Generator end winding vibration monitoring

The principal focus of this paper is the development of regression based models for the prediction of vibration levels in the end windings of a 288 MVA gas turbine generator. These models are part of a wider end-winding monitoring system. For the purpose of this study, a regression approach was applied. This approach was selected, as it did not require any additional major investment in hardware or software once the original system had been installed. Equipment already available at the plant was utilised to enhance the capabilities of the monitoring system. These models should prove a valuable and straight-forward tool in assessing the condition of the machine end-windings.