Islanding Detection for DC Microgrids Based on Episode of Care Severity Index

Lack of an effective protection scheme is a major barrier to increased integration of renewable energy sources. Despite the system instability and personnel safety concerns due to unintentional islanding, detection of this abnormal condition has not been adequately addressed in the DC microgrids. To overcome the large non-detection zone (NDZ) and power quality degradation of previous islanding detection techniques, this paper proposes a two-stage hybrid islanding detection technique for DC microgrids. In the first stage, a disturbance is detected by quantifying the severity of voltage change using the proposed superimposed voltage-based episode of care severity index. Then, a superimposed voltage-based current perturbation is injected to distinguish an islanding condition from other disturbances. The proposed technique uses the available distributed generation measurements. It has a near-zero NDZ and does not degrade the power quality during normal operation. The effective performance of the proposed islanding detection technique is verified through several case studies in a +/- 750 V DC microgrid test system.