Demand Response: Experts and Businesses Discuss the Challenges Ahead
California's big three utilities have 2,660 MW of demand response enrolled, ahead of the regulators' policy goal for next summer. Enrollment is not the same as participation. Do companies have enough price elasticity to participate? If so, is demand response cost-effective for utilities? Can participating companies recover the cost of the necessary equipment? Who pays the price?
Blogged live from a meeting entitled "Demand Response: Simple Solutions, Real Savings, Fast Payback," this article grew as the meeting proceeded. It's still in its rough form, but nonetheless full of interesting information and perspectives. I invited readers to post questions using the comments form at the bottom of this article, and I posed some of them to the speakers and panelists.
March 16, 2007
Background: California is aggressive about using demand response to reduce the need to construct new power plants. The state plans to expand demand response resources in the summer of 2007.
DR is often called the low-hanging fruit of conservation, so it's easiest in territories that are not already energy efficient. California is one of the most energy-efficient states in the United States, so it makes a statement by pursuing DR in a big way. If CA can get benefits from DR, then so can any other state. Hundreds of U.S. utilities offer DR programs today, but most do not.
The state's utilities are under a "loading order" from the utilities commission to give preference to efficiency and conservation over new construction. DR is one way of achieving conservation.
Companies that are enrolled in DR programs can shed a certain amount of load, and those megawatts are considered resources by the utilities. PG&E, SCE and SDG&E have enrolled participating companies representing a total of 2,660 MW. The California Public Utilities Commission has set a DR enrollment goal of five percent of the anticipated peak load, which would be about 2,300 MW. So these utilities are collectively ahead of their regulators.
Promoting demand response is the impetus behind this meeting, "Demand Response: Simple Solutions, Real Savings, Fast Payback." The organizers, the Silicon Valley Leadership Group, invited me to attend this event in San Jose CA and specifically to blog about the presentations. So this is a different format than most EP articles and blog entries. Post your questions using the comments form below, and I'll try to get them answered.
Introductory remarksJustin Bradley, Director of Energy Programs, Silicon Valley Leadership Group, introduced the first speaker.
Anders Axelsson, vice president of sales and marketing, Echelon Corporation
"Smart metering" at first meant automated meter reading. Today, smart metering enables connecting and controlling devices of all kinds. Once the devices are independently communicating with each other, you need applications to manage them and achieve efficiencies.
Efficiencies and emissions: 70 percent of a building's energy use is for lighting. Reducing it by 10 percent reduces emissions by an amount comparable to that of 26 million cars.
Chris King, CSO, eMeter Corporation, and Chair, Demand Response Subcommittee, Silicon Valley Leadership Group
King showed a load duration curve for CA, a familiar graph to much of the audience, and used it to explain how DR reduces the need to build capacity to meed a peak load that lasts only seconds. CA could avoid building 10 to 15 percent of the necessary power plants -- or avoid rolling blackouts, also familiar here.
There are as many different DR programs as the imagination will allow. In GA, 80 percent of businesses voluntarily participate in a price response program. In a Canada residential pilot, customers get graphs on their bills that show daily electricity use and the rates and rebates. This kind of information gives people the ability to conserve about 10 percent of their energy use.
King hopes today will include a discussion involving the audience, which appears to be about 90 people in an interesting cross-section of customers and service providers.
Live Demonstration: Web Enabled Building Management and Automation SystemMark Kendall, President, Kenmark Real Estate Group
Kendall decided not to try a live demo in his presentation. A tour scheduled for later might include a demo.
When Echelon selected Kenmark to build the company's HQ (where this meeting is taking place), Echelon asked Kenmark to install the control systems. Kendall had been fascinated with control technologies for a few years, so his company took on the challenge. His company built a control center, where tenant, owner, and partner communications systems are aggregated on a single platform.
Lawrence Berkeley National Lab contacted Kendall about their control center and systems. Could DR signals be transmitted from PG&E to Kenmark's control system, and initiate a curtailment? The first tries were rough -- full stop, full start -- but it worked.
"I'm not a tech guru, I'm a builder," Kendall said. "Embracing this kind of tech is frightening, but it's doable. If I can do it, anybody can do it. And it works."
Kenmark can now take a signal from PG&E and "instantly" shed 30 percent of the building's load. Lighting and HVAC are adjusted, varying the amount of load reduction for each section of the building (i.e., less HVAC reduction, and more light reduction, on the south side).
