Stochastic-gradient-based control algorithms for power quality enhancement in solar photovoltaic interfaced three-phase distribution system

Here, stochastic-gradient-based adaptive control algorithms have been discussed and employed for power quality enhancement in a Photovoltaics (PV) integrated distribution system. Least mean square (LMS), least mean fourth (LMF), sign-error LMS, and epsilon$\epsilon $-normalised LMS (epsilon$\epsilon $-NLMS) have been implemented as control algorithms for the estimation of fundamental load current. The performances of these adaptive algorithms are compared under steady-state and dynamic conditions under the non-linear load conditions in a closed-loop three-phase system. The main aim of implementing these algorithms is reactive power compensation, power quality enhancement, and load balancing in a single-stage three-phase grid-tied PV system. The hysteresis current control (HCC) technique is used to generate switching pulses for the three-phase Distribution Static Power Compensator (DSTATCOM). An MPPT is also employed to ensure maximum power delivery from the solar PV array. PV integrated three-phase single-stage distribution system with adaptive control algorithms is implemented in MATLAB/Simulink environment as well as in experimental environment to achieve the goals per standard IEEE-519. 1. Performance of least mean square (LMS), sign-error LMS, & varepsilon;-NLMS, and least mean fourth (LMF)-based adaptive control algorithms with different learning rates has been analysed.2. PV array has been integrated into a single-stage topology with an maximum power point tracking (MPPT) facility.3. Reactive power compensation and harmonic elimination have been achieved along with active power injection to the grid.4. Performance of LMS, sign-error LMS, & varepsilon;-NLMS, and LMF have been compared and discussed under different conditions. image