flashover voltage
The variation of flashover voltage according to different conductivities is presented in Figs. 12 and 13.
It is noted that the flashover voltage reveals a clear reduction for conductivities lower than 30.5 mS/cm and is more slowly beyond this conductivity. On the other hand, the reduction in the flashover voltage becomes less accentuated when conductivity is higher. It could be deduced that the isolating system (insulator) is more rigid when the conductivity is weak.
Leakage current
Figures 14–23 present the variation of leakage current according to the level of pollution for each conductivity studied and applied voltage. An increase in current is noted. This is explained by the reduction in the surface resistivity of the clean zones which depends on their temperature. A reduction in the leakage current length caused by the level of pollution is also confirmed.
Figures 24–29 illustrate the variation of the leakage current as a function of the conductivity and the voltage applied.
For low conductivities, it is noted that the increase in leakage current is relatively low for the lower voltage levels applied to 15 kV (lower 25% of the flashover voltage), and discharges at the active electrode are not yet intense, which can explain the low values of leakage currents.
For important conductivity, values of the leakage current are significantly greater with respect to other conductivities, once the voltage level of U= 25 kV (greater than 50% of flashover voltage) is attained.
This is due to the activity of discharges that becomes intense when they exceed 50% of the flashover voltage, which may explain the sudden increase in the leakage current. In addition, the rigidity of the insulation system decreases when conductivity pollution increases.
Conclusions
In this paper, the principal experimental results relating to the influence of the discontinuity of the polluting layer on the behavior of a model insulator of high voltage cap and pin of the 1512L type was exposed and an experimental model was proposed and tested under an alternating voltage. Many tests were conducted in the laboratory for the two models (real one and the proposed model). Several solutions of various conductivities (1.823 mS/cm, 3.33 mS/cm, 8.02 mS/cm, 2.61 mS/cm, 16.32 mS/cm, 30.5 mS/cm, 50.4 mS/cm, and 93.7 mS/cm) and a distribution of pollution (discontinuous by pulverizing the surface of the experimental model and filing different zones of the real one) were applied.
During the experimental tests, the variations of the voltage flashover as well as the leakage current were followed. It is concluded that the flashover and the leakage current change according to the conductivity and the width of the polluting layer.
Indeed, the flashover decreases with the increase in the surface conductivity of pollution. Consequently, the insulator is less rigid when it is applied to a polluting layer. It is obvious that the flashover voltage is affected by the surface quality of the insulator. The flashover is more important in the case of the dry state than in the polluted case. According to the width and the conductivity of pollution, the low voltage of the flashover is obtained for the level of pollution 8.02 mS/cm and conductivity 93.7 mS/cm.
The leakage current increases with the applied voltage, the conductivity of pollution and the width of the polluting layer. There exists a critical mode from which the leakage current increases brutally. This mode is obtained for the great widths of the polluting layer for high applied voltage.
By comparison, from the two models (real and experimental), it is concluded that the experimental model does not reflect exactly the behavior of the real insulator, and considering the complex shape of the insulator 1512L, it is represented by a circular model which is equivalent. However, this model facilitates, on the other hand, the observations and the necessary measures for a good analysis of the physical phenomena of flashover.
PDF
(1615KB)
