| Abstract: |
Artificial intelligence integration into power grid fault management is a groundbreaking strategy of improving the reliability of electrical infrastructure and its efficiency of operations. This work explores the use of AI-based techniques in fault detection and localization in sophisticated power grid models, especially the focus on machine learning algorithms, deep neural networks, and smart monitoring systems. The main goals are to compare the effectiveness of different AI methods, determine the accuracy of fault detection in various grid configurations, and determine the difficulties involved in implementing AI in developing countries such as India. The research approach is a system review and a quantitative analysis of secondary sources gathered in form of grid operators, research departments, and international energy organizations. The hypothesis is that AI-powered systems represent much better fault detection and localization resistance than traditional schemes of protection. The findings have revealed that convolutional neural networks have a fault classification accuracy of over 98% whereas the support vector machines have a high fault localization in a transmission line with less than 1.2 errors. A discussion shows that there were significant gains in response time, decrease in outage time, and increased grid stability using AI. The conclusion has made it clear that AI-based solutions provide scalable answers to modernizing the power grid protection infrastructure in a wide range of contexts. |
