A Multi-sensor Fusion Method for the Detection of Cavitations in the Hydropower Turbine

An on line monitoring and diagnosis system for a modern power plant is essential as hydropower turbine failures are rarely single mode events. Normally, hydropower turbine deterioration is a slow process involving many factors. Cavitations are most severe, eventually forcing a shutdown for repair of affected parts. Cavitations phenomenon in an operating turbine are ubiquitous and unpredictable, and much different in one type turbine from others. Cavitations are often connected with efficiency change, increase of noise level and vibrations. Consequently, early studies on cavitations detection in hydro-power turbines are indirectly based on observation or feeling of vibration pulsations on surface of the guide-vane stem, the draft tube wall, the lower guide bearing etc or intensity of acoustics emission on the turbine casing. The most feasible approach to detect cavitations must be of less influence to the normal turbine operation conditions, thus we must gain insights into its condition with a nondestructive technique (NDT). The aim in the present paper is to develop an adequate monitoring and diagnosis system on line based on NDT to detect the cavitations occurrence and to quantify its aggressiveness in order to optimize turbine operation and to extend inspections and repairs interval.