Facing growing uncertainty from intermittent resources, MISO and other grid operators must increase use of demand resources and provide market participants with tools to hedge risks, academics, consultants and others told the RTO’s Market Symposium this week.
“The challenge facing ISOs around the world in the next decade or 15 years is quite unprecedented,” said Andy Sun, associate professor at the Georgia Institute of Technology’s H. Milton Stewart School of Industrial and Systems Engineering. “We’re moving into territory I think we haven’t seen before in terms of the uncertainty that will be in the system. Demand and supply are all becoming a lot more stochastic — which means it’s both uncertain and also dynamic.”
Scott Harvey of FTI Consulting said California’s and Australia’s challenges in managing higher levels of solar output foreshadow MISO’s future and the limits of markets.
“One of the key things we have to look at is: Are the markets consistent with the actions the operators are taking? Are they reinforcing those actions?” he said. “What we’re seeing in California is [that] the failure of some of the tools to deliver — to perform as intended — has led to a continuation of operators taking out-of-market actions.”
Harvey and Sun were among the panelists in a discussion Tuesday on “Addressing Uncertainty, Variability and Risk via Power Markets and Operations.”
Need for DR
Harvey and Ross Baldick, professor emeritus in the Department of Electrical & Computer Engineering at the University of Texas, Austin, also called for more emphasis on demand resources.
More than 90% of demand response in MISO is only available in the lead up to an emergency, and there is little DR in much of MISO South, according to testimony by former Ohio regulator Paul Centolella that was filed with a complaint over state DR opt-out policies in MISO on Tuesday. (See DR Firm Challenges FERC, MISO on State Opt-out.)
“I think there is an incredibly invaluable role that needs to be played by the demand side,” Baldick said. “We’re simply, in a high renewable world, moving away from the presumption that we can serve all demand no matter what. And the way I would like to see that worked out is that we have a lot more price-based demand response that gives us gigawatts and gigawatts of flexibility.”
Harvey said the most important areas for MISO to improve in the operational time frame are DR and ensuring the demand curve and operating reserve demand curve provide resources needed to balance variations in net load and reduce operators’ “ad hoc” actions.
“We need to get demand response into the market — not just have demand response be something you pay for providing phantom response, but actually be there, providing regulation [and] reserves that the operators can see and count on.
“If we’re going to get into energy-limited situations, I think we’re going to need demand-side resources that can provide 30-minute reserves or longer; reserves that can drop off for periods of time to enable us to conserve energy, and not necessarily during emergency conditions,” he continued. “We may need to conserve energy before we get into the emergency so we can get through the emergency. So that’s another [thing] we need to study.”
In a panel discussion Wednesday on “Infrastructure as an Enabler,” Julian Leslie, head of networks for National Grid Electricity System Operator in the U.K, said grid operators will move from dispatching generation to meet demand to dispatching load to meet generation.
“So, on those windy days, the hydrogen electrolyzers are running; people are charging their electric vehicles; they’re charging their batteries in the lofts of their houses. And on those non-windy days, they’ll be using all that stored energy,” he said. “We’ve got one supplier here in Britain that offers you a negative pricing tariff, so when it’s really windy and the demand is low, say on a Sunday afternoon, this supplier will pay its customers to consume electricity.” About 150,000 customers have signed up for the service, he said.
“This is a great resource; it’s cheap,” said James McCalley, a professor of electrical and computer engineering at Iowa State University who appeared on the panel with Leslie. “The only problem there is [is] to identify the right kinds of loads to go after. Not all loads are created equal.”
One desirable type of load, he said, is water treatment plants. “They have an inherent flexibility in when they use their energy as a result of the ability to store water,” he said. “And they’re ubiquitous. … Every region has many, many water treatment plants, so this is a very good target.”
New Products
Harvey said the future will require new products, calling on grid operators to start “radically thinking about regulation in next 10 years.”
“We may need two types of regulating resources — those that can only regulate up five minutes because they’re a battery and they’re going to run out [of energy] … and others that can keep reg up, and as we need more and more and more, they can keep going up and up and up. That’s something the ISOs, as they encounter this, need to talk to each other and do joint research,” Harvey said.
