Intelligent Prediction of Alarm Response Time in Modular and Data-Centric Transmission Grid Operation

This paper presents the theory and implementation of techniques to predict the time available for the Control Center personnel of transmission and distribution system operators to respond to an alarming event related to a grid asset. The described techniques include trendline, linear regression, value-at-risk and k-means classification-based prediction and are implemented to support decision-making even with poor quality SCADA data. These techniques have already been deployed in a modular transmission control center alarming and logging system and can be applied for a variety of assets in power systems as well as in other utilities and process industries.