Long Short-Term Memory Customer-Centric Power Outage Prediction Models for Weather-Related Power Outages

Abstract

Severe weather phenomena have become more prevalent resulting in frequent and significant power disruptions and outages. Several weather-related power outage prediction models have been developed. However, most of the developed models focus on predicting outages at the utility’s equipment level, and not at the customer’s level. This paper introduces Long short-term memory (LSTM) power outages prediction models, with high prediction accuracy that have the potential to predict outages at a single customer’s location. The developed models can be deployed into smart energy agents to assist customers in preparing for outages