| چکیده انگلیسی مقاله |
Extended Abstract
Introduction
Power transmission lines are essential components of modern electrical networks, playing a crucial role in ensuring a continuous and stable supply of electricity to various sectors, including industries, homes, and essential services. Insulators within these systems are responsible for maintaining the necessary separation between electrical conductors and the supporting structures. This separation is vital for the reliability and safety of power networks, as any failure in the insulators can lead to serious issues, such as power outages and safety risks. Insulators are constantly exposed to various stressors, including mechanical wear, harsh environmental conditions, and electrical surges, all of which can degrade their performance over time and lead to faults. Traditional methods for inspecting power transmission line insulators are not only labor-intensive but also costly and risky for workers, as they often involve climbing transmission towers or using ground-based equipment like binoculars or cameras. These approaches are time-consuming, prone to human error, and limited in scope, leading to inefficient maintenance and the potential for missed faults. The increasing complexity of modern electrical networks demands more efficient, reliable, and safer methods for detecting insulator faults. As such, there is a need for advanced solutions that can overcome the limitations of traditional inspection methods and ensure the continuous operation of power transmission systems.
Materials & Methods
This research proposes a new method for monitoring and detecting faults in power transmission line insulators using aerial imagery captured by Unmanned Aerial Vehicles (UAVs) and analyzed through the YOLOv8 deep learning model. UAVs equipped with high-resolution cameras provide a safe, efficient, and non-invasive way to inspect large sections of transmission lines, eliminating the need for workers to scale towers or navigate hazardous environments. The aerial images captured by UAVs are processed using the YOLOv8 model, which detects insulators and classifies them as intact, broken, or arcing. YOLOv8 model is selected for its real-time object detection capabilities, speed, and high accuracy, making it particularly suitable for analyzing large datasets in a short time frame. The method involves capturing aerial images of transmission lines, labeling the insulators within the images, and using this labeled data to train the YOLOv8 model to detect insulators and classify their condition into three categories: intact, broken, or arcing. Once the model is trained, it can automatically analyze new UAV-captured images to identify potential faults, significantly improving the efficiency, accuracy, and cost-effectiveness of power line inspections.
Results & Discussion
The results of this study demonstrate that the proposed approach is highly effective. The YOLOv8 model achieved an overall accuracy of 93.9% in detecting insulators within the UAV-captured images. Furthermore, the model demonstrated an impressive 99.2% accuracy in classifying the detected insulators into the three predefined categories. These results suggest that the YOLOv8 model is not only capable of accurately identifying and classifying insulators but also of distinguishing between different types of faults, such as broken or arcing insulators. This high level of accuracy ensures that potential issues can be identified early, allowing for timely maintenance and repair before they escalate into more serious problems. In comparison to traditional methods of monitoring power transmission lines, the UAV and deep learning-based approach presents several significant advantages. Firstly, it offers a much faster and more efficient means of inspecting large areas of transmission networks. UAVs can cover vast distances in a short period, and the automated analysis of the captured images further reduces the time required for fault detection. Secondly, the method is highly reliable, with deep learning models providing a level of accuracy that far exceeds manual inspections. Thirdly, it is more cost-effective, as it reduces the need for manual labor and expensive equipment while minimizing the risk of accidents and injuries.
Conclusion
In conclusion, the use of UAVs and deep learning models, specifically the YOLOv8 architecture, offers a powerful and efficient solution for monitoring and detecting faults in power transmission line insulators. This approach overcomes the significant challenges posed by traditional inspection methods, such as labor-intensive processes, high costs, and safety risks to personnel. UAVs allow for quick and comprehensive coverage of large transmission networks, while YOLOv8 ensures highly accurate and real-time detection of insulator faults. The combination of these technologies provides substantial benefits in terms of accuracy, speed, and reliability, significantly reducing the likelihood of undetected faults and enabling timely maintenance. Additionally, it is a cost-effective approach, as it minimizes the need for manual inspections and emergency repairs. As electrical networks expand and increase in complexity, integrating such advanced technologies will play a crucial role in maintaining the stability, safety, and security of power supply systems in the future. |