When the brownouts started, about four years ago, a tenant company president contacted Kendall and asked if there wasn't something that could be done.
"Isn't there somebody you can call?" Kendall recounted. "Now I can say, yes, we are doing something."
Q: How did he handle the complaints from building occupants?
Kendall: Groveling, begging for forgiveness, and reminding them how much they're saving.
Leveraging Demand Response Revenue into ReinvestmentsRandy Haines, Energy Manager, Thomas Jefferson University
Background: Randy Haines manages DR at a large hospital and campus in Philadelphia, termed by industry folks as "PJM territory." His facility is four million square feet, 12,000 employees, peak load in excess of 20 MW, and sensitive electrical loads in the hospital. Payments for successful load reductions have reached up to 15 percent of his monthly power bill, which translated into $150,000 last year. He claims there is no impact on patients, students, or staff.
His facility got involved in DR to make money and control costs. They get better insight into when they peak and why. The full-blown BAS and metering system now gets utilized more fully.
The local utility signed TJU up for a program, but never gave him the tools to use it, and called it only twice.
When it got involved with aggregator ECI, the university was attracted by the simplicity of enrolling and participating. A bid screen shows real-time hourly prices and lets TJU decide whether to participate. The Siemens BAS reduces fan speeds and resets air temp in certain buildings. The university's peak is late morning, so their peak is declining as the peak pay rates are coming up. An Encelium system reduces lighting levels with dimming ballasts.
More on the systems: 50 data loggers, 50 real meters, 50 virtual meters, sub-meter groups for buildings and the entire health facility. Everything has an IP address. A web server gives occupants access to the data that pertains to their buildings.
The monitoring application makes it easy for Haines to see the daily results and report to ECI. It takes only a few minutes of his time to read the color-coded display and make the necessary decisions.
"Curtailments don't affect students or patients," Haines said. "When we adjust comfort, people don't even notice."
In 2006, TJU received a check for more than $150,000, about five percent of the participating load power bill. TJU is ready to sign up for Synchronized Reserve, where the facility's backup generators are dispatchable by the utility (which sounds similar to the idea behind Celerity, an innovative company acquired by EnerNOC in May 2006).
Q: Did he run into resistance, and how did he overcome it?
Haines: Didn't run into any resistance when he first advocated participating in DR, partly because he wasn't risking anything, and partly because he already had the BAS and controls systems in place. No additional capital investment was required, and nothing was at risk.
Panel Discussion: Customer Strategies That WorkDirk van Ulden, Utility and Energy Management, Sustainable Development, Cushman & Wakefield (Symantec, Fireman's Fund, and other clients)
Wayne Wiebe, Vice President, Kenmark Real Estate Group and Echelon's campus property manager.
Eric Giles, Gilead
Mukesh Khattar, Oracle
Moderator: Chris King
King: What is your program, why did you decide to participate, and what were your decision criteria?
Wiebe: PG&E funded Kenmark's participation in a test by LBNL. Then Kenmark signed up for a PG&E DR rate structure.
Eric: Gilead was involved in a critical peak pricing program, where the utility can curtail on 12 days. PG&E approached his company when he was installing Siemens systems in the labs. Labs were easier to curtail than the office buildings. Last year the company saved about $30,000.
van Ulden: When savings didn't convince participants, the threat of brownouts did.
Khattar: Reliability is the primary motivator. Plus the company is made up of tech geeks who wanted to play with the systems. Cisco agreed to do a voluntary test, together with 10 other local companies. PG&E reported 10 percent reduction of loads, around 10 MW. Cisco also participated in "Blackout Busters," a now-silent group that it formed.
King: What challenges did you face in getting started?
Khattar: Not very difficult, because reliability was so important to Cisco. Participating in automated DR and CPP programs was not as easy, because the company needed to invest about half a million dollars to upgrade its pneumatic controls systems and get granular control of lights. Demand tended to peak when each curtailment was over; it's as important how you come out of a curtailment to avoid a spike.
van Ulden: There's a mountain of paperwork to sign up with PG&E. The facility management people, who get the hot and cold calls, also resisted. He got employees and the CEO behind the program, to reduce the calls. You have to have passion, and sell the ideas internally. You need to talk to the business unit managers and get them on board. When employees are on board, and know there will be a peak the next day, they're generally willing to do what they can.
