Deep Learning Topology for Islanding Detection for Distributed Generation (Sags-Swells)
| Author | A ALMUBARAK, ABDULRAOUF ABDULLAH | |
|---|---|---|
| metadata.dc.date.accessioned | 2026-09-20T15:51:52Z | |
| Publication date | 2024 | |
| Degree level | Doctoral | |
| Abstract | Abstract : Distributed generation (DG) presents a viable strategy to meet the rising demand for energy by producing power at the point of consumption through conventional energy or renewable sources. The integration of DGs into microgrids leads to improved power supply reliability for end users. These microgrids can function in either a grid connected or in an autonomous islanded mode. Identifying islanding events in which a DG segment becomes electrically isolated from the main grid is critical for safely linking a DG with the larger utility network. From an engineering management perspective, implementing reliable islanding detection systems requires careful consideration of technology selection, system integration, and maintenance protocols. Additionally, effective coordination among engineering teams, regulatory bodies, and utility operators is essential to establish clear standards, ensure compliance, and optimize system performance. By addressing these management considerations, the seamless integration of DG into the utility network can be achieved while prioritizing safety, reliability, and operational efficiency. Numerous strategies for detecting islanding have been introduced to enhance the detection speed and precision. However, these methods require an infusion of intelligent capabilities to ensure accuracy during low power discrepancies and in the presence of electrical noise. This research is introducing the use of Artificial Intelligence based approaches and algorithms for islanding detection. It involves transforming electrical data from the point of common coupling into images and time series signals. A comprehensive dataset representing various islanding and non-islanding scenarios has been compiled through dynamic system for the development and validation of machine learning and deep learning algorithms. The study uses Convolutional Neural Network (CNN) topology for islanding detection techniques. The innovation is the early islanding detection feature which monitors for faults and normal operations, exploiting the brief interval between relay activation and B circuit breaker disengagement to identify islanding occurrences. The proposed model has been validated using a special dataset made for islanding that has cases where the power is noisy, and the data is messy. The study used 1000 trends across various parameter protections to train, validate, and test the model. Overall accuracy rates were mostly above 94%, highlighting the potential of these models as reliable tools for islanding detection across various protections, demonstrating the high effectiveness of these methods in islanding detection. | |
| metadata.dc.identifier.other | https://digitalrepository.uob.edu.bh/id/492a8bfd-ac50-4740-b9d1-545461a7fad4 | |
| metadata.dc.identifier.other | Drupal node 2060 | |
| metadata.dc.identifier.other | https://digitalrepository.uob.edu.bh/en/dar/deep-learning-topology-islanding-detection-distributed-generation-sags-swells | |
| metadata.dc.identifier.uri | http://165.227.56.122/handle/123456789/7208 | |
| Language | en | |
| Link | https://digitalrepository.uob.edu.bh/en/_flysystem/fedora/2025-08/deep_learning_topology_for_islanding_detection_for_distributed_generation_sags-swells.pdf | |
| Title | Deep Learning Topology for Islanding Detection for Distributed Generation (Sags-Swells) | |
| Type | Doctoral Dissertation |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- node-2060-1-deep_learning_topology_for_islanding_detection_for_distributed_generation_sags-swells.pdf
- Size:
- 8.53 MB
- Format:
- Adobe Portable Document Format