AbstractThe methods for determining transmission line energisation overvoltages incorporate the traditional analogue Transient Network Analyser approach and the more recent mathematical techniques developed for the digital computer. The combination of the two technologies suggest a hybrid T N A system where switching and monitoring of the analogue transmission line model and its response are controlled by a decision-making microprocessor.
This study describes a new approach where the microprocessor is used to reduce the Gibbs phenomena present in uncompensated model line energisation responses. This has previously been achieved by incorporating resistor compensation in the model but the subsequent increase in attenuation reduces waveform rates of rise and affects global response.
Experimental Sigma Factors, analogous to those developed in Fourier Transform analysis, have been devised to locally average the digitised uncompensated line response. They have been applied to uncompensated line receiving-end waveforms resulting from ideal and inductive source energisation of single and 3-phase (simultaneous closure) open-circuited model lines.
These factors significantly reduce the magnitude of the oscillations and achieve the overvoltage accuracy obtained from the compensated line. The resulting processor responses however give steeper rates of rise at response discontinuities as line length increases and also improved global response.
Significant differences in peak overvoltage, between processor and compensated waveforms resulting from inductive source energisation, have been determined for source values of 0 - 0.2H the improvement in accuracy being apparent over an increasing range of source inductance as line length is increased.
Spectral analysis of the model line waveforms was also undertaken and has improved the understanding of the model. Attempts at reducing the Gibbs content by Digital Filtering were not successful.
The Experimental Sigma Factors do not completely reduce the Gibbs content but the results obtained indicate that a modified approach for their implementation may further improve the processor responses already obtained.
|Date of Award||Oct 1986|