Why is it that more than five decades after beingintroduced, there remains lack of a broad consensus on whether compositeinsulators always offer a valid alternative to ceramic insulators? In theopinion of industry expert and observer, Alberto Pigini, the underlying issuebehind contradictory views on the long-term reliability of composite insulatortechnology lies in the fact that performance of any insulator depends greatlyon choice of design given its intended service environment.

Unfortunately, in the case of composite insulators thishas not always been done correctly. For example, inappropriate or toosuperficial a specification from the electrical standpoint can lead toflashover of ceramic insulators. But in the case of composite insulators the resultcould be permanent damage.

Composite insulators offer a number ofwell-documented advantages. However, contrary to what was promoted during theirearly years, they are certainly not ‘indestructible’. Therefore, to ensureperformance comparable to or better than that expected of ceramic insulators,great care must be taken in their:
• specification,
• handling, and
• installation.
With respect to specification, a large proportion of the problems reported withthese insulators over the years can be traced back to deficiencies in selection– especially from the electrical point of view. This is because electricaldesign of composite insulators should not be made looking solely at theirflashover performance during short-term tests. Rather, it must ideally be basedon the risk of surface degradation from partial discharges that, over the longterm, can cause tracking, erosion and eventual failure.

This is a critical shortcoming sincecomposite insulators are highly vulnerable to damage should there be continuouspartial discharges and arcing activity on or near their surfaces. For example,many reported cases of failure have been due to insulators being installedwithout suitable shielding electrodes to limit electric field gradients neartheir high voltage end and even at their earth end in the case of very highsystem voltages.

Similarly, failures have sometimes beenthe result of inaccurate estimation of the real pollution service environment.CIGRE Brochure 142-1999 explained that experience from laboratory ageing testsas well as field trials has confirmed that there are three classes of leakagecurrent on composite insulators under normal wetting conditions:
1. a low-value, highly intermittent class;
2. a relatively high average current of a few mA, but far from values typicalof pre-flashover conditions;
3. a high current value class (i.e. some hundreds of mA) indicating that theinsulator is nearing flashover.
While ceramic insulators are designed looking mainly at the ‘c type’ class ofleakage current, composite insulator units should be designed instead taking ‘btype’ currents into account. In fact, research has indicated that while class‘a’ currents have little influence on long-term performance, class ‘b’ currentscan lead to tracking and erosion and possibly permanent failure.

As a result, there should always besufficient design margin between withstand severity and actual environmentalpollution whenever selecting composite insulators. The critical need is tolimit leakage current over the full service life taking into account thepossible influence of service stresses on surface hydrophobicity andwettability. Therefore, in the case of composite insulators, whether AC or DC,the conventional approach based on pollution classes as defined in IEC 60815can be considered questionable.

Rather, to assure satisfactory serviceperformance, a statistical approach must be made that accounts for bothenvironmental parameters and specific insulator characteristics. In particular,specification in terms of required creepage distance alone is itself notsufficient. For example, the efficiency of a profile may become low if too muchcreepage is forced on a given arcing distance. Indications as per IEC 60815should ideally be regarded more of an ‘orientation tool’ than a substitute forthe information that comes from testing.

Forthose composite insulators already installed on lines and where it is too lateto change specifications, diagnostics based on measuring leakage current alongselected units can help identify possible insufficiency in design and triggerwashing should average leakage current values reach the destructive class ‘b’type.

Whileonly the aspect of electrical design is considered here, suitable specificationfrom the mechanical point of view is of course also important and possibly evenmore so than for ceramic insulators. Again, many reported failures, especiallyaffecting recent generations of composite insulators, have been due toinaccurate mechanical specification or by mishandling and improper installationpractices that do not take into account the permanent damage that can occur.

Inprinciple, the maturity and intrinsic reliability of composite insulators cannow be considered satisfactory and of the same high level as ceramic insulators.However, reliability in practice will depend on whether the electrical andmechanical specifications are accurate and also account for their specialcharacteristics, response to specific service stresses and methods ofinstallation.