A multi-task learning for cavitation detection and cavitation intensity recognition of valve acoustic signals

Yu Sha, Johannes Faber, Shuiping Gou, Bo Liu, Wei Li, Stefan Schramm, Horst Stoecker, Thomas Steckenreiter, Domagoj Vnucec, Nadine Wetzstein, Andreas Widl, Kai Zhou

Abstract: Highlights•A multi-task learning framework for simultaneous cavitation detection and cavitation intensity recognition is proposed.•The 1-D double hierarchical residual blocks (1-D DHRB) with large kernel are proposed as a feature extractor on valve acoustic signals.•The sliding window with Fast Fourier Transform (Swin-FFT) data augmentation method is proposed to tackle the small-sample problem.•The proposed method is tested on three different real-world datasets of acoustic signals and compared with the state-of-the-art methods.•The impact of different sampling rate of the acoustics signals on our proposed method for cavitation recognition is investigated.