Signal processing for extraction of characteristic features of induction motor stator currents at rotor faults

The rapid growth of the signal processing theory resulted in the rise of new methods in diagnostics of the drives. This paper focuses on the application of the complete systems of orthogonal functions, for which fast computational procedures are applicable. It is shown that a spectral analysis employing bases of the Haar functions is highly suitable for both analysis and synthesis of the diagnostic signals in an on-line diagnostics of induction motor rotor faults. The resulting computational structures providing modularity of the fast transform algorithms may be set up of simple digital filter structures. Such an approach enables the synthesis of effective data processing algorithms, specifically when the goal is the speed of computation. These algorithms may be implemented in general purpose multiprocessor systems as well as in multiprocessor systems employing dedicated RISC CPUs.