Turbine Logic的动态

Looking forward to our own Chris Perullo's talk about digital twins and their use in industry at this year's UPISC Workshop.

Meet Scott Sheppard! As a Senior Tech Lead at Turbine Logic, Scott architects solutions across a wide range of energy generation technologies, with a recent focus on utility-scale solar modeling. His deep passion for clean energy drives his work, and he’s led some of our most impactful PV projects. Outside of work, Scott enjoys hiking and immersing himself in nature. Whether spotting birds, shy forest critters, or even the occasional tarantula, he’s always keen to explore new environments. Recently, Scott took on the trails near the Matterhorn in Switzerland, capturing some incredible views along the way. Check out these photos from his adventures!

What if you could instantly know how your solar site compares to its peers? Our latest benchmark score does just that. After analyzing 23,500 unit-years of performance data using our U.S.-wide PV digital twin, we've developed a grading system to evaluate solar site degradation. The scoring ranges from A to F, based on how much a site deviates from the median degradation rate. With this data-backed grading system, solar sites can now track their degradation and make informed improvements, ensuring long-term efficiency. Want to dive deeper into your site's grade? Reach out, and we’ll walk you through the details.

Ever wonder how solar power plants hold up over time? We’ve analyzed a decade of data to find out. We used our U.S.-wide PV digital twin, which models every large-scale system in the United States, to track how these plants perform since achieving peak output. The data shows annualized performance loss rates across different rated capacities, with median degradation illustrated by black lines. Most plants perform well, maintaining a loss rate between 0% and 1%. However, smaller sites tend to experience higher degradation rates, with a broader range of performance—some even facing severe loss. Meanwhile, larger utility-scale sites generally see lower degradation, although a subset in the 20-50 MW range requires closer attention. Although most solar plants degrade within expected ranges, smaller sites and mid-sized plants (20-50 MW) present more variability. Further analysis is needed to explore how factors like plant age and owner-operator management affect performance. What trends are you seeing in solar performance? Share your experiences or thoughts on how plant age and management could be influencing degradation.

Imagine being able to track and predict the performance of every solar power plant in the country. Turbine Logic’s own Steven Koskey, Scott Sheppard, and Chris Perullo just created a digital twin for every PV Plant in the United States. We’ll use these digital twins to identify broad trends across the solar industry and focus on specific plants that can benefit from improved operations and efficiency. More to come soon!

Meet Chris! As Turbine Logic’s Director of Engineering, Chris architects solutions and manages the day-to-day work for our clients. He’s also responsible for developing our team of engineers. His leadership keeps everything running smoothly. In his free time, Chris enjoys grilling. He loves using different woods – mesquite, cherry, peach, and pecan – to bring out unique flavors in his BBQ. Like all the great pit-masters, Chris cooks his brisket overnight, waking up a few times to check on it. Now that’s dedication to getting the job done right.

In solar plant monitoring, accuracy is crucial for catching real issues without being overwhelmed by false alarms. The goal is to achieve a high True Positive Rate (TPR) while keeping the False Positive Rate (FPR) low. Take a look at the plots: TPR is on the x-axis, FPR on the y-axis. The ideal spot is the lower right corner—100% true faults detected and minimal false alarms. The lines show the ratio of false positives to true positives. Here’s the catch: if only 2% of your hardware is faulted and your system has 90% TPR and 10% FPR, you could see five times as many false alarms as real ones. Even with 5% faults, false alarms still outnumber true ones. Since a majority of the hardware at PV plants is unfaulted, even a small FPR can cause costly disruptions. Deploying a monitoring with a low false positive rate is absolutely critical.

Not every alert in your solar plant is created equal. Monitoring systems identify true positives and false positives. But what does that mean? Imagine sending a truck to perform maintenance, only to find out the inverter is working fine. That’s a false positive—an unnecessary cost in time and money. Multiply that by several incidents a year, and operators begin to question their monitoring systems. True positives, however, catch real faults early and prevent downtime. The challenge? Balancing between catching real issues and avoiding false alarms. Too many false positives lead to distrust, while too few true positives let issues slip by. Next, we’ll dive into how often you can expect false positives versus true positives, based on fault prevalence.

Tracking degradation is one thing, but what about accurately identifying hardware faults? Monitoring hardware is essential to keeping solar power plants running at peak performance. By tracking system performance in real time, these systems help operators detect and address faults, optimize maintenance, and ultimately, ensure maximum energy output. Inverter-level faults are often easy to spot—error codes or noticeable drops in power output make it clear that something’s wrong. But not all faults are so obvious. Smaller issues, like a single panel malfunction, are harder to detect and can silently reduce efficiency. While catching faults is crucial, monitoring systems must also distinguish real issues from false alarms. Too many false alarms lead to wasted resources and distrust in the system. Stay tuned as we dig deeper into understanding the relationship between accurate and false alarms.

Aerial scanning for asset inspection is both expensive and irregular. These challenges hinder consistent monitoring and reliable data collection. Is there a more cost-effective solution available? Integrating existing sensors with advanced analytics is the answer. Combining technologies ensures comprehensive and continuous asset monitoring.