An efficient and cost-effective statistical voltage-based method for passive islanding detection in microgrids

Abstract

Unintentional islanding is defined as the condition in which a section of the power grid continues to be energized by distributed generators after disconnection from the utility grid, for example, due to a fault. This situation remains a significant protection challenge in microgrids with high penetration of inverter-based distributed energy resources (DERs), particularly under near power-balanced conditions where conventional passive methods may exhibit reduced detection sensitivity. This paper proposes a voltage-based passive islanding detection algorithm using a dual exponential moving average (EMA) statistical framework. The method monitors normalized voltage dynamics using a slow EMA for long-term baseline estimation and a fast EMA for short-term statistical variation tracking. An adaptive statistical mechanism is then used to distinguish islanding conditions from non-islanding disturbances. The algorithm requires only basic arithmetic operations, enabling efficient implementation on embedded industrial controllers. The proposed method was validated through both software simulation and hardware-in-the-loop (HIL) testing using an NI CompactRIO interfaced with an RTDS-based multi-DER microgrid. Experimental results demonstrated average islanding detection times ranging from 7 to 20 ms under the investigated operating conditions, while no false detections were observed during load switching, capacitor switching, and DER switching events. Under severe noise conditions (40 dB SNR), four incorrect islanding declarations were recorded. These results indicate that the proposed method remains effective under moderate noise levels but exhibits reduced selectivity under highly noisy operating conditions. From an economic perspective, the proposed approach is cost-effective due to its low computational complexity and minimal hardware requirements, reducing the need for high-performance processors and advanced signal-processing units. The obtained results satisfy the IEEE 1547–2018 islanding detection time requirements and demonstrate the feasibility of real-time implementation on embedded industrial controllers.