FCS-MPC Multi-objective in NPC with LCL Filter in GCPVS Under Irradiance Transients

This work presents a multi-objective finite control set model predictive control (FCS-MPC) strategy for a two-stage grid-connected photovoltaic system (GCPVS), composed of a DC-DC boost converter, a three-level neutral-point-clamped (NPC) inverter, and an LCL filter. The proposal addresses the challenges of neutral-point stability and dynamic performance under rapid irradiance variations. The multi-objective cost function integrates the regulation of grid current ( \(i_g\) ), inverter current ( \(i_i\) ), filter capacitor voltage ( \(v_c\) ), and neutral-point balance. In addition, a selection of 19 non-redundant switching vectors is incorporated, improving neutral-point balancing and reducing computational cost. The scheme is implemented in MATLAB/Simulink considering an irradiance profile with steps at 1000, 700, and 450 W/m2, an integration step of \(T_p = 10^{-6}\) s, and the ode1 (Euler) method. The performance is evaluated using standard metrics such as settling time, root mean square error (RMSE), and total harmonic distortion (THD). The results show stabilization times below 50 ms, reactive power ( \(Q_g\) ) close to zero, minimal ripple in the DC-link voltage ( \(v_{dc}\) ), and THD between 2.39% and 4.97%, all within IEEE-519 limits. Overall, the proposal provides a robust and efficient alternative for active power injection ( \(P_g\) ) in photovoltaic multilevel inverters, validating its effectiveness under irradiance transients.