Enhancing generator monitoring systems with machine learning for early fault detection

One of the most effective ways of improving hydro plant maintenance efficiency is to install, enhance or modernize condition-based maintenance (CBM) systems. This paper describes how a conventional hydro monitoring and diagnostic system has been upgraded with an early fault detection module based on machine learning methods. The effectiveness of this upgraded monitoring system is demonstrated through a real life case study at a hydropower plant in Croatia, highlighting its potential for use in early fault detection on generators.