Global Maximum Power Point Tracking for PV Conversion Systems Under Partial Shadings: NNIDA Based Approach

Due to the nonlinear characteristics of photovoltaic (PV) cells, it remains as a challenge to generate stable maximum power from PV conversion systems. Under partial shading circumstances, PV characteristics present multiple local maximum power points. In this paper, a neural network improved dragonfly algorithm (NNIDA) based approach is proposed to improve the performance of tracking the global maximum power point (GMPP). Specifically, the specific points on the
$I-V$
curves are sampled and analyzed such that the various ranges of irradiance and temperature are covered. Then, the iteration best solution of MPP (MPP
$^\mathrm{IB}$
) can be quickly acquired by utilizing the fast approximation characteristics of NNIDA based approach. This process is independent of the configurations of PV modules and has no requirement for irradiance and temperature sensors. Further, the NNIDA based approach can locate the global best solution of MPP (MPP
$^\mathrm{GB}$
) rapidly and precisely in a small interval. Particularly, an adaptive convergence factor is introduced into the NNIDA based approach to accelerate the convergence rate. Furthermore, the inertia weight is developed to improve the tracking accuracy. The comparative simulation and hardware-in-loop (HIL) examples validate the effectiveness of the proposed approach.