Prototype
It was desired to analyze the power output of the solar panel in various tracking modes to determine which provided the best cost-to-energy ratio for the customer. This could be done by hooking up a load to the solar panel and recording measurements of the voltage across the load as well as the current flowing through it. The power output from the panel could then be calculated using:
However, given that the prototype would be exposed to the weather and sitting unattended for long periods of time, it was decided that using a standard data acquisition system feeding into a computer would not be wise. This equipment could easily be stolen or water-damaged if left outside. Instead, the team borrowed a voltage datalogger roughly the size of a USB thumb drive from the FSU Mechanical Engineering department. This device could easily fit inside the prototype’s box (or even the junction box on the solar panel) and record voltage measurements every minute over the course of many days, or even weeks. This made taking data very easy, but made the task of computing the panel’s maximum power output more difficult because current was not directly measured.
According to Kyocera specifications [11] the maximum power output of the KC-85T panel is a function of both temperature and solar irradiance, as shown in Figure 22. The datalogger was set up to measure the panel’s open-circuit voltage (I = 0 amps), which increases as the solar irradiance increases and decreases as the ambient temperature increases.
![Text Box: Figure 22 – I-V curves for the Kyocera KC-85T [11] , with maximum power interpolated](prototype_clip_image006.gif)
 
The data shown in Figure 22 was used to generate curves relating open-circuit voltage to estimated max power output for three temperatures, shown in Figure 23. Best-fit curves were generated for the data of form:
where C1 and C2 are functions of temperature as shown in Figure 24. Because the specifications from Kyocera only included panel performance for three temperatures these equations may not perfectly predict C1 and C2, but they will produce usable results.
 
Because equipment to record ambient temperature was not available, data from the nearest weather station (KTLH) was downloaded from www.weatherunderground.com. The Matlab code shown in Appendix F was used to import this data, which contained hourly temperature measurements. The temperature data was used to calculate the power equation in terms of voltage for each voltage reading. The panel voltage was then plugged in to obtain an estimation of the maximum power available at each voltage reading. These power values were averaged with other power values taken at the same time of day over the course of the testing to produce a graph of the panel’s power production during an average day. The area underneath this curve would then be the power production in kilowatt-hours, the standard metric used by utility companies to provide electric power. A sample of this data is shown in Figure 25.
 
It was desired to test the prototype using five different tracking modes: stationary flat, stationary angled, one-axis motor tracking, one-axis reservoir tracking, and two-axis tracking. The prototype was set up near an existing solar array at the FAMU/FSU College of Engineering. The angle and direction for the angled stationary test was roughly 30o with respect to horizontal and facing South, as shown in Figure 26(a). This replicated the angle and direction of the existing solar array, which is optimized for solar conditions in Tallahassee. The panel was set to the same angle with respect to horizontal for one-axis tracking using the motor (Figure 26b), and to the same compass heading for one-axis tracking using the reservoir system.
 
|
Related Links
|
