DOE and Gage R&R studies help evaluate solar panel efficiency
While working on the Corporate Sustainability team at Collins Aerospace (formerly Rockwell Collins) back in 2011, I was approached by some design engineers who thought the glass technology used in cockpit aircrafts could work well in solar modules. I connected them with my friend in the solar industry, and we took a tour of a nearby solar panel manufacturer to watch their process, so we could understand how to make our own cells for this experiment, and connect with a potential manufacturer if things went well.
We crafted up a designed experiment (DOE) to test out some of the different glass materials and adhesives. The experiment would test the solar cell efficiency to see which combination produced the highest results.
The 24 Factorial experiment tested Glass Type, Glass Thickness, Adhesive Type and Adhesive Thickness. We created 16 different panels with these 4 factors.
Before we proceeded with the experiment, we first needed to perform a Gage R&R study on the testing method, to ensure we could get consistent measurements of the efficiency. This was a new process, which required a special lamp that simulated natural sunlight. We had the two interns take measurements of 10 different solar cells using the test setup they created. The results showed that they used different methods and each got slightly different answers when measuring the same cells. Based on the results, we came up with a standard testing approach, and moved forward with running the experiment.
The results of the experiment resulted in 2 statistically significant factors: Glass Type and Adhesive Thickness. The other two factors were insignificant.
The results showed that it did matter what glass type we used, so we should use PV instead of PVAR (higher efficiency is better). It also showed that higher adhesive thickness was better, so we should use 0.04 thickness. The insignificant factors were also important to understand. Since there was very little difference between the glass thickness, we should use the thinner glass to save costs and material. In addition, the type of adhesive did not matter, so we should go with the less costly or more eco-friendly option.
Ultimately, the efficiency results were not good enough to outperform the current performance in the marketplace. We observed about 17 as the highest results in our experiment, but we could commercially purchase cells that perform over 20. Therefore, we decided not to continue to pursue this option.
