‘Virtual Inverters’ Are the Future of Solar Power Plant Design

“By the end of 2017, I think it’s going to be the only thing we see in the architecture of large systems.”

Last fall, Cao Renxian gathered 30 employees of the Chinese inverter manufacturer Sungrow together for an impromptu meeting. 

Renxian, who is the founder and chairman of Sungrow, was eager to set an aggressive goal for the 2016 Solar Power International conference.

“We kicked some ideas around and came up with the first 1,500-volt string inverter as the next step in the evolution of cost-effective, large-scale solar power systems,” recalled Allan Gregg, director of application engineering at Sungrow. “Most of the time people think of Chinese companies as copying what’s out there and making it cheaper or adding features to compete. Sungrow proved with this that it wasn’t afraid to lead.” 

The real proof of leadership, of course, didn’t come from wanting to develop the first 1,500-volt string inverter; it came from actually pulling it off, which is exactly what Sungrow did when it unveiled its newest string inverter at SPI in Las Vegas last September.

At the same time, by releasing the first 1,500-volt, 125-kilowatt string inverter, Sungrow also advanced the virtual central inverter concept, which combines the command-and-control advantages of a central inverter with the flexibility and low-cost of string inverters (more about that later). 

Engineering a low-cost string inverter

Gregg is the first to admit that the technical hurdles involved with creating a 125-kilowatt capacity, 1,500-volt string inverter weighing just 130 pounds were not insurmountable. After all, 1,500-volt central inverters have existed for some time, and Sungrow already offers four different 1,000-volt string inverters. 

The real challenge was to design and build a higher-voltage string inverter that was not too expensive. “The achievement is not so much 1,500 volts; it’s the packaging to take advantage of the higher volts and still increase the functionality of the power conditioning system,” said Gregg. “And keeping the price down to where it is going to be directly competitive with central inverters.”

Making a lightweight, low-cost string inverter meant finding the right components, particularly the switching devices that make up the IGBT stack of an inverter. “The heart of any inverter is the switching, the IGBT stack,” said Gregg. “The switching devices and the circuit includes an inductor and a capacitor and to create 3-phase, you have three sets of those and a gate drive circuit that runs them.”

As the string inverter’s voltage goes up from 1,000 to 1,500, the current required decreases, which is what allows the same IGBT stack to deliver more power. “The caveat is those devices need to be able to switch at a higher voltage than what they did at 1,000 volts,” said Gregg.

For Sungrow, the key was to find silicon carbide switching devices that could withstand a higher voltage rating. “We couldn’t pay too much of a premium for them because we didn’t want to push our ultimate product cost too high,” he said. “It’s money engineering and mechanical packaging, so you can put everything in a small box and have it adequately cooled and operate at full power all the way up to 50 degrees Celsius, and then still run in a derated mode up to 60 degrees Celsius.” 

The best of both worlds: The virtual central inverter concept

The 1,500-volt string inverter from Sungrow also enables the virtual central inverter concept. In the past, very large solar power plants -- meaning 100 megawatts and up -- have exclusively used central inverters, usually between 2.5 and 3 megawatts. For solar power system designers, it was a logical choice.

“From an architecture standpoint, that means you have a 100-megawatt system with a reasonable number of command-and-control interfaces, meaning SCADA interfaces,” said Gregg. “But the disadvantage is if you lose one, you lose 2.5 or 3 megawatts of capacity and need to fix it right away, because the production of the system has been affected.” 

There are other disadvantages to central inverters. One is the fact that large and heavy central inverters have high installed costs because they have to be mounted on concrete pads and often require the use of a crane to install and in order to do maintenance. Another problem is uncertainty. “You can run into a situation, which has come up in the last few years, where an inverter company goes out of business and then you still have to maintain an obsolete system for the 20 or 25 years of the project.” 

String inverters, by contrast, are lower-cost than central inverters and sufficiently lightweight to require neither a crane to lift in and out of place nor a concrete pad to rest on. Because string inverters are relatively low capacity, when one fails it doesn’t have a large impact on generation. 

They’re also easy to replace -- a modestly experienced technician can swap one out in as little as one hour. But in large-scale solar power plants, there are disadvantages to string inverters. “You are going to have thousands of inverters for a 200- or 500-megawatt system,” said Gregg. “That means there are thousands of points of command and control, which can become very difficult and costly to manage.” 

That is, unless those string inverters use Sungrow’s virtual central inverter.

Instead of using a single 2.5-megawatt inverter, the virtual central inverter concept utilizes twenty 125-kilowatt string inverters. The difference here is that the twenty string inverters have a single command-and-control point.

“Now I’ve got a 2.5-megawatt inverter that is made up of twenty power modules, but my command and control is through one interface,” said Gregg. “You have instant control over all twenty power modules of my virtual inverter, and you maintain the advantage of smaller string inverters but add to it by having a single command-and-control point.” 

Solar power plant design flexibility

For solar power plant designers, this provides some distinct advantages. No longer is it necessary to connect all the module strings of an array to one inverter.

“With 20 modules, you can choose to put them where you want,” said Gregg. This matters when system designers could benefit from some flexibility.

“Now you can get more creative and adapt to irregular terrain. If you’re working in the desert and you have less than 2 percent grade and a perfect geometric solar array, then it’s not that big a deal,” he said. “But if you start to see irregular patterns and you’re working in and out of shade structures or in hilly terrain, then all of a sudden, using distributed architecture looks good.”

The advantages are so compelling that Gregg believes the virtual central inverter concept will catch on quickly. A number of system designers were already talking excitedly about the possibilities at this year’s SPI conference, he said. 

Some were contemplating the combination of central inverters and string inverters. But even that might not last. 

“As soon as people see these systems installed and working, that is going to be the way to go. By the end of 2017, I think it’s going to be the only thing we see in the architecture of large systems.”