Giles: Scientists don't mind the occasional curtailments, where labs go from 70F to 74F. Fume hoods can't be scaled back, for occupant health and safety reasons. The savings helped him to get buy-in. In one building with 2,000 occupants, by just turning off some lights, closing shades, and raising the spaces to 74F, they can easily meet the CPP curve and no one really complains.
van Ulden: When I was at Sun, there was no interest in this. It was all about the money. At Fireman's it's about being green. The costs are paid for by the marketing department.
Khattar: A commitment from managers is not the same as a commitment from facility managers. Energy management is just one of their responsibilities. A fully automated building might not perform as well as one where the managers are committed to demand response.
Q: What about reducing supply air temp?
Khattar: The major savings come from fan speed and power, rather than temp. He has VFDs. It depends on the HVAC system.
Q: What about reducing demand on continuing operations?
van Ulden: The CEO at Sun asked why you wouldn't do this every day, if it's so beneficial. The CPP compensations are based on the past ten days' load, so it doesn't lend itself to continuous curtailment.
Q: Do any of you have cogen, and have you considered using it?
A: None of the panelists has on-site generation or thermal storage.
Q: I'm told it's not necessary to keep my data center as cold as the data center manager thinks. How do I approach the DC manager about this?
A: There's an ASHRAE standard. The problem isn't the temp, but the air flow management. Show the DC manager that by adhering to the standards for air flow, they can reduce the need for cooling.
Q: Have you tried pre-cooling?
Khattar: We pre-cooled a building by four degrees when we had a day-ahead notice. Our VAV boxes start reducing the air flow; we need to come up with a mechanism to override that during the pre-cooling process. It's quite expensive, requiring the vendor to write custom code.
Q: What DR programs would you like to see that aren't available?
Khattar: Weather adjustments for DR. Programs are structured to compare a baseline, but the weather in some regions (particularly coastal) can make that rolling average quite low. Then you have to curtail not five percent, but ten percent or more, to comply.
King: The CPUC is looking at this.
van Ulden: The support from the utility could be better. The metering systems are inadequate. At Sun and Fireman's we had to put in submeters at our own cost.
Giles: While the CPP event is going on, you don't have live information, so you don't know where you are. We put in our own meter at the main. I can go to Interact and read the meter, then go tweak the systems. That's added cost for us to monitor our power, to help PG&E meet their demand.
Wiebe: We can't count on reimbursement from PG&E, so we can't put hard numbers on a spreadsheet that we show to tenants.
Demand Response, Summer 2007 and Beyond
Steve McCarty, Director of Demand Side Resources, PG&E
Sean Gallagher, Director of Energy Division, California Public Utilities Commission
Jim Detmers, VP of Operations, California ISO
Moderator: Ken Abreu, Principal Regulatory Analyst, PG&E
The panel got off to a rough start, and finally discovered the audio system's power was off. The DR jokes were flying, while people fiddled with the amplifier.
Gallagher: We need to educate Californians about time-sensitive rates. (The CAISO load curve made another appearance on the big screen.) CA is paying $50 to $100 million a year for power plants that run less than a hundred hours to meet peak demand.
The system benefits of DR are market price impacts, lower long-term investments, greater reliability, and environmental benefits. From the customer side, there are financial benefits -- bill savings, or direct payments -- and increased reliability for all customers.
IOUs offer economic DR programs -- e.g., critical peak pricing, where price is used to get participation -- and emergency day-of DR programs -- e.g., interruptible tariffs, where the utility can call with 15 to 30 minutes' notice. Interruptible programs have largely been closed recently, as the strategy is to move more customers to the economic DR programs.
The CPUC approved measures to increase demand response following the heat wave of July 2006. Some of the measures applied immediately that summer; others directed utilities to propose how they would augment their existing DR programs in summer 2007. All these changes will give the ISO another 270 MW of DR for the coming summer of 2007. Another 225 MW will be added to the AC Cycling Program. As of January 2007, somewhere between 343 and 1,050 MW are enrolled in economic DR programs (the goal is 2,500 MW), and 1,600 MW are enrolled in emergency DR programs (the goal is zero).
