In a past contribution to INMR, Prof. Ravi Gorur explained that utilities sometimes face outages caused by de-capping of porcelain insulator strings – a process whereby the porcelain body separates from the insulator head. The principal mode of failure in this situation is known as insulator puncture or dielectric puncture, defined as a disruptive discharge through the body of a solid dielectric. In the case of insulators, this results in permanent loss of dielectric strength.
Primary insulation in cap & pin insulators is provided by the porcelain ‘sandwiched’ between the end fitting hardware. Since the inter-electrode distance is shortest here, this region is subject to much higher electrical stress than other parts of the insulator. Puncture is evaluated using standard methods such as ASTM D149 or IEC 243-1. Here, thin sheets of the material to be tested are placed between spherical or plane electrodes and exposed to power frequency voltage. The sample is immersed in oil to ensure that any discharge occurs inside the material and not across its surface.
The typical value of dielectric strength for ceramics, as per the literature, is 15 kV/mm. This value is mainly of academic interest since most insulators in practice employ a thickness in the range of 15-25 mm for each bell. That means they are subject to comparatively low electric stress compared to their inherent dielectric strength.
ANSI C 29.1 and also IEC 60383 required the puncture test be performed as a sample test for insulators at time of manufacture. Generally, more than 110 kV is required to cause puncture in a new insulator while, depending on its position in the string, the voltage across each bell varies from only 5 to 20 kV. So what causes the porcelain to puncture at electrical stresses far below the strength measured in the new condition?
Electrical grade porcelain is made by firing a heterogeneous mixture of various ingredients and contains numerous pores and interfaces. While there is a test for porosity in the industry standards that is performed on sample insulators when manufactured, this test will not provide any indication of their internal condition after they have been in service for many years. Indeed, punctures have been known to occur as long as 15-20 years after an insulator has been placed in service.
Many factors are responsible for insulator puncture beyond their composition and firing details during production. These include magnitude and duration of 50/60 Hz overvoltages, lightning and switching surges, mechanical vibration and environmental conditions. It’s therefore very difficult to predict if and when insulators will puncture during service.
Once a puncture occurs, its path can involve either a single or multiple partially conducting channels. These paths grow gradually and in many instances never completely bridge the distance between the cap and the pin. Hence resistance measurements on such insulators can vary over a wide range. For example, good porcelain bells have internal resistance in the range 1010 to 1012 Ω. By contrast, defective porcelain bells are known to yield values over a far wider range (i.e. 0 to 108 Ω), depending on the extent and nature of defect.
From the viewpoint of live-line maintenance, even a few defective units in a string can be tolerated given that such work is almost always carried out under fair weather. It is also possible to assess the actual number of sound units prior to start of maintenance. For example, at 69 kV the number of permissible defective units in the string is one, while in the case of 765 kV this increases to as many as 10. However the presence of even a single defective unit can cause catastrophic failure (line drops) should there be flashover. The resulting power follow current is fairly high (several hundred Amperes) and its path is dependent on the insulator’s internal (i.e. body) and external (i.e. surface) impedances.
On sound units, the internal impedance is much higher than the external value. This causes the current to flow along the surface of the bell. On defective units, however, the current divides and part of it flows inside the bell. It is this internal current and the great electromechanical forces it generates that cause the porcelain to separate from the cap.
External impedance is hard to determine since it depends on thermal ionization (which is a function of arc temperature), moisture (e.g. relative humidity, rain, dew) and contamination. With so many variables involved, it is not surprising that strings with defective units are able to function normally in some locations whereas in other locations there are line drops.
It helps to keep tower footing resistance to a low value (e.g. by using counter poises) since de-capping seem to be more prevalent in areas with high lightning intensity. However, the core of the problem lies in the quality of the porcelain body, which can vary greatly from one manufacturer to the next. A steep front impulse test (as described in IEC 61211) is a good way to evaluate the integrity of the dielectric although it is not always evident whether or not this test can identify poor quality insulators at time of manufacture. Still, since it is often not found in user specifications, it would be prudent to include such a test as well.
A more practical way for users to overcome the problem is as follows: whenever there is scheduled maintenance, remove those units that have been in service for varying periods of time and subject them to resistance measurements, steep front impulse, thermal-mechanical testing, etc. A database can then be created listing test results along with supplier and year of manufacture. This way, the power utility can determine if the problem is manufacturer specific or if it is mainly generic.
With demands for high reliability, timely identification of defective porcelain insulators that are approaching puncture is becoming critical and this cannot be done by visual inspection. There are several suitable instruments for this purpose and their sensitivity increases as the defective insulator’s condition approaches short circuit. Unfortunately, this is rarely the case and well-trained, experienced personnel are needed to correctly interpret test results.
