Artificial Intelligence and Electrical & Electronics Engineering: AIEEE Open Access
Uncovering Hidden Causes of Turbine Bearing Vibration Using Neural Network
Abstract
Por Gabor
This paper presents a practical application of Artificial Intelligence (AI) to identify the root cause of excess vibration in a turbine bearing at a conventional power plant. For several years, the vibration occurred sporadically without any physical explanation. Traditional diagnostic methods failed to determine the cause of this excessive vibration in the 5th bearing of the turbine shaft. To address this, we applied a neural network program, using all available measured data from the plant to train the AI system. After training, the AI revealed that the unexpected eigenfrequency of the condenser, combined with the rotational frequency of the shaft, was the actual cause of the event.