Model predictive torque and flux control has shown some advantages over the classical methods. However, one of the challenges that still need to be investigated is the establishment of a control balance between the torque and flux which leads to better switching state selection. Traditionally, a weighting factor is used in the classical model predictive control (MPC). There are some new techniques that tried to avoid using a weighting factor in order to select the optimal switching state. However, in most of them, new optimization problems are added to the method. In this research, a simple discrete optimization technique is proposed for weighting factor and switching state optimization. The proposed method can be applied to the full range of operating points. The simulation and experimental results show the validity of the proposed method.

A Multisource Inverter (MSI) comprises several DC sources in the input that can be combined together with varying voltage levels to operate at different loads to reduce the battery size of electric vehicles. Different multisource inverter configurations have been presented recently in the literature. These structures use two DC sources to operate at different demand loads by using a low battery size. The weakness of these multisource inverters is that they use a high number of power switches. In addition, these structures cannot connect two DC used sources together in series to operate under a heavy load, which increases the battery's size, increases high power losses, and reduces efficiency due to a high number of switches. This paper proposes a new topology for multisource inverters that reduces the power electronics switches and the battery size due to generating four combinations between the two used DC sources in the proposed technique. The proposed multisource topology is controlled by the model predictive due to its popular advantages. The performance of the proposal is verified through simulation results in Matlab. The results show that the MPC is a good alternative for such applications due to its simplicity, high performance, and low harmonic content.

The transformer-based inverters in PV systems increase the weight, size, and cost of the inverter while reducing efficiency. This research presents a new PV inverter topology to increase efficiency using a reduction of dc-link. The proposed multilevel inverter is comprised of six power switches, one discrete diode, and three capacitors to produce five voltage levels. The proposed inverter is connected to a PV panel at input and a local grid at output to inject a sinusoidal current waveform into the grid. To control the grid current, a finite set model predictive control is needed to evaluate the proposed inverter. A comparison study is carried out between the proposal and other five-level inverters to verify the strengths and weaknesses of the proposed multilevel inverter. Finally, to demonstrate the performance of the proposed multilevel inverter, the simulation results are presented in the MATLAB/Simulink environment.

Model predictive control brings many advantages and it simplifies the control scheme in power electronics. However, tuning the weighting factor is one of the important open discussions on this topic. There are online and offline methods that have been introduced to select the weighting factor. The online methods are preferred because they are more feasible. In this article, an online weighting factor optimization method based on the simulated annealing algorithm is proposed. The energy of the ripple is used as a convergence criterion. The presented method can be converged in a few steps and it does not impose cumbersome computations. Therefore, the optimal voltage will be identical for a range of the weighting factor. Furthermore, the used search algorithm is parameter independent. The proposed method is implemented for an induction motor but it is also applicable for other applications. The proposed method is validated by the experimental tests.

This paper presents a novel model predictive control (MPC) method for the single-inductor multi-output (SIMO) converter. The novel MPC method combines the conventional MPC strategy and the event-triggered control strategy, i.e. event-triggered MPC (ET-MPC). On one hand, the proposed ET-MPC method inherits the feature of fast dynamic response of MPC to reduce the cross regulation of SIMO converter; On the other hand, the MPC scheme is activated when the state of the SIMO converter triggers a preset triggering condition. Therefore, the unnecessary online computation and switching actions can be avoided, since the MPC scheme is suspended if the triggering condition is inactive. Consequently, the ET-MPC method has two advantages over the conventional MPC method: i) lower computational burden, ii) and less switching actions which contribute to lower switching losses. Moreover, the weighting factor for tuning the switching frequency can be deleted because the unnecessary switching action are all removed. The steady-state operation and dynamic performance cases are studied. The results demonstrate that PS-CRS is able to regulate the SI-MIMO DC-DC converter effectively and robustly.

This paper presents an adaptive stator current disturbance observer based on the model predictive current control (PCC) algorithm. For the various disturbances, the disturbance observer (DO) algorithm estimates the lumped disturbance and makes a feedforward compensation to correct the output. However, the discrete characteristic of current prediction error is neglected in conventional stator current disturbance observer (SCDO). Based on the detailed analysis of the stator current prediction error, the proposed method structures a parallel strategy to observe the discrete disturbance related to each voltage vector separately, and designs an adaptive compensation strategy to eliminate the prediction error. The effectiveness of proposed method is verified using experimental tests in the permanent-magnet synchronous motor (PMSM) system.

The model predictive control is increasingly used as a high performance control strategy in converters and drives. This paper presents a new sensorless predictive power control method for the active front end rectifiers. Sensorless application of predictive control method faces more challenge compared to conventional control methods because the accurate prediction is dependent on the accurate voltage estimation. The model based estimation method is used in the predictive control technique in this research. This technique creates two problems, i.e., the derivatives of the currents, and the need for accurate values of the parameters of the model. The first problem is diminished by using the proposed filters. On the other hand, the inductance is estimated based on the model reference adaptive system observers to improve the accuracy of the line voltages estimation, although the proposed sensorless control is stable even without the inductance estimation. For a robust parameter estimation, a new adaptive function is achieved via Lyapunov technique. Simulation and experimental results verify the performances of the proposed methods.

Finite Control Set Model Predictive Control (FCS-MPC) allows to deal with non-linearities of the system and obtain a fast dynamic response. Therefore FCS-MPC is a good alternative to govern complex power converters or when fast transient operation is required. A problem about the implementation of FCS-MPC is the computational cost, which is bigger for multilevel converters. This paper proposed a new method to implement FCS-MPC reducing the necessary iterations in a 3-phase 4-level Flying Capacitor converter. Simulation results show that the proposed FCS-MPC strategy produces an effective control of the load current, while keeping balanced capacitor voltages with lower necessary iterations.

This paper proposes a model predictive speed control of a permanent magnet synchronous machine (PMSM). The control scheme has a cascade architecture, where the inner loop uses a finite set model predictive control scheme (FS-MPC) for the electrical subsystem, and the outer loop uses a dead-beat model predictive control for the mechanical subsystem. Due to the discrete nature of the control platform an accurate discrete model of the systems is necessary. In this work both systems, electrical and mechanical, are discretizated with a second order Taylor method. Simulation results are presented to validate the proposed control strategy.

Multilevel converters are an attractive alternative for medium voltage applications. The Stacked Multicell Converter (SMC), in particular, is a multilevel converter that allows to increase the output voltage level compared with the classical Flying Capacitor Converter, while decreasing the stored energy in the converter. This paper presents the application of Finite Control Set Model Predictive Control (FCS-MPC) in a three phase SMC with two cells and two stacks. The strategy controls simultaneously the load currents and capacitor voltages. simulation results show that the FCS-MPC strategy produces an effective control of the load current, while keeping balanced capacitor voltages. In addition, it is demonstrated that FCS-MPC outperforms Phase Shifted Pulse Width Modulation with linear controllers in transient and steady state operation.

This work proposes a new control scheme for electrical drives system, named cascaded predictive speed control (PSC). The strategy seeks to maintain the simplicity of the classic predictive control while excluding linear or other controllers. The control strategy has a cascade architecture, similar to the techniques of classical control (FOC or DTC). The outer loop controls the speed of the machine, determining a reference torque through a mechanical dynamic model, which allows tracking the speed reference. The inner loop controls the stator current with a cost function that selects the state of the converter which generates the best tracking references for the stator current synchronous components. Preliminary simulation results confirm the effectiveness of this approach, which produces produces a high quality drive control.