Interruptible tariffs offer more economic benefit with less chance of being interrupted. Large customers oppose "default participation" inherent in economic tariffs. Many are not convinced that the costs are offset by the benefit potential. There may be too many programs to choose from.
How to increase interest and enthusiasm for price-sensitive tariffs? Increase the incentives, make less frequent modifications to the programs, and make technology accessible.
Utilities and the ISO don't always communicate well, their expectations don't align, and their timing doesn't sync. Price-responsive DR is not a predictable number of megawatts. Disconnects could lead to double-procurement of supply-side resources.
"What is the bang that we get from each buck we spend on DR?" Gallagher asked. One factor is the load impact, which can be quite variable. Another factor is the methodology for determining cost-effectiveness. (The lack of a standard valuation methodology was highlighted clearly in the DOE's report to Congress in response to EPACT05 Section 1252.)
Detmers: Talked about what happened in CA in 2006. The next year, after China had a massive blackout, its energy officials wanted to meet with the ISO. They said they needed to learn from people who knew how to prevent blackouts (laughter). In particular, given China's forecast for load growth, the officials wanted to know how California keeps its load at such a low level -- not just for peaks, but throughout the year.
"Demand response is happening, it isn't just a passing phase," Detmers said. "The problem isn't the peak [another appearance of the load curve], it's managing it. Today 20,000 MW of generating capacity is sitting idle and won't come online until summer. That's why your bills are so high. We keep all this generation to meet peaks a few times a year."
The old way is to focus on having enough generation on the system to meet demand. The new way is to manage loads with programs like DR.
"Demand grows 1,000 MW a year. Let's fix that problem. You're the right people to make that happen."
"Do we have confidence in DR? Yes, we have confidence, and I'm counting on it for this coming summer."
McCarty: The asset utilization of power generation, if found in any other industry, would be considered intolerable. Imagine a trucking company with as many idle assets most of the year. Correcting this is just one of the benefits of DR.
PG&E wants to help companies implement technologies and participate in DR. DR is second in the state's loading order. PG&E has its own programs, and is adding five third-party aggregator programs.
Automated DR will be a critical part of shaving the peak. PG&E has a $30 million budget for technical assistance focused on ADR. They'll do detailed audits of facilities, at their expense, or give a $100/kW incentive to participate. Cash incentives for installation of hardware and software that provide DR capabilities: up to $250/kW of demonstrated load reduction capability. ADR deployment is another $50/kW incentive.
PG&E programs include DR, AC Cycling (directly controlled by the utility), and permanent peak-load shifts. They're installing eight million interval meters on small commercial and residential accounts to support price-sensitive tariffs.
Q: Can PV be used for DR?
Gallagher: No, PV is net metered, it's a different program with a different set of benefits. (There were other questions throughout the day about renewable energy, and none of them initiated any serious discussion.)
Q: Is customer choice the right way to go?
Detmers: I believe customers can make the right choice. This industry can change, can put in the right corrective measures. We have a framework to make customer choice work. Large energy users have been putting pressure on regulators and legislators to open that up. The timeframe is 2011 or 2012. I'd like to see something sooner. The capability to reduce your demand is a resource, just like a generator. Do we not have that today?
Gallagher: We need to see if there's anything we can do under the current framework. Aggregators are a step in that direction.
Detmers: "The ISO wants to be there to give you the information you need. We're going under a market redesign and pumping out information about what it takes to get this job done. It's not just the ISO that makes that happen, it's everyone in this room."
Gallagher: "We are running this case at the PUC, we want to get smarter about these programs, so let us hear from you how to make these programs work for you."
McCarty: "The rest of the world is looking to us. If the people in this room can't do it, no one can."
Demand Response – Water Heaters to HybridsJon Wellinghoff, Commissioner, FERC
Background: The Federal Energy Regulatory Commission regulates the interstate transmission of natural gas, oil, and electricity. FERC also regulates natural gas and hydropower projects. Wellinghoff was recommended by U.S. Senator Harry Reid and nominated by President Bush. He was sworn into office on July 31, 2006. Wellinghoff was a partner with one of Nevada’s largest law firms, where he concentrated his practice in the fields of energy law and utility regulation.
"You can't treat a demand response participant the same way you treat an 800 MW power plant," Wellinghoff began. "And to do so I think would be undue discrimination."
Under current rules, the utility industry must consider demand-side resources the same as supply-side resources. Federal law now has language about smart metering.
