Electronic Thesis and Dissertation Repository

Thesis Format



Master of Engineering Science


Electrical and Computer Engineering


Ajaei, Firouz B


The ever-increasing penetration of distributed energy resources in power distribution systems has led to challenges in the detection of islanding. Among different islanding detection methods (IDMs), passive methods are the least intrusive and typically require the lowest investment cost. However, they generally suffer from larger non-detection zones (NDZs) and higher nuisance detection ratios as compared to active, hybrid, and remote IDMs. This study provides an overview of the criteria outlined in the existing technical literature for the performance evaluation of IDMs, a review and comparison of the existing passive IDMs, and an analysis of the phase-locked loop (PLL) behaviour under grid-connected and islanded conditions using its quasi-static model. Based on the results of these studies, a novel passive IDM is developed that utilizes the PLL error to detect islanding with a small NDZ and high speed.

The performance of the proposed IDM is evaluated under various islanding and non-islanding disturbances. The performance evaluation studies are conducted through simulations in the PSCAD software environment as well as experimental tests using a hybrid microgrid test platform. The study results indicate that the proposed method: (i) can detect islanding in less than two cycles, which is well below the requirements of the IEEE 1547 standard, (ii) leads to a small NDZ and can identify islanding with only ±5% power mismatch, (iii) does not cause false detection of non-islanding disturbances as islanding, and (iv) is robust against noise.

Summary for Lay Audience

Under normal conditions, the utility grid and distributed energy resources (DERs) operate in parallel to supply power to consumers. If for any reason (e.g., a fault) an upstream circuit breaker opens and disconnects a DER from the rest of the grid, an island may form where the DER continues to feed the adjacent loads without the voltage and frequency support from the grid. Unintentional islanding must be rapidly detected since it leads to safety hazards for utility workers and maintenance personnel.

There are several categories of islanding detection techniques such as passive and active methods. A desirable islanding detection scheme is dependable, secure, robust, fast, cost-effective and non-intrusive. None of the existing islanding detection approaches meets all these requirements. For instance, active methods are generally reliable but degrade power quality. In contrast, passive methods do not adversely affect the power quality but may fail to detect islanding in some cases. This thesis introduces a novel passive islanding detection method which is fast and inexpensive. It operates reliably under most islanding/non-islanding scenarios, does not require communication systems, and is robust against noise. The proposed method is evaluated and validated through both simulation studies and experimental tests.

Available for download on Wednesday, December 13, 2023