Renuka Chatterjee, MISO’s executive director of systems operations, said she sees change coming also.
“One of the biggest MISO value propositions is our large footprint. So, we could be having snowstorms up north and hurricanes in the south simultaneously. We operate that market today with 400 MW of regulation, and we don’t use most of it on most days. We operate very, very well within that 400 MW for a 125-GW peak load system.
“I keep wondering when we’ll change the algorithm,” Chatterjee said. “So, the traditional thinking … is about to change.”
Baldick said grid operators might consider a product that addresses the reduced inertia provided by renewables.
“When we’ve defined things like contingency reserves, spinning reserves … we haven’t fully represented all the physics. We’ve captured the essence of the issue, and we find a constraint that is a pretty good surrogate for capturing the issue most of the time,” he said. “Maybe we need to do something like that for inertia or maybe some other products.”
Don’t Act Without Data
Several of the speakers said MISO and other grid operators should put an emphasis on data gathering before thinking about potential solutions.
“Let’s not make choices about how we’re going to build the software or how we’re going to solve the problem too soon. I think [there should be] a lot of looking at this and learning from each other and then deciding which road is the best for everybody,” Harvey said.
He noted that CAISO is trying to calculate the demand for flexibility resulting from intermittent resources, unpredictable load and resources that fail to follow dispatch instructions.
“So, you can try to do that calculation and assign [the costs] to the resources that are creating the uncertainty and perhaps incent them to behave differently,” he said. “But it is complicated, and it’s going to be a burden on your settlement system. … So, you have to make a judgment of: Is this going to make enough difference in behavior and resource choices to improve the outcomes?”
Mark O’Malley, professor of electrical engineering at University College, Dublin, agreed. “Why don’t we experiment? Why don’t we understand it before we do it?” he said. “If you’ve got a product in a market that’s got a zero-price … then you didn’t need it. And I think some of the products thrown out there [are unnecessary]. … I think people go too far with markets sometimes.”
He said grid operators should concentrate on “real measurements after the fact, collecting large amounts of data. … Get enormous amounts of data that’s real and use that. That’s better than any model data.”
Socializing Risk
Baldick said MISO shouldn’t seek to eliminate market volatility resulting from the increased variability of renewables. “We do have to think very carefully about how to navigate who takes on those risks, who hedges them,” he said. “From my own philosophical perspective, the risks that can be borne by market participants are best left with those market participants, and we should only hope to socialize the risks that are truly ones that pose systemic risks to the operation of the system — cascading outages, as a good example. … To the extent possible, [MISO should be] making sure that market participants are seeing those risks and are provided with the right incentives to hedge them whenever its economically efficient to hedge them.”
Even the best data and the most carefully designed market tools will not eliminate the increasing uncertainty, speakers said.
“Ultimately, we need to pay more attention to the underlying statistics of renewables — and particularly wind — to really understand the answers to questions like: If we connect MISO with PJM and average out the wind across that footprint, do we significantly reduce the relative variability or not?” Baldick asked.
“It’s pretty clear to me that there are fundamental stochasticities that aren’t going to be smoothed out by larger and larger footprints until we average coastal [and] far inland wind, for example; until we average storms in the western half of the United States with weather in the eastern half. … So, while I certainly want to improve weather forecasting, it’s not going to take away from the fundamental variability of those renewables.”
“What we need to forecast is how big the variability could be,” Harvey said. “Even if we can’t forecast what load can be, can we forecast how high net load could go? That’s complicated enough.”
Forecasts also need to be matched with an understanding of transmission congestion, he added, citing CAISO’s load-shed events in August.
“It’s not just, ‘How much variability is there [and] when?’ It’s when [and] where. Because congestion is critical. A lot of the reasons why the CAISO’s ramp product has completely failed is because of locational constraints. They had enough upward ramp capability, but it was behind transmission constraints. If we’re going to use storage to balance some of this, we not only need to know how much [and] where but for how long and how much energy we’re going to need.”
