Model-Free Predictive Current Control for Three-Phase Vienna Rectifier Based on Adaptive Super-Twisting Sliding Mode Observer

The three-phase Vienna rectifier is widely used in communication power supply, wind power generation, and electric drive due to its advantages of no dead zone between the same switch bridge, high reliability, and low harmonic content of input current. The traditional control methods applied in Vienna rectifiers mainly include proportional-integral (PI), single-period control, etc. These traditional control methods have simple principles and convenient designs. However, with the increasing requirement of Vienna rectifier performance and the increasing complexity of the application, the traditional control method is difficult to obtain a satisfactory control effect. In order to improve the control performance of the Vienna rectifier under the complex environment, a model-free predictive current control strategy based on adaptive super-twisting sliding mode observer (ASTSMO-MFPCC) is proposed. Firstly, an ultra-local model independent of system physical parameters is constructed by analyzing the mathematical model under the condition of model mismatch. Secondly, a super-twisting sliding mode observer is designed to estimate the unknown part of the ultra-local model, which can effectively suppress the influence of system disturbance. At the same time, an adaptive gain was designed to dynamically adjust the parameters of the super-twisting sliding mode observe to solve the gain selection problem. Finally, a two-step predictive cost function is constructed to realize the model-free predictive current control. To validate the performance of ASTSMO-MFPCC, a model-free predictive current control algorithm based on traditional sliding mode observer (SMO-MFPCC) is realized and compared under steady-state, dynamic, and parameter mismatch conditions. The simulation and experimental results show that the bus voltage of the ASTSMO-MFPCC algorithm can reach the steady state faster, and the amplitude of voltage change is small in the given voltage and load mutation experiments. From the simulation results of load mutation, it can be found that the observer parameters can adjust adaptively after the load changes. In the steady-state experiment, the A-phase current THD of SMO-MFPCC is 6.7%, while that of the ASTSMO-MFPCC algorithm is only 4.26%. In the parameter mismatch experiment, the D-axis current of SMO-MFPCC has a significant change and needs a long time to stabilize, and the current fluctuation is as high as 2.7 A. The ASTSMO-MFPCC has a smaller change in D-axis current, and the current fluctuation is only 2.1 A. According to the simulation results of inductance parameter mutation, the chattering observed by the adaptive super-twisting sliding mode observer is significantly smaller. Simulation and experimental results verify the feasibility and correctness of the proposed method, and the following conclusions are drawn: (1) This strategy combines model-free control with predictive control so that the system is independent of the exact physical parameters of the rectifier during operation. (2) The control performance degradation problem of a system caused by the chattering of the traditional sliding mode observer is solved, and the disturbance suppression ability is improved. At the same time, the adaptive gain is designed, and the observer parameters are adjusted dynamically to simplify the gain selection effectively. (3) The comprehensive comparison with the SMO-MFPCC method confirms that the ASTMO-MFPCC strategy has better dynamic and stable performance and anti-interference ability. © 2024 China Machine Press. All rights reserved.