Abstract
Shunt capacitor banks (SCBs) are critical power system assets that play a crucial part in providing reactive power support, improving voltage profile, increasing system capacity, and reducing system losses. SCBs are widely installed in the system as they are relatively inexpensive compared with transmission or generation system upgrades. SCBs consists of a number of single phase capacitor units connected in series and parallel to meet the desired voltage and VAr ratings. Continuous overvoltage stress results in capacitor unit failures. If faulty units are not detected and replaced quickly they can cause additional overvoltage stress on the remaining healthy units eventually resulting in a cascading failure. Unbalance protection provides the primary protection against unit failures in capacitor banks by detecting unbalances within the bank. Unbalance protection asserts an alarm if the unbalance magnitude is small, but trips the bank if the unbalance is high enough to potentially cause a cascading failure. Using the unbalance phase angle provides an economical way to quickly locate the faulty units and minimize bank outage time.
All unbalance protection schemes have an inherent problem detecting canceling or balanced failures; i.e., units or elements fail in multiple phases or sections in the bank that cancel or reduce the net unbalance. If the canceling failure is symmetric, then the measured unbalance will be zero. If the canceling failure is asymmetric, then the measured degree of unbalance will be decreased compared to failures in one section of the bank. Both symmetric and asymmetric canceling failures results in reduced sensitivity of unbalance protection and affects the reliability of capacitor bank protection.
This thesis proposes an algorithm that can detect the symmetric and asymmetric canceling failures, thereby improving the reliability of the capacitor bank unbalance protection. The research investigates using a combination of the change of unbalance magnitude and unbalance angle before and after a disturbance in SCBs as a canceling failure detection approach. The proposed algorithm is tested and validated using time domain computer simulation with an electromagnetic transients program. The proposed algorithm is economical and practical as it uses unbalance magnitude and phase angle which are available as part of unbalance protection.