Real-Time Detection of Cavitation for Hydraulic Turbomachines

Cavitation is an important problem in hydraulic machines that negatively affects their performance and may cause damages. Image processing techniques have been often used for visual analysis of cavitation conditions. Real-time detection of cavitation inception is more critical. This is usually performed by measuring related vibrations or electrical impedance, but coarse errors occur. Even if vision inspection is one of the first methods to analyse cavitation damages, this paper presents an innovative approach, i.e. a vision-based system for real-time detection of cavitation inception. Experimentation results in real conditions are reported, which were obtained by testing the system proposed. This technique can be applied to any hydraulic turbo-machine, provided that an optical access is available.