Application of Mathematical Models and Sensors for Optimization of Solar Energy Collection Using a Developed Tracking Device

Photovoltaic panels are less efficient when fixed only to a particular angle of inclination with the sun. However, its efficiency improves with the application of a solar tracker which changes the panel orientation automatically in accordance with the sun's position. In this work an automated solar tracking device was developed, whose control was from either predictive models or sensors signal. The data collected on field were analyzed and the graphical representation was used to determine the most suitable trend. The trends were of Fourier series, linear regression, hyperbolic and exponential function which stands as basis for the predictive models formulated. The most appropriate azimuth and altitude panel orientation were achieved by the developed solar tracking device with the aid of microcontroller and two linear actuators. A programmable intelligent computer (PIC) was used for controlling the linear actuators during cloudy and sunny condition with input signals from the predictive models developed and solar sensors respectively. The correlation coefficient of the models was found to be above 0.95 and 0.92 for azimuth and altitude respectively, indicating high degree of agreement of both predicted and actual angle data. The maximum percentage errors were 4.72% for azimuth and 11% for altitude of the models developed showing an appreciable level of accuracy. The tracker developed has an efficiency of 72.25%. The stationary panel generated a total of 1.458 kW of power stored per day while the average power stored by the tracking panel is 1.916 kW per day. The tracking percentage power gain was estimated to be 46.15% at peak period above the fixed panel from the analysis of the data collected.