Application of the Generalized Mean Value Function for Detection of Defects in Metal Cylindrical Slugs

Abstract

In this paper we present a free-oscillation method for acoustic detection of defects in metal rods. For detection of normal rods from rods with defects we use a treatment procedure based on the statistics of the fractional moments. This procedure allows to extract the quanti-tative information from the acoustic signals that are propagated in cylindrical slugs after mechanical strike and having different defects. This information allows separating normal rods (without defects) from the rods having different cuts and differentiating the saw-cut defects with different depth from each other. The analysis of data obtained from this research shows that the proposed method of the fractional moments can be applied for quantitative "reading" of envelopes of different acoustic signals that are obtained by means of free-oscillation method and propagated in dense medium having local defects.