Khoirudin Fathoni, Rizky Ajie Aprilianto, Mario Norman Syah, Abdurrakhman Hamid Al-Azhari, Agus Suryanto
Brushless DC Motor (BLDCM) has become popular for electric drive system due to the advantages compared to brushed DC Motor. Nevertheless, it requires a more complex control system to achieve better performance. Model Predictive Control (MPC) offers improvements in the BLDC drive system. However, model and parameter uncertainties can deteriorate the MPC because of incorrect prediction of the model. This paper proposes Model Free Predictive Control (MFPC) for BLDC drives with Ultra Local Model (ULM) approach. ULM's unknown parts are estimated using Sliding Mode Observer (SMO) to predict the BLDC currents required in the MPC. This paper aims to design the proposed control and investigate the performance in BLDC drive and compare it to traditional MPC in handling parameter uncertainty. The result shows that with proper SMO gain, the MFPCSMO achieved fast dynamic response with lower ripples and steady state errors for both speed and torque performance compared to traditional FCSMPC. In addition, the proposed method is completely independent of BLDCM model as well as motor parameter and produces excellent performance in parameter mismatch conditions with lower ripples. © 2024 IEEE.
Universitas Negeri Semarang, Department of Electrical Engineering, Semarang, Indonesia
