Application of PMU to detect high impedance fault using statistical analysis

Utilities have recently installed phasor measurement units (PMUs) in their systems. In the control room, massive amounts of data come in from PMUs, but it is hard to utilize in real-time. Situational awareness and event detection are necessary to avoid cascading outages. Tangible algorithms need to be implemented in order to make this abundant data useful to a system operator. Several events cause frequency deviation, including load variation and faults. Of these events, changing loads and switching shunt capacitors on and off is necessary. The developed algorithms need to flag faults without giving false positives for other events. Of the fault types, a high impedance fault (HIF) is the hardest to detect since it may not trip a breaker and causes a smaller frequency deviation from lower impedance faults. This paper proposes an approach to detect a HIF using statistical analysis: a null and alternative hypothesis test is conducted to determine whether a system has diverged from the 60Hz average to another less probable value. The approach uses utility PMU measurements at substations to achieve the standard deviation of the system. A high impedance fault is simulated in Matlab/Simulink, and the bus frequency at the fault location is exported for analysis. Real-time PMU measurements and the simulated frequency deviation are then combined to test whether the HIF would be detectable in the real system. The results show the effectiveness of this approach for HIF detection.