With reduced number of components count, Modular Multilevel Converter (MMC) based on ZPUC converter can generate a higher number of voltage levels when compared to well-known topologies. However, the capacitors voltage balancing of ZPUC topology remains a challenge when operating as a submodule due to the high number of capacitor’s voltages to be regulated at the same time in each submodule. This becomes more complicated when several submodules operate and when no self voltage balancing can be achieved such the case of seven-level ZPUC (7L-ZPUC) converter. For this reason, Model Predictive Control (MPC), well known of its superiority in achieving multiple control objectives and superior dynamic performance is introduced in this paper as a controller for the MMC converter based 7L-ZPUC. A Finite Set Model Predictive Control method (FS-MPC) is designed to control a three-phase grid-connected MMC. The proposed FS-MPC is designed to generate thirteen-level output phase voltage and 25-level line-to-line voltages. A detailed real-time model of the converter and its suggested controller is developed and implemented in the real-time simulation software Hypersim to validate the effectiveness of the proposed predictive control. Real-time implementation results verify and demonstrate the good performance of the overall Software In the Loop (SIL) application.

Modular Multilevel Converter (MMC) is a promising topology that gained lot of interest in medium-voltage grid-connected applications due to its modularity, scalability, and adaptability. Although increasing the number of submodules per arm leads to a higher number of voltage levels and better waveforms quality, the size and then the cost of the converter remains a challenge. Therefore, a trade-off between topology complexity and power quality exists. This paper proposes an MMC based on Z Packed U-Cell converter as Submodule (ZPUC-SM) to generate a higher number of voltage levels with reduced number of SMs. When compared to HB-MMC, this topology features reduced components count for generating the same voltage levels. Thus, the candidate topology is considered to be a promising solution for medium-voltage applications connecting the utility. This paper details the design, modeling and control of ZPUC-MMC operating as a three-phase grid-connected converter. The controller task is to regulate output current to fulfill grid connectivity requirements. A Phase shifted-PWM with an integrated voltage balancing algorithm takes in action of controlling the three-capacitors voltages of each arm. Thus, no additional voltage controllers are required. The performance of the proposed converter along with its control system are validated with MATLAB/SIMULINK, and the simulation results validate the robustness and effectiveness of the proposed control during both steady-state and dynamic operations.

In this paper a sensor-less voltage balancing algorithm integrated with phase-shift pulsed width modulation (PS-PWM) is proposed to reduce the size of the flying capacitor (FC) on Packed U-cell (PUC5) converter. The size of the flying capacitor that is achieved by proposed method is the least value heretofore. In conventional modulation methods charge and discharge of the FC is carried out in power frequency. In contrast, in proposed method this process is accomplished in carrier frequency. Increased frequency of FC with PS-PWM carriers leads to highly reduce the FC size. Moreover, all the benefits of Phase-shift modulation such as filters size reduction and total harmonic distortion (THD) reduction is achieved. The comparison between previous literatures and the proposed method is carried out in this paper to compare among the size of FC. Simulation is run by Matlab/Simulink to validate the performance of this proposed method on stand alone and grid connected mode.

Considering the vast application of Power Electronic Converters (PEC) in numerous modern technologies, modeling their behavior is of immense importance. These components are non-linear and modeling their precise behavior requires a relatively substantial amount of computational resources. In this paper, a modeling method based on Gated Recurrent Unit (GRU) networks is proposed that is capable of modeling PECs with the aid of machine learning. Furthermore, the precise datasheets of many commercial over-the-shelf converters are not readily available. Thus, the proposed method is a black-box modeling method that only requires the input and output signals of a converter to model its dynamic behavior. For a performance comparison, the proposed GRU-based method, as well as another similar method based on Long-short Term Memory (LSTM) networks, are used to model a voltage-regulated variable-load DC/DC buck converter. A detailed switching model of the aforementioned converter is simulated in Matlab/Simulink environment to generate the required datasets. This model is also used as a reference for comparison. Ultimately, both of the models are exported to Matlab/Simulink after training to be simulated and assessed.

In this paper, a bidirectional power electronics converter assisted by Photovoltaic Panels for Electric Vehicle battery charging application is presented. The charger is composed of two conversion stages: an AC/DC converter represented by an active-rectifier, and a DC/DC converter illustrated by a Dual Active Bridge. The solar renewable energy is considered an alternative DC source assisting in charging the battery. The charger is tested using MATLAB/SIMULINK under different charging and discharging scenarios. An Electric Vehicle equipped with a bidirectional battery charging system has the ability to act as source or a load.

Multilevel inverters provide various advantages over conventional inverters. However, control and switching these types of inverters is a challenge for an increased number of levels that necessitate a higher number of semiconductor devices. Space Vector Modulation (SVM) switching method is a popular one that provides several advantages over conventional Sinusoidal Pulse Width Modulation techniques. Advantages such as higher output voltage, lower switching loss, lower dv/dt, etc. Despite its advantages, the SVM switching method requires complex time-interval equations. This complexity increases exponentially by increasing the number of levels. In this paper, a new SVM method based on Deep Reinforcement Learning (DRL) is proposed. The latter is based on a Reinforcement Learning Agent (RL-Agent) that tries all the "allowed" switching states of an inverter regarding the reference signal in each time step. Then based on the accuracy of the output, the agent receives a reward or punishment (negative reward). Over several episodes of trial and error, the agent learns how to perform the switching to receive the maximum long-term reward. Since the agent is also rewarded for keeping the voltage of the DC-link capacitors in a given range, this method is capable of voltage balancing. This method is implemented on a three-phase three-level Neutral Point Clamped (NPC) inverter for performance evaluation. The RL-agent is trained in MATLAB while the inverter model is simulated in the Simulink environment. The simulation results demonstrate the decent performance of this method.

Power electronic converters are the crucial components of numerous applications. Due to the switching nature of power electronic converters, they behave nonlinearly. Thus the detailed modeling of such converters imposes a huge complexity and calculation burden on their simulation. This complexity grows exponentially when the detailed models are integrated into the simulation of large networks. In this paper, a large signal modeling technique based on long-short term memory networks is proposed. The black-box approach enables the method to model commercial over the shelf converters. The size of the network provides enough degree of freedom for tuning a trade-off between complexity and accuracy. A DC/DC buck converter is modeled using this technique to validate its performance. The training datasets are obtained from the MATLAB/Simulink switching model of the converter. A long-short term memory network (LSTM) is then trained using MATLAB's deep learning toolbox. The obtained results demonstrate the performance of the proposed method over conventional black box modeling techniques based on neural networks.

This paper proposes a control algorithm that controls a DC/DC converter within a capacitorless battery charger for Electric Vehicles (EVs). The charger is composed of an AC/DC converter connected to a three-phase DAB (Dual Active Bridge) DC/DC converter with no capacitor between those two. The proposed control adds an additional degree of liberty in order to deliver the reference charging or discharging power while mitigating the DC/DC converter's drawbacks, dealing with a varying voltage since there is no capacitor, operating in a wide range power spectrum, and controlling the transitional instants. The control algorithm's performance is verified using Matlab/Simulink, and the results are discussed.