Methodology for characterizing electric power system response and locating the energized capacitor banks using conventional power quality data
A relatively small harmonic current with frequencies near or at the power system parallel resonant frequencies could excite the power system into a resonance condition. While a capacitor bank is not the root cause of the condition, it facilitates and helps cause the problem. This is because when the capacitor bank is energized, the system resonant frequency could shift closer to existing harmonic frequencies produced by nonlinear loads. Therefore, the objective of this dissertation is to quantify the power system characteristics corresponding to the parallel resonant frequencies and system damping. Additionally, since a capacitor bank actively facilitates the resonance condition, the relative or exact location of the involved bank must be determined. This dissertation first presents a practical and accurate methodology to estimate system parallel resonant frequencies by performing spectral analysis of the voltage and current transient data immediately after the capacitor bank switching. The proposed method is also robust in that the accuracy of the resulting estimates is not affected by prevalent harmonics in the system. This dissertation provides two efficient algorithms for estimating the system damping parameters using the Hilbert and analytic wavelet transforms. These algorithms take advantage of the principle of an asymptotic or weaklymodulated signal, for which the signal phase varies much more rapidly than the amplitude. The zero-input voltage response or free response of the capacitor bank energizing can be categorized into these asymptotic signals, and one can assign a unique time-varying amplitude with the system damping information and phase pair by building analytic signals. System model reduction theory also allows us to interpret or quantify this damping as an effective X/R ratio. This dissertation defines two fundamental signatures of shunt capacitor bank energizing. It demonstrates that these two signatures can be utilized to accurately determine the relative location of an energized capacitor bank whether it is upstream or downstream from the monitoring location. This dissertation also presents an efficient and accurate methodology for finding the exact location of an energized capacitor bank. Once a PQ monitor is found to be upstream from the capacitor bank by the relative location finding algorithm, the proposed algorithm can further determine the exact location of the switched capacitor bank by estimating the distance between the PQ monitor and the energized capacitor bank. Thus, one can pinpoint the energized capacitor bank causing the resonance.