An accurate and fast technique for correcting spectral leakage in motor diagnosis

Abstract

This paper presents a technique for correcting the spectral leakage which occurs in windowed time series. The presented technique is aimed to be implemented on an autonomous embedded device based on reconfigurable hardware. This device is intended for being used in remote condition monitoring of motors. As a matter of example, the technique is used to analyse the common frequency components appearing in faulty motors suffering from inter-turn short circuits.