If California grid planners are worried about the famous “Duck Curve,” representing the future disruptions that distributedsolarPV systems will have on the state’s grid, they should check out Hawaii’s “Nessie Curve.”
That’s the term that Hawaii’s utility and grid planners have adopted for the real-world grid disruptions they’re seeing today from the island state’s growing solar PV generation mix. They’ve dubbed it the Nessie Curve after the nickname for the mythic Scottish lake-dwelling dinosaur, the Loch Ness Monster.
Unlike California’s Duck Curve, Hawaii’s curve is happening today, not in the future. Also, unlike California’s projection of reduced, but not erased, demand due to solar PV generation, Hawaii’s curve actually drops “underwater” during peak solar times of the day.
In other words, Hawaii is already seeing enough solar coming onto parts of its grid to start feeding back power onto certain distribution circuits on sunny days, and to drive system-wide demand curves below zero on certain peak days.
This graph is from a presentation made by Dora Nakafuji, director of renewable energy planning for Hawaii Electric Co. (HECO), at last month’s DistribuTECH conference in San Antonio, Texas. This chart represents typical circuits on the island of Oahu, where lucrative incentives have led to a huge increase in distributed solar PV. As of the most recent tally, 10 percent of the island’s customers have rooftop PV, which equates to 29,558 systems with a cumulative nameplate generating capacity of 221 megawatts.
“It’s starting to look like this Loch Ness Monster,” Nakafuji said, pointing to the mid-day sag in residential energy demand, as rooftop solar PV energy supply exceeds the energy demand on those circuits, then the steep curve upward as solar fades away and late afternoon demand increases.
That’s made Hawaii a vanguard in testing the boundaries of how to manage such large penetrations of customer-owned generation, which is outside the utility’s control, and largely unmonitored by utilities and grid operators to manage the balance of energy supply and demand. In a controversial move, HECO has put new interconnection requirements in place for even small-scale rooftop solar PV systems, which has slammed the brakes on new projects and drawn the ire of the solar industry.
As Nakafuji pointed out, one of the clearest representations of this disruption can be seen in the data that tracks daily demand from homes, businesses and other energy users on distribution circuits. Because distributed solar PV causes that demand to drop during sunny mid-afternoon hours and then fades away in late afternoon and evening, it’s giving HECO a much more challenging situation in terms of turning down its oil-fired generators when solar is at its peak, then ramping them up much faster than it’s used to when solar power availability declines.
The radical effect that distributed solar has had on Hawaii’s load curve is demonstrated by the following graph, tracking changes from 2010 to 2013 on average 46kV transformer loading. On one peak day in August, the system experienced a backfeed condition, not just on some individual circuits, but across the system as a whole, she noted.
That’s the same challenge that California grid operators believe may start occurring on that state’s much larger and more integrated grid in coming years, as represented by the well-known “Duck Curve” from California Independent System Operator (CAISO). But while California is facing this effect in the future, Hawaii it facing it today -- and it’s doing so with an island system with far fewer generation resources to tap to manage it.
This becomes a serious problem for managing the addition of new solar systems on heavily impacted circuits, Nakafuji noted. “We’ve noticed that 16 percent of the customers on the feeders who have the PV are already pushing the daytime minimum load up to over 100 percent. That means it’s back-feeding on our system. That becomes a concern,” she said, because circuits that send power back up to distribution transformers and substations cause all sorts of technical and operational challenges, with attendant costs.
Even getting the data to understand what’s happening on Hawaii’s grid has been a challenge, Nakafuji noted. “One of the things about renewable integration is, we really need locational data,” she said. As this map of Oahu’s distribution circuits indicates, certain parts of the island are far more affected by solar PV penetration than others.
“We’ve been deploying sensors out in the field to correlate” information pulled in from existing SCADA systems, she said, since “this was information the traditional system didn’t look at.” HECO is also planning to deploy smart meters, which could add more data to help guide day-to-day grid management decisions, she said.
Then there’s the essential problem of relying on a solar PV fleet for the island’s energy needs, she said -- “What do we do when those renewables aren’t there?” HECO is investing lots of money in weather forecasting systems and distributed energy analysis and management technology, as part of a program funded by the Department of Energy's SunShot grant initiative (PDF).
HECO is also installing grid-scale battery systems, incorporating demand response, and taking other steps to manage this solar challenge. Given that these island problems are expected to become challenges in many more regions in the future, it’s worth keeping an eye on how Hawaii manages them.