In this study, a robust model-free nonsingular terminal sliding mode control (MFNTSMC) algorithm based on the ultra-local model is proposed to reduce the influence of permanent magnet (PM) demagnetization for permanent magnet synchronous motors (PMSMs). First, the PMSM mathematical model in normal and demagnetization is described, and the ultra-local model of speed loop and current loop is constructed based on the input and the output of PMSM vector control system. Then, the MFNTSMC method is proposed and adopted to design speed controller and d-q-axis current controller, and sliding mode observer (SMO) is designed to estimate the unknown terms of the ultra-local model. Finally, compared with the PI control method and Model-Free Control (MFC) method, the results of simulations and experimentations show that the MFNTSMC method can improve the dynamic response while maintaining robustness of PMSM driven system, reduce the dependence of the design of controller on the precise PMSM model, and has fault-tolerant control function for PM demagnetization fault.

A demagnetization fault not only degrades the performance of a permanent magnet synchronous motor (PMSM) drive system, but also may even cause the irreversible demagnetization. This article presents an improved-sliding-mode-observer (ISMO) based equivalent-input-disturbance (EID) method to reconstruct a demagnetization fault and to perform fault-tolerant control for the demagnetization fault in a PMSM drive system. The models of healthy and demagnetized PMSM are first established. Next, an ISMO is constructed by inserting a decoupling term. It completely removes the influence of the motor speed on the reconstruction of a demagnetization fault. Then, the ISMO-based EID approach is developed to reconstruct the fault with the high precision and maintain system control performance even for a demagnetization fault. The stability of the whole system is proved by using the concept of globally uniformly ultimately bounded. Finally, the validity of the method is demonstrated using hardware-in-the-loop simulation.

The parameter mismatch caused by the parameter uncertainties and unknown disturbances degrades the performance of finite-control-set model predictive control (FCS-MPC). This paper presents a model-free fault-tolerant predictive control (MFFTPC) method based on an extended sliding mode observer (ESMO) for the surface-mounted permanent magnet synchronous motor (SPMSM) drive system. First, considering parameter uncertainties and unknown disturbances, a novel ultra-local model (ULM) is established for the PMSM drive system. Next, a finite-control-set model-free fault-tolerant predictive current controller (FCS-MFFTPCC) is designed in the current loop, and the model-free deadbeat fault-tolerant predictive speed controller (MFDFTPSC) is designed in the speed loop. Then, unknown parts of the novel ULM are estimated by the designed ESMO and compensated for the errors caused by the parameter mismatches. Thus, the presented method reduces the dependence on the precise model and eliminates the effect caused by parameter mismatches on the MPC control performance of the SPMSM drive system.

Junction temperature is the main factor affecting the lifespan of insulated-gate bipolar transistor (IGBT). In order to improve the lifespan of three-phase three-level neutral-point clamped (3P-3L-NPC) inverters, this paper proposes a new pulse width modulation (PWM) strategy that reduces inverter power loss based on clamping current. Compared to the traditional space vector pulse width modulation (SVPWM), the proposed method reduces the IGBT junction temperature while ensuring excellent control performance. This will lead to an improved lifespan of 3P-3L-NPC inverter. Furthermore, the proposed PWM strategy maintains the neutral point voltage balance in a wide modulation range. The efficacy of the proposed method is verified by simulation and experimental results.