This article is sponsored by OMCO Solar, A MacLean-Fogg Company. In this Voices interview, Solar Power World spoke with Jake Rhodes, firmware engineer at OMCO Solar, about how ray casting and software-based tracker optimization can help solar plants reduce shading and improve power output. Rhodes shared insights on the difference between simple and advanced terrain backtracking, how ray-casting concepts can be applied to site-wide tracker positioning, the role of simulation and numerical optimization and how software and potentially AI can help improve energy production.
Solar Power World: Please tell us about your background and your role at OMCO Solar.
Jake Rhodes: I’m a firmware engineer at OMCO Solar, where I work mainly on research and development for the OMCO Star® Tracker Control System. I was born and raised in Arizona and earned a Bachelor of Science in Computer Engineering from Embry-Riddle Aeronautical University. Outside of work, I enjoy playing guitar.
A major part of my work is looking at how controls and software can help trackers operate more effectively. In this case, that means exploring ways to improve the positioning of site-wide trackers so solar plants can avoid row-to-row shading and increase power output.
Can you explain what ray casting is and how you came to learn about it?
Ray casting is a computational method used to trace ray paths through a virtual environment. It is commonly used in video game graphics and 3D simulation.
A simple way to picture it is to imagine a point or line source in 2D or 3D space. The source casts ray lines down toward a surface. If an object is in that space, the rays intersect with the object. We can calculate where that intersection happens using relatively simple math.
Once we know where the ray intersects the object, we can determine where shadows are. Anything past the surface of the object, along the path where the ray would have continued, is essentially in shadow.
My connection to the idea came from seeing how well lighting and shading are handled in video games. Those graphics use computational methods to define where light hits and where shadows fall, and the same concept became useful when thinking about solar tracker shading.
What is backtracking?
Backtracking is a control method used to avoid shading between rows of solar modules when the sun is low in the sky, in the morning or evening.
There are two forms of backtracking that we use. The first is simple backtracking that applies when the site terrain is flat, and the trackers are at equal height. When the sun is not low in the sky, trackers keep the modules perpendicular to the sun.
But when the sun gets low, keeping every module perpendicular can cause one row to shade its neighboring row. Simple backtracking uses a relatively straightforward calculation to apply a fixed angle offset, so the trackers avoid shading each other while staying as close as possible to perpendicular to the sun.
The second form is advanced terrain backtracking that is used when site elevation varies significantly. In that situation, one tracker may be much higher or lower than its neighboring tracker, which creates a more complex shading problem. The trackers may need to be at different positions when the sun is low.
For example, a tracker at a higher elevation and closer to the sun can shade a lower adjacent tracker much sooner than it would on flat terrain. In some cases, a higher elevation tracker can shade two, three, four or more trackers.
If the sun is on the other side, that same higher elevation tracker may receive little or no shading. There is a full spectrum of possibilities depending on terrain, and the complexity is what makes advanced terrain backtracking harder to solve.
How did you discover the link between ray casting and backtracking?
One weekend, I was playing video games. The week before, we had been talking about advanced backtracking and investigating the different problems and possible solutions around it. While I was playing, I noticed how good and well-defined the lighting and shading were in the 3D graphics.
This led to the thought: why not use this for a solar site? If we could apply ray casting to a site, we could know where light hits each tracker, when it hits the tracker, and where the shading occurs on the tracker. It would give us a full site-wide picture at each sun angle of where light is present and where shading occurs.
But the technical concern was computational complexity. As tracker counts increase, there are many tracker angles to evaluate. Finding the best angle for every tracker could take a long time, especially as sites scale.
The solution is to use optimization techniques that converge on the best solution for each sun-angle position and each tracker angle. The goal is to find the configuration that gives the best site-wide power output for each position of the sun.
Compared with our previous solutions, the improvement is primarily in site-wide power output. Our simulated output suggested up to a half-percent increase in total site-wide power output for simple backtracking.
For advanced terrain backtracking, it is harder to give one exact number because terrain can vary so much. It could be a little more or a little less depending on the site. As the terrain gets closer to flat, the result approaches the simple backtracking number. As the terrain becomes more uneven, the expected power benefit can increase from there.
How can ray casting improve solar power plant operations?
The benefit is that ray casting can determine better tracker positions without adding hardware. Everything we are implementing is software, which means we can update, simulate, re-update and refactor as needed.
Ray casting lets us evaluate where the sun is, where light hits, and where shading occurs across the site. Because we know those relationships, we can evaluate how much power the site generates at a specific time and under specific sun-angle conditions.
The most important periods are when the sun is low in the sky on either side of the site. We can look at how shading changes as the sun moves, then analyze how adjusting the trackers on the west side or east side affects the entire site at the same time.
Right now, we are using numerical methods that take an aggregate view of the possible tracker plane and the different tracker angles for every tracker on the site. The system can then evaluate possible configurations and identify better ways to avoid row-to-row shading and improve total power output.
There is room for AI to improve optimization techniques in the future. Ray casting can generate many possible configurations, and a model could help find the most optimal solution for the entire site.
It is a strong fit for AI because the problem is fundamentally numerical: the model can learn from available configurations, sun-angle positions, shading conditions and resulting power output.
How does this make solar energy less expensive, and by how much?
It makes solar energy less expensive by helping a site produce more power from the same installed equipment. If a solar plant can generate more energy without adding hardware, the cost per unit of energy can improve.
It is difficult to put a dollar estimate because the financial impact depends on the site, the terrain, the scale of the project, the value of the power produced and other variables. Our simulated values indicate an improvement in power output.
Measuring the effect of this approach in the field is also complicated because weather, seasonality, and other variables can change power output. This is why we rely heavily on simulated perfect condition comparisons to understand the before-and-after delta.
We are currently running tests in our research and development lab to determine how the approach performs in practical terms. From what we can see so far, the numbers are solid, and the results are consistently better.
We are still optimizing it. At this point, I would say the approach is about 99% of what it could be. Finding more and better optimization techniques helps push that final 1% and moves us closer to an almost perfect position for every tracker.
OMCO Solar, A MacLean-Fogg Company is America’s premier factory-direct manufacturer of solar trackers and fixed-tilt solutions for solar projects. The company’s expertise in fixed-tilt and single-axis tracker systems stems from decades of steel manufacturing, beginning in 1955 when OMCO Holdings, its parent company, was founded.
To date, OMCO Solar has delivered over 12 GW of factory-direct, high-quality solar mounting structures within the US. With a manufacturing footprint spanning coast-to-coast, the company delivers their American made solar structures built from 100% US steel to developers nationwide. For more information, please visit OMCO Solar at omcosolar.com.