California has gigawatt-hours of grid-balancing available in homes, offices, factories and farms that can "shape" and "shift" their energy usage, and a smaller but still valuable market for fast-responding "shimmy" resources. But it has very little demand for “shed,” or conventional demand response -- unless it can capture value at the distribution-grid level.

These are the findings from a Lawrence Berkeley National Laboratory study, now going before energy regulators and the public, which envisions a ground-up reorganization of how California does demand response. This week, the Department of Energy lab will release the first round of public comments on its latest draft, giving utilities and the demand response industry a chance to weigh in on it.

Known as the Demand Response Potential Study, it’s built on terabytes of smart meter data from California’s three big investor-owned utilities, and uses more than 200,000 customer load profiles to determine just how much flexible customer load is available at different costs. In fact, LBNL has been able to disaggregate this data down to individual loads such as heating and air conditioning, water pumps and industrial processes, said Jennie Potter, one of the LBNL scientists involved in the study, in a Thursday briefing.

It then matches these models against the changes coming to California’s grid between now and 2025. These include more and more solar power at utility and distributed scale, the increasing shift of demand from midday to evening, the closure of the state’s last nuclear reactor, and the proliferation of smart thermostats, behind-the-meter batteries, electric vehicles and other future grid-enabled technologies.

This 10-year timescale allows LBNL and project partner Energy and Environmental Economics Inc. (E3) to compare very different ways of getting people to alter their energy use, Potter said. These range from the time-of-use (TOU) or peak pricing programs that eventually shape customer energy demand curves, to the opportunities available in day- and hour-ahead energy markets and fast-responding ancillary services programs.

Finally, to clarify just what it’s worth, LBNL has done away with California’s current welter of demand response options like interruptible loads and critical peak pricing, and replaced them with four conceptual frameworks for the state’s demand-side potential -- “shape, shift, shed and shimmy.” Or as the report puts it, “four flavors of demand response, each with a unique character complementing the needs of the grid.”

Breaking down values for shape, shift, shimmy and shed

In terms of timescale, shape -- demand response that reshapes customer load profiles through TOU prices and incentives, energy efficiency programs and behavioral campaigns -- takes the longest to make its effect known. But as “load-modifying DR," it shows up on the demand side of the balance sheet, which could reduce how much faster-acting demand response California will need in the future.

"Shift" is the resource that could be tapped by getting people to move their energy consumption from high-demand times to those when there is surplus of renewable generation. In California, this means dealing with the “duck curve” -- absorbing excess midday solar generation and reducing afternoon and evening ramps in peak energy consumption.

"Shimmy" is for loads that can dynamically adjust within minutes or seconds to react to short-run ramps and grid disturbances, such as through ancillary services or frequency regulation.

Finally, "shed" can “be thought of as conventional demand response,” Potter said. These are loads that can be curtailed to provide peak capacity and support the system in emergency or contingency events.  

Most of California’s demand response falls into the "shed" category today. Unfortunately, that category also has the dimmest future, according to LBNL’s system levelized value analysis, which found that “there is close to zero value created related to avoiding investment in the generation fleet” from load shedding as it’s done today.

That’s mainly because the rapid deployment of renewable generation, slow power plant retirement schedules, and energy-efficiency investments that modify the system load curve mean that there will be enough generation at peak times to meet system-wide demand.

That means “no opportunity for accounting for value from avoided investment in new capacity,” such as a new natural-gas-fired peaker plant -- the economic proxy of value in the old-school demand response world.


Part of this drop in shed value has to do with how much of it can come from shaping, through the time-of-use prices coming to California over the next five years. By 2020, every customer of a major utility in the state will be paying different prices depending on the hour of the day, and those prices can be designed to shape load curves in ways that help the state’s energy supply-demand balance.

The shape resource can be modeled as either shed or shift, depending on whether it’s geared toward reducing peak hours or getting people to move to off-peak hours. But either way, it’s a significant resource -- 1 gigawatt for the shape-as-shed, critical peak pricing-based model, and about 1.8 gigawatt-hours per day by 2025 for the shape-as-shift model that incentivizes usage at off-peak hours.

That’s a relatively small slice of 600 to 700 gigawatt-hours of average total daily load projected for California by 2025. Even so, as a marginal addition to the state’s grid balancing act, it’s worthwhile.

Shimmy resources are more of a specialty demand response product, with about 1.2 gigawatts of resource providing $43.5 million to the system by 2025, according to the report. However, “the value of advanced DR will increase over time, as the CAISO system integrates additional renewables, and curtailment becomes more significant during the midday hours,” it noted.

Seeking balance at system-wide and distributed scales

Shift resources -- loads that can be incentivized to move from an on-peak hour to an off-peak hour -- are by far the biggest source of future flexible load, accounting for up to $700 million per year, the report found. But there are significant regulatory and market barriers to hitting this value.

California’s three investor-owned utilities have about 20 gigawatts of demand response assets that could meet the shift category’s needs today, including big industrial loads, agricultural water pumps and commercial air conditioning.

Of that, only about 2.5 gigawatt-hours per day would be cost-effective at a price of about $20 per kilowatt-hour shifted, according to the report. But that number could go way up if load-shifting resources are able to capture additional benefits, such as getting paid for reducing local grid congestion in place of upgrading a substation.

“[A]ssuming that some shift DR is also optimally located in the distribution system so that it avoids building new infrastructure to handle load growth, and that some resources provide site-level benefits, the increase in shift DR that is cost-competitive is significant,” rising to more than 25 gigawatt-hours per day by 2025. 

And even traditional shed resources, which the report projects could fall to less than a gigawatt by 2025, could find some value if they’re used for distribution grid needs. “Accounting for possible service to local distribution system capacity needs can flip the potential back to a significantly large value,” adding up to 1 to 4 gigawatts of resource by 2025, the report found. 

Mark Martinez, emerging markets and technology manager at Southern California Edison, said on Thursday’s call that “when we heard that shed had little or no value, we were kind of shaking our heads -- this is demand response, this is what we do.”

At the same time, SCE is working on multiple fronts that conform to the shape, shift, and shimmy paradigms laid out by LBNL, said Martinez. On the shape side, it’s submitting a rate case proposal that includes big changes to its time-of-use hours and pricing. On the shift side of things, it’s procuring hundreds of megawatts of energy efficiency, demand response, energy storage and thermal load-shifting resources to meet both system-level and local needs. 

“Things have changed, and they’ve changed rather dramatically,” he said. “We have the internet of things, bring your own thermostats, batteries, electric vehicles. And the needs for demand response have changed as well.”

Finding the local value for these resources will also be critical, said Martinez, given the state’s push to have utilities allow distributed energy resources to stand in for distribution grid investments.

“It’s not so much a matter of trying to trim the load -- it’s about flexibility, it’s about a dynamic system,” he said. “We know there’s more value at a local level, where there may not be value at a high-level market level. But we need to know where it is.” 

Join GTM for actionable conversations on the future of electricity in our nation's most innovative state. California's Distributed Energy Future 2017 will be held in San Francisco, March 8-9. Learn more here.