The effect of rotor design on sensorless speed estimation using rotor slot harmonics identified by adaptive digital filtering using the maximum likelihood approach

Adaptive digital filtering has been demonstrated [1] as an effective technique for extracting a real-time, sensorless, speed signal from rotor slot harmonics (RSHs) embedded in the line current waveform of induction motor drives. It is known that sensorless speed estimation techniques using RSHs may exhibit poor performance with certain motor designs. This paper examines the reasons for that poor performance which reflects differences in the magnitude of the slot harmonic signals consequent upon rotor design. Experimental results for a 30 kW motor with 6 different rotors are presented. The significant parameters are the number of rotor slots, skew and the accuracy of construction. It is clearly shown how inferior performance can arise. Conversely, improvements in reliability of speed estimation and transient response can be obtained by recognising those aspects which provide an enhanced signal, by minimising the background noise of the inverter/machine, or by enhancing the adaptive filter. The recursive Maximum Likelihood technique [2] is presented as an improved algorithm for tuning the digital filter which aids transient response and reliability of speed estimation. Real time, experimental transient performance is demonstrated for the different rotors used in this paper and the performance failure of a particular slot combination demonstrated.