Flow Design Bureau AS Cavitation Intensity Measured on

Flow Design Bureau AS Cavitation Intensity Measured on a NACA 0015 Hydrofoil with Various Gas Contents Jarle V. Ekanger Norwegian University of Science and Technology Xavier Escaler Universitat Politécnica de Catalunya Morten Kjeldsen Flow Design Bureau AS Ellison Kawakami University of Minnesota Roger E. A. Arndt University of Minnesota

Flow Design Bureau AS Project scope • Jarle Vikør Ekanger, MSc in Process Engineering • Currently: Product Engineer in FDB & Ph. D Candidate (NTNU); «Water Quality in a hydropower context» (2011 -2014) – Industrial Ph. D, partly funded by the Norwegian Research Council.

Flow Design Bureau AS Objective • Detect variation of cavitation intensity for a wide variety of cavitation types (i. e cavitation numbers), due to changed water quality. • Verification of cavitation detection using external sensors. • Use frequency spectrum analysis and amplitude demodulation to determine (relative) cavitation intensity.

Flow Design Bureau AS Experimental setup • 3 x B & K 4397 Accelerometers – Uniaxial, fc = 25 k. Hz • 1 x PAC R 6α Acoustic Emission Sensor – fc = 60 k. Hz • All mounted externally Acc. mounted along each geometrical axis on a rod extedended from the foil base mounting plate. Pic before AE sensor moved to here after initial tests

Flow Design Bureau AS Recording signals • Acoustic sensor and acceleremeter signals were recorded using c. DAQ equipment from NI. Periods of 10 s length were logged. • Test section static pressure, test section water velocity, water gas content and water temperature was logged separately, using the water tunnel proprietary system.

Flow Design Bureau AS Finding evidence of cavitation • Collapse of cavitation bubbles is associated with high frequency vibrations and acoustics. • Frequency analysis of the measurement data to detect increased high frequency activity. • The layout of the water tunnel allows visual evaluation of cavitation activity level. – I. e: We had cases that were considered cavitation free by visual inspection. – Comparison of cavitating and non-cavitating cases provide frequency bands that are assumed to hold the cavitation noise.

Flow Design Bureau AS Cavitation frequency band • • (Presumed) Cavitation free signal compared to signal with cavitation. Both signals to same axis scale. 45 k. Hz to 55 k. Hz appear to hold cavitation noise, but there is a general increase of energy in the entire spectrum!

Flow Design Bureau AS Cavitation frequency band • • Scaling the presumed cavitation free signal (new axis on the right), reveals shape similarity between the signals. Suggests that a small amount of cavitation is present, but not visually detectable in the presumed cavitation free case.

Flow Design Bureau AS Confirming cavitation by amplitude demodulation • In the presence of local pressure variations, cavitation activity should be seen to fluctuate. – Examples: Vortex shedding from a hydrofoil, guide vanes in pumps and turbines, penstock dynamics. • Matching amplitude modulation frequencies to hydrodynamic frequencies present in the system provide basis for further analysis.

Flow Design Bureau AS Demodulation analysis • A(t) = abs{fbandpassfilter(t) + i * Hilbert[fbandpassfilter (t)]} • Frequency analysis of A(t) will reveal any dominant modulating frequencies. • Results did not show clear modulation frequencies. • Multiples of the pump frequency were found, should not be transferrable through the water. • It was concluded that intensity variations could not be verified. • Further analysis: Crest Factor

Flow Design Bureau AS Crest factor • Used an algorithm to find the peak amplitude values corresponding to discrete bubble collapses • Averaged peak values relative to the cases identified as cavitation free were plotted against cavitation number.

Flow Design Bureau AS Results

Flow Design Bureau AS Conclusion • Expected intensity modulation effect from vortex shedding could not be verified. • Crest Factor analysis confirms that cavitation noise is affected by the gas content in the water. – This is caused by air content variations in the bubbles at different dissolved gas pressures.

Flow Design Bureau AS Special thanks to: • Roger E. A Arndt – For providing access to the SAFL cavitation tunnel • Ellison Kawakami – For providing operational assistance at the SAFL cavitation tunnel • Xavier Escaler – For good discussions and assistance on analysis.