Comparative study of multi‐objective finite set predictive control methods with new max–min strategy applied on a seven‐level packedU‐cell inverter

Abstract

This article studies the design and implementation of multi-objective predictive control (MO-PC) of a grid-connected seven-level packed U-cell (PUC7) inverter for minimising the line current total harmonic distortion (THD) and capacitor voltage error simultaneously. The weighting factor method is usually used as a simple method for solving the control problem in the literature. However, there are some difficulties and shortcomings in the calculation of weighting factors. Here, max-min selection strategy with together priority is adopted to reduce these deficiencies and improves the system performance. The switch model of the PUC inverter is derived and then applied in designing the MO-PC for grid-connected applications, where a controlled active or reactive power is injected into the utility. A comparative study among three strategies of weighting factor, fuzzy decision-making and max-min selection is performed to distinguish the proposed method superiority. Experimental results are given to validate the practicality of the applied controller in regulating the line current and capacitor voltage of the grid-connected PUC7 inverter.

Proportional‐resonant based control strategy for grid‐connected packed‐E cell inverters with Lyapunov filter‐based PLL

“…Various model predictive control (MPC) methods are proposed to solve the control problem of grid-connected PUC inverters [22][23][24][25][26]. Although the MPC methods offer some advantages which include easy implementation, good current (or voltage) quality, and low distortion, they suffer from the disadvantages such as sensitivity to parameter variations and weighting factor requirement [22][23][24][25]. In [26], a Lyapunovbased MPC is proposed for grid-connected PUC inverters.…”

“…On the other hand, a seven-level voltage can be produced when the dc capacitor voltage is balanced to the one-third of dc input voltage. Owing to the reduced number of switching devices and dc capacitor requirement, the performance of PUC inverters is investigated in various applications such as active filter [14,15], dynamic voltage restorer [16], rectifier [17,18], stand-alone inverter [18][19][20][21], and grid-connected inverter [20][21][22][23][24][25][26][27][28].…”

“…However, the grid current is controlled by a proportional-integral (PI) controller whose performance is not satisfactory for ac signals. Various model predictive control (MPC) methods are proposed to solve the control problem of grid-connected PUC inverters [22][23][24][25][26]. Although the MPC methods offer some advantages which include easy implementation, good current (or voltage) quality, and low distortion, they suffer from the disadvantages such as sensitivity to parameter variations and weighting factor requirement [22][23][24][25].…”

Solving predictive control problem of fast‐varying multivariable systems by incorporating unknown active dynamics generated by real‐time adaptive learning machine

“…One of premier strategies in robust MPC is incorporation of min‐max calculations. While min‐max algorithms are amongst the least conservative control algorithms and coincide with strategies regarding consideration of worst‐case configurations in switched systems, they also require significant computational burdens (Abedi Pahnehkolaei, Vahedi, Alfi, & Al‐Haddad, 2019; Hu & Ding, 2019) which may be particularly problematic in MIMO systems given high calculation cost of MPC for systems with increased dimensions. Additionally, such methods may not be easily implemented in case disturbances are of unbounded variation.…”