Abstract
This paper presents a comprehensive sensorless control approach for interior permanent magnet (IPM) motors, integrating high-frequency injection (HFI) and model-based observer techniques to ensure accurate rotor position estimation across a wide speed range. Two HFI strategies—pulsating and rotating—are investigated experimentally and compared in combination with two observer structures: the conventional Sliding Mode Observer (SMO) and Adaptive-Gain SMO (AG-SMO). The AG-SMO dynamically adjusts its observer gain according to the estimated back-electromotive force (back-EMF) amplitude, significantly reducing chattering and improving estimation performance under varying load and noise conditions. A Frequency-Adaptive Complex Coefficient Filter (FACCF) and an Orthogonal Phase-Locked Loop (PLL) are incorporated to eliminate phase delay and enhance demodulation accuracy. Simulation and experimental results obtained using a 30 W, 20 V IPM motor demonstrate that the pulsating HFI + AG-SMO configuration achieves superior stability and noise immunity, while the rotating HFI + AG-SMO provides smoother and more accurate estimation. Overall, the proposed hybrid control framework achieves robust, high-precision, and sensorless operation for IPM motors over the wide speed range, offering a practical solution for applications such as inverter-driven compressor systems operating in noisy environments.