Analysis and Compensation of Hall Sensor-Based Position Detection Errors in TPMSLM

The air gap flux density in tubular permanent magnet synchronous linear motor can be approximated as a sine function related to mover position. Therefore, analog hall sensors are used to measure flux density and detect motor position. This technique exhibits the advantages of low cost and small installation volume. The major drawback of this technique is that many interference factors (e.g., magnetic field harmonics and assembling tolerances) degrade position detection accuracy. To compensate for position detection errors, the influences of interference factors on the magnetic field model are analyzed, and a unified formula for position estimation error is derived. A compensation algorithm based on the Luenberger observer and an Adaline neural network (ANN) is further proposed. The Luenberger observer is utilized to estimate raw position, and ANN is used for error compensation. In the experiments performed on a prototype motor, the proposed method is compared with typical algorithms under different operating conditions. Experimental results validate the effectiveness and robustness of the proposed method. The position estimation error of the proposed method is within 120
$\mu \text{m}$
under different operating conditions.