Underwater Solar Photovoltaic (Pv) Systems Performance Analysis and Modelling

Abstract

Photovoltaic (PV) Solar systems have become attractive for powering autonomous systems and various devices. So far, the installation and usage of solar PV systems has been limited to either land or space. Lately, underwater solar PV power generation has attracted interest due to some of its unique application in powering underwater devices. The thermal control and cooling that result makes it more dependable for underwater devices such as underwater autonomous vehicles, cameras, and long-term remote sensing missions. A detailed study was performed aimed at establishing the performance of PV solar panels placed at different water depths for possible applications in underwater power generation. In this work, a set-up comprising of a 1 metre high container customized to create an underwater environment was designed and used in this work. Real-time measurements of the solar irradiance, panel temperature and ambient temperature a t different water depths ranging from 0.0 - 0.6 m were obtained. Data collection was carried out between 11.00 am - 3:00 pm East African Time (EAT) at 30 minutes intervals for 21 days. The solar irradiance was obtained using a solar 02 device connected to the solar I-V analyzer while the ambient temperature and panel temperature were measured using a double PT 300N temperature probe. The results obtained show that panel temperature reduced by 15.48 % at the rate of 0.062 , ambient temperature decreased by 5.13 % at the rate of 0.02 , solar irradiance decreased by 63.79 % at the rate of 9.25 W/m2/cm and the output power decreased by 64.00 % at the rate of 0.13 W/cm with water depth. It was also observed that even at a depth of 0.6 m, a reasonable amount of power ranging between 2.126 - 7.137 W was generated and the maximum power of 19.5 W was generated when the solar panels were floating. A predictive model employing the support vector machine (SVM) and random forest (RF) techniques that was able to predict the expected power output for the system at different water depths was developed. The model proved to be able to simulate the power generated at different water depths, solar irradiance, ambient temperature and panel temperature. Solar irradiance and water depth were the most active parameters affecting the performance of the PV panel while ambient temperature was the least. The predictive model was cast into a predictive app using the shiny app package in R that was tested and validated using the measured data.