Strong parameter dependence and large current ripple are two shortcomings that obstruct the development of model predictive control for voltage source inverters (VSIs). To solve these problems, in this article,a double-vector model-free predictive control (MFPC) method for VSI with visualization analysis is proposed. First, the ultralocal model of the VSI is established and a full-order sliding model observer is designed to estimate the lumped disturbance. MFPC is achieved with enhanced parameter robustness. Then, a double-vector MFPC method is further proposed. By applying two vectors per control period, the current ripple is reduced significantly. A detailed visualization analysis method is proposed, which verifies the effectiveness of the proposed double-vector MFPC method in theory. Besides, the proposed visualization analysis method has the potential to be used to analyze the effectiveness of other types of multivector model predictive control method with different cost functions. Detailed comparative experimental studies verify the effectiveness of the proposed double-vector MFPC method.

To reduce the common-mode voltage (CMV) and eliminate the CMV spikes for 2-level voltage source inverters (2L-VSIs), an improved model predictive control (MPC) based CMV reduction method is proposed considering dead time effects. First, an improved voltage vector (VV) preselection strategy is proposed to reduce the CMV spikes caused by the dead time. Second, a new hybrid VV preselection strategy based on a delay compensation strategy and a modified current sector division strategy is proposed to completely eliminate the CMV spikes. As the proposed hybrid VV preselection based MPC strategy makes full use of the six nonzero VVs of the 2L-VSI, the current total harmonic distortions and ripples are reduced. The simulation and experiments are carried out to verify the effectiveness of the proposed strategy.

A model predictive control (MPC) strategy for grid-connected power converters is proposed in this paper to reduce the current ripples without using AC voltage sensors. Firstly, a sliding mode observer is designed to estimate the grid voltage with an adaptive compensation algorithm. Thus, grid voltage sensors are removed. The novelty of the proposed grid voltage observation method is that it is inherent frequency adaptive without using the accurate grid angular frequency. By implementing the MPC strategy using the estimated grid voltage, the current ripples are reduced, especially under distorted grid voltage conditions because of the low-pass filter used in the observer. Next, to further reduce the current ripples, a double-voltage vector (VV) based MPC strategy is proposed based on the principle of the modulated MPC. Detailed theoretical analyses are also carried out to show its effectiveness for the first time. Finally, experimental studies are carried out to verify the validity of the proposed AC voltage sensorless MPC strategy.

Model predictive control (MPC) method has been developed as a simple and effective current control technique for voltage-source inverters (VSI). The conventional MPC method applied to VSI adopts all the seven voltage vectors (VVs), including one zero VV and six nonzero VVs, to implement the predictive current control in order to improve the current control performance. However, the common-mode voltage (CMV) is large due to the use of zero VV. Although the MPC strategy only using six nonzero VVs can reduce the CMV, its switching frequency is relatively high due to too many switches between nonadjacent VVs. In this paper, a new MPC-based CMV suppression strategy for VSI is proposed and applied to a permanent-magnet synchronous generator (PMSG)-based control system. Only four nonzero VVs, including three adjacent nonzero VVs and one nonadjacent nonzero VV, are utilized to perform the optimization in every sampling period, resulting in switching frequency and calculation effort reduction with no adverse effect on the current response. Moreover, the influence of the MPC-based CMV reduction algorithm on the speed control of PMSG is tested in this paper. The experimental results confirm the effectiveness of the proposed strategy.

To suppress the common-mode voltage (CMV) and reduce the current total harmonic distortions (THDs) and ripples of two-level voltage source inverters (2L-VSIs), a hybrid voltage vector (VV) preselection based model predictive control (MPC) strategy is proposed in this paper. Six synthesized VVs (SVVs) are first defined based on the six nonzero VVs of the 2L-VSIs and a new SVV preselection strategy is proposed to reduce the CMV spikes caused by the dead time effects. Then, by dividing the three-phase current plane into seven sectors with an asymmetric hysteresis, a hybrid-VV preselection based MPC strategy is finally proposed by considering the one step delay effects to completely remove the CMV spikes. Moreover, to further improve the parameter robustness of the proposed method, a control error compensation algorithm is also designed in this paper. Simulation and experimental results show that the proposed hybrid-VV preselection based MPC strategy can not only completely restrict the CMV within $\pm$ udc/6, but also reduce the current THDs and ripples significantly. Additionally, the influences of the parameter variation on the proposed method are also reduced obviously.

Conventional model predictive control (MPC) strategy for voltage source inverters has large current ripple since only one voltage vector is applied in each control cycle. Although the multi-vector modulated MPC strategy can reduce the current ripple, its lack of theoretical basis makes its optimality questionable. To solve this problem, a new visualization analysis method is proposed in this article to evaluate the effectiveness of the conventional multi-vector modulated MPC strategy, based on which a new hybrid multi-vector modulated MPC strategy is further presented to improve the current control performance. First, the principle of the proposed visualization analysis method is introduced, based on which it is verified that the conventional multi-vector modulated MPC has better control performance than the single-vector MPC. Second, based on the proposed visualization analysis method, it is also found that both the conventional double-vector and triple-vector modulated MPC strategies cannot achieve optimal control. Thus, a new hybrid multi-vector modulated MPC strategy is further proposed. By selecting an optimal voltage vector from 18 alternative voltage vector combinations, the current control performance is improved. Detailed comparative experimental results based on experimental platform and StarSim hardware-in-the-loop platform validate the effectiveness of the proposed hybrid multi-vector modulated MPC strategy.