The widely used PI-controller imposes certain performance limitations when used for the grid-connected VSC systems. Model predictive controllers (MPC) can eliminate these inherent issues of PI-controller but in trade-off with an increased computational burden. Research focuses extensively on finding an intermediate solution by employing artificial intelligence (AI) techniques. This paper discusses the implementation and performance of an artificial neural network (ANN) as an inner current controller. Firstly, the PI and MPC schemes are implemented as a supervisory controller for the grid-connected VSC system. Afterwards, a feed-forward ANN structure is trained with the data set of PI and MPC schemes. The resulted ANN structures are termed ANN-PI and ANN-MPC, respectively. For a fair comparison, Levenberg-Marquardt based backpropagation algorithm is used to train both the ANN controllers with the same number of neurons and hidden layers. All four controllers are implemented in the Matlab/Simulink environment to analyze the performance under different dynamic test scenarios. Further, the performance of all the controllers is evaluated using the OPAL-RT based real-time simulator.

The standard vector control is widely used for grid-connected voltage source converters (VSCs), however, the inability to satisfactorily operate at the desired level under different dynamic scenarios and grid conditions limit its application. This paper presents an artificial neural network (ANN) based vector control for the grid-connected VSCs. The Multilayer Perceptron-Recurrent Neural Network (MP-RNN) approach is used in this work which generates the reference current for the current controller of the VSC. To verify the effectiveness of the proposed control, the MP-RNN-based vector control is implemented for a grid-connected VSC in the MATLAB/Simulink environment and the MP-RNN structure is trained by the Levenberg-Marquardt based backpropagation algorithm. Simulation results are presented and compared with the vector control with and without the proposed ANN-aided control. The results clearly show that the proposed control has a better dynamic performance in damping the oscillation introduced by the dc-link dynamics of the VSC. Further, the dynamic performance of the proposed control has been verified with a model implemented in the OPAL-RT environment considering different dynamic test cases.

Model predictive control (MPC) as a current controller has gained attention in a grid-connected power electronic converter system. The multi-objective predictive control is in greater demand when the photovoltaic (PV) sources are integrated with the utility grid since the inverter alone must ensure all of the control objectives such as grid current following, constant power factor operation and accurate maximum power point tracking (MPPT) performance. In this regard, this paper aims to implement the multi-objective Finite Set Control (FCS) - MPC for a single-stage grid-connected PV system considering the two-stage system is less cost-effective, more loss and less reliable. Moreover, the classical MPPT fails to extract the maximum power under fluctuating and shaded environmental condition. Therefore, an improved MPPT algorithm is proposed in this work which is again combined with the multi-objective FCS-MPC to ensure the extraction of the maximum power from the PV and good transient performance of the grid-side voltage and current. Two different approaches for the dc-voltage tracking are compared, one is implemented with an outer-loop proportional integral controller and the second one is implemented directly in the MPC objective function. The output findings clearly show the single-stage PV system with the proposed control can ensure a fast and accurate maximum power tracking and maintenance of stable output signal throughout all the transient conditions.

Multilevel inverters (MLIs) have been extensively employed to improve the power quality of the photovoltaic (PV) systems. However, the need for large number of components, higher standing voltage, and high harmonic content in the output of a conventional MLI greatly affects the system efficiency. Asymmetrical MLIs have been therefore developed as a suitable alternative to address these issues. This article aims at developing such a hybrid asymmetrical structure suitable for PV application that has a high level per component ratio and minimum standing voltage. The proposed MLI is assembled using a reduced switch H-bridge-based (RSHB) MLI structure with n asymmetrical repeating units and different level doubling circuit (LDC) combinations. The two dc sources used in the repeating units are in the ratio of 1:n voltage ratio, and using n such units, the proposed MLI structures, i.e., PS1 and PS2 can synthesize 4n + 5 and 4n + 7 levels, respectively, at the output instead of 2n + 3 levels with only RSHB MLI. Comparative analysis reveals that both PS1 and PS2 have fewer switches, low standing voltage, less power loss, and lower cost. A 3.9-kW standalone solar PV system is considered for performance evaluation of the PS1 structure applying both the selective harmonic elimination and carrier-based pulsewidth modulation control schemes. In light of this, dc-link voltage balancing and self-voltage balancing mechanism of the LDC are warranted. Extensive simulation of the proposed MLI is performed in MATLAB/Simulink platform under a change in modulation index, sudden load change, frequency change, and step change in solar insolation. Furthermore, theoretical and simulation findings are validated experimentally by performing similar tests on a prototype of the proposed 17-level MLI.

High-penetration of renewable sources and wide application of electric drives has increased the research effort in power converter technology. Multilevel inverters (MLIs) are the well-known power converters that are suitable alternative to the traditional three-level inverters for power quality improvement in different applications. Nevertheless need of more number of switches and dc sources are the shortcoming in conventional MLIs. To address this issue, this paper presents an asymmetric compact module MLI using fewer switch count and less number of dc sources. In addition, the proposed MLI has low standing voltage and less loss as compared to few state-of-art MLI topologies. The developed module can be cascaded further to increase the number of levels. Using n such modules, 16n+1 level output can be obtained. Selective harmonic elimination (SHE) PWM technique is analyzed to verify the operation of a 17-level MLI. MATLAB/SIMULINK platform is used for extensive simulation under different conditions such as at different modulation indices, change in frequency and load dynamic conditions. The theoretical and simulated findings are further validated experimentally on a prototype developed in the laboratory.