Soft computing technique for industrial drive failure identification using JavaNNS and Lab VIEW

TLDR: In this article, the authors used feed forward neural network where the input data's are obtained from the positive and negative sequence component derived from hardware circuit to identify the stator fault and side band frequency of input motor current is obtained from Tektronics Power analyzer is used to identify rotor fault.

Abstract: The application of three phase squirrel cage induction motor as an industrial drive is a common practice. With passage of time these industrial motors are subjected to incipient faults which if undetected can lead to a major fault. Recently artificial neural network, fuzzy logic and genetic algorithm have been employed to assist the diagnosis task and to interpret the data for machine condition. In this paper JavaNNS and LabVIEW have been used as soft computing tools to identify the induction motor faults. Feed forward neural network where the Input data's are obtained from the positive and negative sequence component derived from hardware circuit to identify the stator fault. The side band frequency of input motor current is obtained from Tektronics Power analyzer is used to identify the rotor fault. The result thus obtained is compared with the conventional technique results and have been found much more accurate in identifying the machine internal condition.