Department of Electrical Engineering, Lamar University, Beaumont, TX, USA.
Research Article
World Journal of Advanced Research and Reviews, 2024, 21(03), 1631–1641
Received on 09 February 2024; revised on 16 March 2024; accepted on 19 March 2024
Power grids are endangered yearly by disasters that could well affect our livelihood. This paper investigates an AI (Artificial Intelligence) solution to understand better and ultimately predict disasters using the data from previously observed disasters and artificially generated data. The data is fed to the AI in the form of resiliency calculations, which are handled by another sub-program and then go through stages of linear regression models, normal curving models, and decision tree models. Afterward, the main body assigns a value percentile to each of the predictions and reaches a final prediction that is reliable, fast, and accurate. The AI then reviews its decision and checks it with the actual data to see if the forecast is accurate and refines itself according to the newly gathered data. This Algorithm has massive potential implications for decision-making while facing a disaster. It can also help test out different approaches to the problems in a safe environment before the actual procedure initiates, saving time and reducing costs. This paper is an extension to this lab’s effort to simulate a real-life disaster and grid via disaster data and multi-layer analysis to predict the steps of disaster and the grid’s responses. The raw data (generator’s generations, load’s amount, lines current, etc.) is refined through the resiliency equations introduced by this lab to be evaluated via the algorithm.
Artificial Intelligence; Machine Learning; Natural Disasters; Power Grid; Resilience Analysis; Restoration.
Preview Article PDF
Hossein Rahimighazvini, Milad Hadizadeh Masali, Sahand Saeidi and Reza Barzegaran. Disaster impact prediction in the power grid using artificial intelligence based on Texas synthetic grid data replication. World Journal of Advanced Research and Reviews, 2024, 21(3), 1631-1641. Article DOI: https://doi.org/10.30574/wjarr.2024.21.3.0881