Power Plants Failure Reports Analysis for Predictive Maintenance

Vincenza Carchiolo, Alessandro Longheu, Vincenzo di Martino, Niccolo Consoli


The shifting from reactive to predictive maintenance heavily improves the assets management, especially for complex systems with high business value. This occurs in particular in power plants, whose functioning is a mission-critical task. In this work, an NLP-based analysis of failure reports in power plants is presented, showing how they can be effectively used to implement a predictive maintenance aiming to reduce unplanned downtime and repair time, thus increasing operational efficiency while reducing costs.


Paper Citation