Investigations on application of k-factor method for evaluation of impulse waveforms generated during impulse tests on transformers and reactors

摘要

The lightning impulse (LI) voltage test is carried out as routine test on transformers and reactors. The inductive nature of the test object causes superimposed oscillations and/overshoot near the peak of impulse waveforms, which are defined as non-standard waveforms by IEC 60060-1, ed.2. The evaluation of impulse parameters viz. test voltage peak, front time, tail time and overshoot, for non-standard waveforms was subjective and had certain limitations. Experiments were carried out, at different international laboratories, to evaluate the effect of superimposed oscillation and/overshoot on breakdown of different insulation media and a correction factor (k-factor) curve was proposed. Recently, IEC has revised the standard and proposed a new procedure, based on digital filtering, for automatic evaluation of non-standard LI waveform parameters. A low pass digital filter was designed to process the impulse waveforms, and based on frequency of superimposed oscillations and/overshoot the effective test voltage peak is calculated. The impulse test is a destructive test and it is very important to accurately estimate the test voltage, which otherwise may not adequately stress the insulation to required BIL level or overstress the insulation and lead to unintended failure of test object. A program was successfully developed for the implementation of the evaluation procedure proposed by IEC 60060-1, ed.3. Investigations were carried out on practical data, generated during testing of transformers, reactors, bushings and gas insulated substation (GIS) equipment, to estimate the impulse waveform parameters by the developed program. The impulse parameters calculated by earlier and revised evaluation methods are compared. In certain cases it was observed that the evaluation of test voltage peak by revised IEC standard prevents the overstressing of transformer insulation by 5%, which is very critical at higher test voltages. It is very difficult to generate lightning impulse waveforms within the allowable front time for capacitive test objects. The inconsistencies in estimation of time to front in case of GIS test objects by the revised method are discussed in the paper.