| || Expert System for Signature Analysis
Author : Kalyani, K. ;Lyyengar, K.R.S.;Krishnabrahmam, V.
Source : Defence Science Journal ; Vol:37(4) ; 1987 ; pp 495-505
Subject : 681.3 Computer Science
Keywords : Structureborne ;Machinery
Abstract : Vibrating Machinery Noise,-airborne, structureborne or waterborne-has a signature characteristic of the combination of machinery, its mounting, load coupling, and the structure. FFT and RTA are standard techniques for obtaining power spectral density (PSD) of the noise from the source. In this paper a technique of extraction of features from the PSD, and classification of the sources is presented. An expert system is reported which classifies if the same sources has appeared second time. It also gives the degree of confidence to which the classification is made to the operator.