Abstract:
A solar photovoltaic (PV) system is currently gaining prominence as a cost-effective
renewable energy source, making a significant contribution to the energy sector. Its
environmentally friendly nature further enhances its appeal. The main challenge of the PV
system is it depends upon the solar irradiation and any changes in the incoming solar
irradiation will affect badly on the output of the PV system. However, the ongoing effort to
improve rural electrification by enhancing the performance of solar systems continues in
order to effectively utilize the better maximum power point tracking (MPPT) controller. The
Zeta converter, due to its low input and output current ripple, low electromagnetic
interference (EMI), better reliability, input-output isolation ability and better adaptability in a
variety of applications, is used for this study. This work consists modeling of solar panels,
charge controllers, battery system, and Artificial Neural Network (ANN) based MPPT
controller. The proposed approach was considered to design the PV based power generation
system for Arba Minch Zuria woreda Zigit-Bakole health center and Zigit-Merche primary
school, Gamo Zone, Ethiopia. The irradiation of the sun is subjected to vary as time progress
in a day and the required irradiance data was collected for the proposed site. As the collected
data, reveals the proposed site has peak 6-8 hours sunshine with the maximum demand of
about 1197W and 1352W for both sites respectively. Lead acid battery isconsidered to supply
the load during night time. The ANN-based MPPT algorithm is designed to change the duty
cylce of the converter to maximize PV output. The PI-controller based inverter is modeled
for supplying the AC loads in both sites. The results of the proposed work have been
compared with and withoutANNand Pertrub and observe (P&O) based MPPT controller for
Zeta converter. The better performance ANN-based MPPT of the Zeta converter is suggested
for prototype development and implementation.
