A New M‐POS‐ANN Controller for the Solar Photo Voltaic Inverter Controlling in the Grid Integrated System

Abstract

ABSTRACT The major challenges with solar photovoltaic‐based grid‐connected systems are poor sensitivity, poor selectivity, improper voltage control and poor power quality parameters like total harmonic distortion (THD), due to which synchronisation between SPV and the grid is difficult to maintain. In the proposed research, the solar photovoltaic is connected to the grid via a boost converter and inverter. The perturb and observe (P&O) method is used for extracting the maximum power from SPV and boost converter controlling. The existing controllers have complex mathematics and occupy a computational burden. In order to overcome such a problem, inverter controlling has been assessed with the implication of a novel design of mixed poison distribution and an artificial neural network (M‐POS‐ANN) controller. Such a novel design of controller is simple with less data computation and easy to apply in comparison to existing methods. The controlling of the inverter is also assessed with the cuckoo search algorithm (CSA) and fuzzy logic controller (FLC). The proposed M‐POS‐ANN controller improves the selectivity to 0.082 p.u., enhances the sensitivity to 5.3 p.u., provides a minimum total harmonic distortion of 3.4% and the least DC link voltage deviation of 2.9% in comparison to CSA, FLC and other existing methods. The effectiveness of the M‐POS‐ANN controller is also tested and validated in the hardware setup of a 2 kW solar module, a 3 kW boost converter, a 5 kW inverter on an STM32F407 controller.