Composite reliability analysis of electric power systems considering climatic and spatial conditions using self-organizing map

Abstract

This article presents an innovative approach by integrating different models that represent the failure rates of components related to weather events, considering the geographic location and its variation over time, to a customized Monte Carlo simulation model to ensure the draw synchronized with the load and generation curves, and then quantify the impact of these events through the power flow. The study focuses on using a neural network is trained to classify/filter system states in order to reduce the need for electrical studies and speed up the reliability assessment process. The approach is demonstrated with reference to the IEEE-RTS network test case. The results obtained indicate that the inclusion of renewable sources and climatic conditions have a significant impact on reliability indicators, as failure rates subject to these meteorological conditions tend to increase, providing a greater risk of simultaneous shutdowns, and contributing to the worsening of the system severity index. Therefore, the proposed model allows for a faster and more realistic reliability assessment, directing the planning of expansions and reinforcements and allowing risk monitoring during system operation.