Assessment of induction machines rotor fault severity by different approaches

The spectral components at frequencies (1plusmn2s)f of stator current are commonly used to assess induction machines rotor conditions in the motor current signature analysis (MCSA) diagnostic procedure. The sum of the amplitudes of these two components can be used as a reliable diagnostic index. A further step is the quantification of the fault severity, i.e. the definition of an asymmetry factor, able to link the diagnostic index to the number of contiguous broken bars. The quantification of the fault severity is a key item, that is not yet fully established. Several machine models were used to this aim: dynamic deep models for faulty cage machines, like mesh rotor models; d-q or symmetrical instantaneous components models. The latter is preferred as it requires a lower number of parameters. Moreover if system constraints don't allow speed ripple as a consequence of the torque unbalance due to rotor asymmetry, only the component at frequency (1--2s)f appears and steady-state forward-backward models can be used in order to compute this component. In this case analytical relationships can be developed under suitable assumptions. A simple asymmetry factor may be the definition of an equivalent additional resistance DeltaRr linked to the number of contiguous broken bars. In this paper some proposals reported in the literature about the quantization of DeltaR r or about the diagnostic index are reviewed and compared with reference to the case of one broken bar. Experiments prove the effectiveness of the authors' proposa