Acoustic pressure threshold prediction in cavitation field based on image and signal processing technique

A novel method is proposed to predict acoustic pressure threshold in cavitation fields by utilizing image processing techniques and parametric resampling technique. Cavitation structure within a water layer of depth λ/4 inside a transparent container was recorded by a high-speed camera, and it was found that a hemispherical bubble cloud attached to the container's solid bottom can affect the morphology of the branched bubble structure beneath the water surface. Due to bubble interactions, the two may bridge together. According to the sequence of binarized image, the structure evolution can be quantitatively predicted. As bubbles coalesce, some large bubbles exist within the bubble clouds. By applying the P-PRTF transform, cavitation noise can be separated from hydrophone detection signals, enabling prediction of the primary acoustic pressure thresholds during cavitation structure bridging: 96.3 kPa at 28 kHz and 110.1 kPa at 40 kHz. It should be noted that more potential factors, such as acoustic frequency, pressure, and liquid properties can influence the merge and separation of the two bubble clusters. The prediction thresholds were also verified through theoretical analysis of the coupled models of bubble oscillations. It reveals that the occurrence of such cavitation events depends on the chaotic threshold. Large hemispherical clusters exhibit a stronger attraction on the floating branched structures, thereby enhancing structural stability. However, increased spacing between the two clusters weakens the vertical component of their interaction force, leading to reduced stability, which closely matches experimental observations. The presented methodology and results will be helpful for further investigations of cavitation erosion prevention.