Assessment of Remote Cavitation Detection Methods with Flow Visualization in a Full Scale Francis Turbine

This paper describes the experimental investigations carried out in the Francis turbine at Svorka power plant operated by Statkraft in Norway. The unit, with a head of 260 m, can deliver a maximum output load of 25 MW. The rated flow rate is 11 m3/s and the machine rotates at 600 rpm. The turbine runner shows cavitation pitting on the suction side of the blades but some blades present more erosion than others. Moreover, preliminary studies based on remote monitoring of vibrations and acoustic emissions in this particular unit have predicted risk of erosion at high loads and the presence of a draft tube swirl affecting the cavity dynamics. In order to assess the sensitivity of these methods and the validity of the predictions, several acrylic-glass windows have been installed on the draft tube wall to visualize the runner outlet flow during operation. A high speed camera has been used to record the flow field during the tests with rates up to 5000 frames per second. A cavitation detection system has been installed comprising three high-frequency uniaxial integrated electronics piezoelectric (IEPE)-type accelerometers and an acoustic emission sensor, mounted in the turbine guide bearing pedestal and a guide vane arm. In particular, a series of measurements at different operation conditions have been carried out to correlate the simultaneous camera observations with the acceleration and acoustic emission overall levels in high frequency bands. The preliminary analysis of the camera records permits to certify the existence of erosive blade cavitation with the closure region close to the eroded areas at high loads. It can be seen that cavitation appears only in some blades and that it presents different cavity sizes for the same operation condition. As the load increases towards maximum powers, both the number of blades with cavitation and the size of the cavities grow. Moreover, the overall vibration levels also rise as expected.