What if grid operators were unable to track the energy generation from a 2.7-GW solar project – the equivalent of several nuclear plants? Imagine the difficulties they would face.
This scenario is unfolding now in California. However this invisible power is not from one solar project, it’s from the combined generation of a rapidly growing number of rooftop arrays that now amount to 2.7 GW.
In February, one count calculated around 130,000 solar systems generating power from “behind the meter” within the California ISO. By mid-June it was over 150,000.
These systems are mostly net metered, so utilities only recieve enough data to credit the owner the sum of their production/consumption on their bill.
But as these numbers grow, utilities and grid operators are increasingly concerned about the impacts on planning and operations.
“In order for utilities to accurately plan and forecast the impact of distributed PV on their distribution systems, they need production data,” said John ‘Skip’ Dise, product manager at Clean Power Research. Dise works with utilities, ISOs and the California Solar Initiative (CSI) to find a solution to the invisible solar generation on the grid.
However, most distributed PV systems don’t have a production meter, measuring all the power generated, only a net meter. Third party solar, like programs offered by SolarCity and Sunrun, generates production numbers, but it’s not sent to the grid.
“Utilities like PG&E need to see all the power generated to plan distribution, identify which transformers and substations to upgrade, and what they need to allow for back flow when there is so much solar on the system that it’s actually more than the load on a given distribution feeder,” he says.
CAISO Sees Only Big Power Stations
Currently, only utility-scale power stations generate this data, whether it’s from solar, wind, geothermal, nuclear, natural gas or coal.
“Everything on the grid has telemetry, which says ‘I’m on and I’m producing this amount, and this is where I am,’” said Steven Greenlee, spokesman for the California Independent System Operator (CAISO). “It’s data exchange.”
The job of maintaining the grid is about balancing demand and supply. System operators need to know what the load will be to match generation to load. But CAISO can’t see distributed solar generation.
“We are not getting that,” Greenlee says. “That’s all behind the meter, so we can’t see it. And that means we have a harder time producing our load forecast. That is probably one of the major things that we are talking about at this time with utilities and the resource developers and regulators.”
Historically, only utility-scale renewable generation counted towards California’s RPS, so what initially seemed like just a sprinkling of rooftop solar arrays didn’t need to be included in generation data. But now it does.
SolarAnywhere Data and Forecast has long been used by the industry to generate data using satellite on a 1-km grid in California (and 10 km nationally) with unique irradiance measurement in real time. Clean Power Research analytics incorporate current local temperatures, wind speeds, and passing clouds with hourly or half-hourly irradiance measures in real time.
Wind speed and temperature both affect cell temperature, which affects efficiency. Heat reduces efficiency, and cells themselves also generate heat, but on windy days the breeze pulls heat away, increasing efficiency.
It’s all live and responds to the changing weather. “We look at where clouds are currently for real time data, and project them forward; forecasting where they’re moving,” said Dise.
Making the Solar Fleet Visible
By adding analytics to the data streams from both offerings, Clean Power Research, partly funded by CSI, now can supply grid operators with a comprehensive view of all generation with FleetView — including the previously invisible behind-the-meter systems.
PowerClerk includes CSI data on the performance of any given array, at any orientation, using any make of panels and inverters, at any location. It is designed to reward effective solar designs and penalize poor ones. Regulators in more and more states require it to ensure that subsidies aren’t paid for poorly performing systems.
With this, grid operators can determine what they need to do and what they don’t need to do, for example, putting off upgrading a particular substation or transformer if local peak load is now being supplied locally.
“As you’re getting higher and higher penetrations in neighborhoods on specific feeders,” Dise explained, “utilities must make new decisions about how they’re building out their distribution system.”
Utilities already use sophisticated planning software, but “distributed solar has thrown a monkey wrench in that because all of their tools are built around modeling load, but not net load” he said.
But energy modeling can only go so far. San Diego Gas & Electric (SDG&E), with the most solar rooftops in the nation, created its own solar forecasting. Their tools are a work in progress, beginning with smart metering, but have limits, according to SDG&E smart grid manager Josh Gerber.
“You can fine tune them, take data sets from smart meters and run them through algorithms and come up with predictive answers about the future, take actual results, compare them to predictions, and figure out ways to tweak the model for accuracy,” Gerber pointed out. “But until we have a way to actually monitor performance, we’re still dealing with limited validation.”
Gerber believes that the solution is better data at the source: the inverter.
Smart inverters are controllable, can change the power factor, and input or receive reactive power to manage voltage, like capacitors that drive voltage up or down. They tell the grid what they are producing “in a much more real time basis, say a minute, or even faster.”
Germany, like California, has seen rapid residential solar growth. But Germany is already switching its solar customers to smart inverters, so its rooftop solar production remains visible as it grows.
“If all my customers had a solar inverter that we were pulling data off of,” said Gerber, “it would be much easier for us to validate what we predicted against what actually happened.” The data that SDG&E gets is “too late, and too granular.”
“It’s not the way that a utility or district operator would have the visibility into the real time or near real time performance of those systems,” he explained. “It’s a very coarse-grained view.”
Smart Inverters Make Solar Visible — and Valuable
Germany is spending hundreds and hundreds of millions of euros to retrofit its several hundred thousand customers’ inverters to smart inverters.
These cost a bit more than regular inverters, adding an additional cost to a system.
To offset this cost, Gerber suggests that California solar customers with smart inverters could receive a payment, like the German feed-in tariff, for their grid balancing service.
Currently, California solar customers are paid only for their kilowatt-hours of power. But with smart inverters providing a valuable voltage smoothing service, they could also be paid by the kilovar.
“In a future with services unbundled, they could get paid for services they provide to the grid: reactive power and voltage support,” suggested Gerber. “They would have a way to earn more than just the raw energy credit on their net metering bill.”
Lead image: Magnifying glass via Shutterstock