Yapay sinir ağı referanslı PID kontrol yöntemine sahip fotovoltaik panel için maksimum güç noktası izleyici tasarımı

In this thesis, a Proportional Integral Derivative (PID) controller based Maximum Power Point Tracking (MPPT) method is proposed for photovoltaic (PV) power systems, which uses the Artificial Neural Network (ANN) output as a reference value. The proposed method is adapted for the traditional MPPT methods, change by perturb and observe (P&O) and Incremental Conductance (IC) MPPT algorithms. In the proposed method, ANN determines the required maximum power point (MPP) by looking at the instantaneous solar radiation and PV panel temperature, and thus an adaptive reference value is obtained according to the variable solar radiation and panel temperature. An error value is then generated by taking the difference between this reference MPP value and the PV instantaneous power. Finally, this error value is passed through a PID controller to obtain a control signal. This control signal is used instead of the normally fixed duty ratio in P&O and IC algorithms. Thus, adaptive duty ratio is produced according to variable solar radiation and panel temperature and according to a reference value. This can eliminate the disadvantages encountered in P&O and IC methods. The study has been tested by simulation studies on a 100 kW PV array connected to the grid via a boost type DC-DC converter and an inverter. Different radiation and PV panel temperature conditions were taken into account in order to represent the real ambient conditions in the tests. According to the findings, it has been observed that the proposed method has less power fluctuation, high speed response and less ripple ratio in duty ratio compared to traditional P&O and IC methods.