A promising new clean energy policy may not be as effective as promised, a group of researchers contends.

The clean peak standard jumped from policy proposal to law of the land in Massachusetts in just a couple of years. It compels electricity retailers to buy credits for clean power that can be delivered during the hours of peak grid demand. Properly timed renewables and demand response count; since wind and solar cannot dispatch on command, though, energy storage stands to play a major role.

But Massachusetts’ Clean Peak policy will do little to reduce carbon emissions from energy storage plants, according to a study published earlier this year. The results challenge, or at least complicate, the role of this policy in shaping a cleaner grid in that state and elsewhere. (For a deeper dive into the theory and implications of the critique, look for this week’s Storage Plus column on GTM Squared.)

The typical explanation of the idea behind a clean peak standard goes like this: Wind and solar are great, but the grid still relies on dirtier fuels to supply the hours of peak demand, which are usually driven by cooling loads in the height of summer or heating needs in the heart of winter. A clean peak incentive should shift clean power from other times of day into those crucial high-demand hours.

But the study, by researchers at Columbia University, New York University and the nonprofit WattTime, challenges the notion that there are clear windows of low-emission generation that can be arbitraged to high-emission peak hours.

Instead, they say, the relevant metric for measuring a battery plant’s greenhouse gas profile is marginal operating rate of emissions: Look at the power plant that the grid operator would dispatch to fill the battery with power, rather than the average cleanliness of the grid at the time.

“The emission consequences of your action depend on the emission rate of whatever generation is reacting to your behavior,” said co-author Burcin Unel, energy policy director at NYU’s Institute for Policy Integrity.

Clean Peak standard vs. carbon pricing 

When Unel and company analyzed marginal grid emissions, for a period spanning March 2018 through February 2019, they found that the New England grid often lacked clear differentials between low emission hours and high emission peak hours. If a battery taps a gas plant to charge up and discharges later to displace another gas plant, losing the typical 15 percent of electricity to roundtrip power conversion, then it’s actually emitting more carbon than if it did nothing.

Since gas generators set the marginal rate of emissions much of the time within the study period, the price signal from Massachusetts’ Clean Peak policy only reduced a simulated battery plant’s emissions by 5 percent.

In contrast, the researchers modeled a $50-per-ton carbon price, equivalent to the social cost of carbon, and found it reduced the battery’s emissions by 65 percent.

This points to a disjoint between the real-time fluctuations of the wholesale market and the rigid nature of a four-hour peak window designated by a state agency.

“Marginal emissions rates are highly dynamic, so when you have a static policy like [Massachusetts’] Clean Peak, they’re unlikely to capture the right incentives for emissions reductions,” Unel said.

A carbon price, by contrast, achieves that dynamism by pricing emissions into the least-cost dispatch algorithm. A battery loading up on cheap power and dispatching during periods of high power prices would be using lower-carbon power to displace higher-carbon power.

That’s a compelling argument in the context of an academic model. In the real world, Massachusetts lawmakers managed to pass the Clean Peak rule, but even commendations from scores of economists have yet to carry a carbon price through the legislature.

Beyond batteries 

Short-term battery emissions are one metric of the success of Massachusetts’ Clean Peak, but they’re not the whole story. For one thing, the rule applies to resources besides energy storage, but it has another mission, too. Gov. Charlie Baker originally proposed it to tackle the disproportionate costs to utility customers from just a few peak hours: as of 2018, the top 10 percent of hours drove 40 percent of the energy costs for customers in Massachusetts. 

The conventional utility approach to serving those high-demand hours would be to build more gas peakers, even if they hardly ever run. Nudging batteries to fill that role as well as perform other services throughout the year could save a lot of money in the long run. And if those batteries knock some of the oldest and dirtiest peaker plants out of business, they could improve the marginal rate of emissions over time, even if they produce more net emissions in the early years.

Speaking of time, the study offers a snapshot from 2018 and 2019, but Massachusetts is building gigawatts’ worth of new solar and offshore wind in the coming years. A robust solar incentive is likely to encourage the growth of the duck curve that’s already appearing in the regional grid — that’s the pattern when midday solar production pushes down net load, which can prompt curtailment. The addition of clean resources will affect the marginal dispatch of power plants and could well change the outcomes compared to the time the study looked at.

The critique of Massachusetts’ Clean Peak policy acknowledges that it won’t do harm: A battery plant will emit less with a Clean Peak price signal than without it. But the authors offer some suggestions to amp up the possible emissions reductions from such a rule.

One is to align Clean Peak credits with the hours of peak marginal emissions, rather than overall system demand. These are distinct metrics, though they often overlap. In Massachusetts, the strongest on-peak emissions spike occurs in the winter, when heating puts pressure on the natural-gas supply. In other seasons, a day’s emissions peak regularly falls outside of the peak demand window identified by the policy. It’s also crucial to revisit those designated peak hours as the grid evolves to ensure their continued accuracy.