Abstract

A cavitating two-phase flow of water in a pipe with area shrinkage was experimentally investigated, acquiring at high sampling rate pressure signals and images of the cavitating flow field. The time series of the pressure fluctuations was analyzed in terms of power spectral density and related to the cavitation regimes. Furthermore, the fluctuations of the pressure measurements were also decomposed using the wavelet transform to analyze the frequency distribution of the signals energy with respect to the flow behavior. The energy content at each frequency band of the acquire signals is well related to cavitation flow-field behavior. Moreover, the artificial neural network and the least squares support vector machine (LS-SVM) were implemented to identify the cavitation regime, using, as inputs, the power spectral density distributions of the pressure fluctuations, and some features of the decomposed signals, as the wavelet energy for each decomposition level and wavelet entropy. Results indicate the most accurate model to be used in the cavitation regime identification, underlining the enhanced capability of LS-SVM trained with the input data set based on the wavelet decomposition features

Autore Pugliese

• Lay Ekuakille Aime , Ficarella Antonio , De Giorgi Maria Grazia

Tutti gli autori

• DE GIORGI M.G. , FICARELLA A. , LAY EKUAKILLE A.

Titolo volume/Rivista

IEEE SENSORS JOURNAL

Anno di pubblicazione

2015

ISSN

1530-437X

ISBN

Non Disponibile

Numero di citazioni Wos

Nessuna citazione

Ultimo Aggiornamento Citazioni

Non Disponibile

Numero di citazioni Scopus

11

Ultimo Aggiornamento Citazioni

28/04/2018

Settori ERC

Non Disponibile

Codici ASJC

Non Disponibile