Identification of Abrasive Waterjet Nozzle Wear Based on Parametric Spectrum Estimation of Acoustic Signal

Abstract

The abrasive waterjet nozzle is one of the most critical parts that influences the technical and economical performance of an abrasive waterjet system. In order to control the uniformity of cutting results in milling and cutting, it is necessary to devise a sensing system that can sense on-line the nozzle wear. This paper presents an on-line technique for monitoring the nozzle wear which is based on monitoring the acoustic signals generated by the abrasive waterjet. The autoregressive moving average (ARMA) spectra are used to estimate the nozzle wear. It has been shown that the ARMA spectra can reveal more features of the nozzle wear than the conventional fast Fourier transform (FFT) method. It was found that the amplitude of the spectra and the frequency of the spectra peaks of ARMA models have a high sensitivity to a small variation of the nozzle exit geometry. A method based on on-line acoustic signals is proposed for identification of the nozzle inside diameter.