Design of a Solar Tree

For a brief summary of this project, watch the video above! More details are below.

This is an independent research project that I have been conducting under the supervision of Dr. Adrian Rodriguez since August 2020. A major issue in using solar panels for energy generation is the land area required for solar installations. That land usually cannot be used for anything else, and land is valuable. While wondering how this problem might be solved, I was inspired by a paper that I came across by Dey et al. online regarding solar tree designs.

The basic idea comes from the premise that a fixed arrangement of solar panels should be optimally oriented to extract maximum power from the sun. The authors of the original paper conducted an analysis to determine such optimum orientations and used genetic algorithms to arrange panels to minimize shading in a tree-like arrangement of panels. However, I wanted to take this a step further and incorporate real patterns found in trees in nature, such as phyllotaxy (the arrangement of leaves on a branch) and the golden ratio.

To begin with, I conducted a literature review of optimum orientations for solar panels and research that has been done on solar trees, to see what approaches others have been taken and help me decide my own avenue of inquiry. I decided to use real insolation data to design a solar tree for Dallas, Texas, which would first involve assessing the solar resource available in Dallas throughout the year. For this I created a typical meteorological year (TMY) dataset relevant to solar insolation for Dallas for my project using data from the National Renewable Energy Laboratory’s National Solar Radiation Database (NSRDB) and Solar Position calculator (SOLPOS).

I identified a model for the insolation striking a tilted surface and wrote a program in MATLAB that implemented this model to determine the orientation that would maximum solar capture for a solar panel for a variety of timespans (months, seasons, semesters and the whole year). My yearly optimum orientation's results were within the ballpark range of other analyses and results that have been done for Dallas, and the patterns seen in the results of the optimum orientation over the months of the year are explicable given how the sun is known to travel in the sky.

In the present setup, the tree performs somewhat worse overall than array-like orientations, which is to be expected since not all panels in the tree are optimally oriented due to their phyllotaxic arrangement and there is shading between the panels due to their arrangement in 3D space. However, from April to August, the tree performs comparably to other array-like options, which is encouraging. Additionally, the tree power output curve at least qualitatively follows the electricity sales curve, suggesting that such a scheme could be feasibly used if scaled properly.

This project is still ongoing, as I want to explore other arrangements of panels and also consider the supporting structure for the panels that would help quantify the land area reduction achieved by using a solar tree. However, I have presented my results so far at the Longhorn Energy Club Student Research Competition, where I won third place in the Energy Generation category, and the Longhorn Undergraduate Research Poster Competition, where I received an honorable mention for my poster design.