التحكم في متتبع الطاقة الشمسية الهوائي باستخدام نظام الاستدلال العصبي الضبابي التكيفي

Controlling the pneumatic solar tracker by using adaptive neuro-fuzzy inference system

Abstract

The sun is one of the most important and indispensable sources of clean energy on our planet. Photovoltaic (PV) energy is nowadays the most viable and accepted solar energy. Solar photovoltaics involves the conversion of sun energy into electrical power. One way to boost the solar PV system’s efficiency is to increase solar radiation time that directed to the solar panel, which means harnessing more sunlight onto the solar panels. This is achieved using sun trackers, which maintain solar PV panels always oriented to the sun’s beams at the optimal angle.
This thesis aims to design and build a prototype of a dual-axis solar tracking system driven by two pneumatic actuators on a smaller scale, which can simulate real time system with an actual load because the pneumatic system has advantages over other actuators due to the simplicity of control, cost-effective, low maintenece and can operate on a wide range of harsh environmental locations. A group of tests were conducted on the tracking system on the day of September 16th 2024 in the city of Baghdad with latitude of (33.367°N) as this location the simulation and the experimental part were conducted.
Addtion to that, a comparison has been made between a fixed solar panel and one driven by pneumatic actuators to show how much power can be produced for a given local load. First, the components of the tracking system will be modeled, which mainly contains the PV panel, the pneumatic actuator system, and the control system. Three control methods has been simulated for the tracking system using Matlab/Simulink: Adaptive Neuro Fuzzy Inference System (ANFIS), Fuzzy Logic (FL), and the compare method to control the pneumatic actuators. The simulation showed that the ANFIS method increased the effeciency of the solar system by approximately 44.6% compared to the fixed panel system that because the ANFIS controller have adaptive features and can predicts the environmental changes due to data training and fuzzy reasoning, while FL method showed a 30.6% increase in solar energy harvest compared to the fixed panel system due to non adaptive properties. The comparison method is a simple process of comparing the readings of two identical optical sensors and has achieved 38.15% greater efficiency than fixe panel since this method took the data directly form the sensors. From the results above, the ANFIS is higher effeciency than FLC controller in about 10.7%.
Comparing the performance of the solar tracker setup in the experimental work under the ANFIS system with the corresponding simulated results, the experimental results showed a decrease of 6.51% over the simulation results. Also, the experimental results of the system under the control of FL and the comparison method gave about 7.6% and 7.25% less than the results of the simulation results, respectively.