In this paper, a model predictive control (MPC) is investigated for a three-level H-bridge D-STATCOM. To predict the future behavior, a discrete-time model of the D-STATCOM is used. Here in the MPC, a single objective cost function is used to avoid the weighted factors in a multivariable objective function, which in turn leads to complex optimization problem. A single objective cost function controls the error in the reference and actual current. A separate proportional integral controller and a capacitor balancing algorithm are used to self-support the DC capacitor voltage under load variation. The reference current for D-STATCOM is generated by using d-q control strategy for the compensation of reactive power and current harmonics. The switching state which minimizes the current error and the corresponding redundant switching state which meets for the DC capacitor voltage balancing are applied at the next sampling instant. The performance of the investigated MPC control scheme is developed in MATLAB-SIMULINK model and steady-state performance is experimentally verified for non-linear load.

This paper investigates the individual DC capacitor voltage balancing algorithm of a single-phase cascaded H-bridge (CHB) STATCOM. The main purpose of STATCOM connected to the grid is to inject or absorb the reactive power but in the process, it draws active power also to meet the losses in the system. The losses accrued in the H-bridges are uneven because of asymmetries in the modules or because the switching and conduction losses are different in the H-bridges. The individual balancing algorithm equalizes reactive power among the H-bridges but the active power remains unequal in different H-bridges. A combination of PLL and Fourier series method is used to extract the phase information as well as coordinate transformations. The entire system is modeled by an MATLAB-SIMULINK environment with the switching losses represented by using a resistive load across the DC capacitors.