A Hybrid-Harmonic-Filter-Based Position Estimation Method for an SRM With Embedded Inductive Sensing Coils

In this paper, a position estimation method is proposed to accurately measure the rotor position of a switched reluctance motor equipped with embedded sensing coils under both unsaturated and saturated conditions. This method effectively solves three key problems existing in traditional methods based on embedded sensing coils, which are the mutual induced voltage effect, the main magnetic circuit saturation effect, and low accuracy of the rotational speed estimation in near-zero-speed regions. To address the first problem, the sensing coils are connected into a special designed structure. Then, the sensing coil inductance, which varies with the rotor position, is encoded by some voltage signals. After that, a novel hybrid harmonic filtering algorithm based on a delay filter and a coordinate transformation is proposed to remove the dc, even harmonic, and 3
n
th harmonic components of the demodulated voltage signals, according to which the rotor position can be estimated continuously and accurately. By applying this hybrid harmonic filter, the main magnetic circuit saturation effect on the position estimation has been indirectly avoided. To solve the third problem, a starting algorithm based on the voltage thresholds is proposed, and a rotational speed estimation method is developed. Detailed experimental results are provided to verify the validity of the proposed methods.