"CA is way ahead and should be commended for what they're doing to make sure advanced meters are available to all customers," Wellinghoff said. Section 1241 says FERC can encourage generation, and 1223 says it can encourage advanced transmission technologies, such as load control, distributed generation and storage.
EPAct 05 made demand response the law -- Congress called for time-based pricing, technology deployment, and removal of barriers. FERC has assessed these matters and issued a comprehensive report available at ferc.gov.
The grid is the world's most complex machine, with 4,000 power plants, 935 GW of generation, 351,000 miles of transmission, and 21,688 substations. DR can play a huge role in making it more efficient.
"It's not simply the role of reducing prices, it's a role of providing an array of other services to the grid that are provided by generators, like VAR support and spinning reserves. There's a way that DR can provide those services and make the grid much more reliable than it is today."
DR can help avoid blackouts. The Northeast blackout took nine seconds to span multiple states. Generators weren't able to respond fast enough to prevent it. DR provides better data, more frequent readings, and allows fast responses that could stop a blackout and save billions of dollars.
Grid-friendly appliances take the capabilities of commercial buildings and bring them down to the residential level. PNL is working on controllers that respond to frequency, shedding their load without the need for communications. (I saw a prototype clothes dryer that did this, at a GridWise meeting about two years ago.) Homes on the Olympic Peninsula in Washington are doing this as part of a pilot.
$10 billion goes into spinning reserves each year. That's money that could go into DR. We need to treat DR as a permanent solution and develop tools for putting DR into our overall system.
Wellinghoff presented the idea of the "cashback hybrid:" Combine a plug-in hybrid petrol-electric vehicle with intelligence that can recharge from the grid and supply power to the grid. The smart plug has a communications chip that provides the intelligence. The car decides whether to buy or sell power. (I heard about this concept at last year's Greenbuild conference, where some NREL scientists presented the idea in great detail.)
"The grid benefits are as large as the transportation benefits. You have all these grid benefits -- peak demand planning, ancillary services, dispatchable reactive power -- and you can store wind, and you can use it as an emergency power supply"
"You have the Holy Grail of the energy industry: storage. If you can store power, you can do tremendous things you couldn't do otherwise."
PNL determined that we could take 75 percent of light vehicles and turn them into cashback hybrids, and we would not have to expand our generating capacity at all. It would also displace half our oil imports. Consumers would pay less per kW for electricity than they pay for the equivalent in gasoline now.
Benefits of the cashback hybrid, from the utility's perspective: Greenhouse emissions would be reduced by a quarter, and the remaining smokestack emissions from power plants would be easier to clean up. The investment is the car owner's responsibility. Consumers would buy more electricity, to charge their cars. The fixed cost of generation is distributed over more volume.
Barriers: Need to disconnect retail pricing and wholesale markets; change retail and wholesale rules; and end utility disincentives for DR.
FERC plans to hold workshops; it needs information on barriers to DR and how it can break down those barriers. That input will come from having more DR participants in those workshops and other stakeholder processes. All the stakeholders today are generators and grid operators, so the input is one-sided. FERC needs input from end users.
Q: If users saw real-time true costs of the energy they use ... why isn't it available?
Wellinghoff: It's available on a limited basis in some areas. Ideally the pricing should be all the way down to the substation level. The next step will be the interface that allows utilities, ISO and PUCs to set those prices.
Q: Will DR drive growth in the distributed generation sector, either renewable or otherwise?
Wellinghoff: Not so much in the renewables sector, I think PV is taking off on its own. DR will be one of the big drivers for non-renewable DG. DG has other benefits, such as security, reliability, and efficiency related to line losses. We need other drivers to get more DG. Ever-escalating prices will drive DG as well. If we can get the tariffs in place, it will help a lot. A national RPS needs to be a portfolio standard that includes efficiency and DR, where utilities would have standard levels they would have to meet on all of those.
Rapid, Reliable Demand Response: What Aggregators Need to Grow Their DR Portfolios
Larry Owens, Assistant Director of Electric Utility Energy Distribution, Silicon Valley Power Santa Clara
Claude Godin, EnergyICT
Alan Gartner, Vice President, Western Region, EnergyConnect, Inc.
Rich Counihan, Senior Director, Corporate Development ( Western U.S.), EnerNOC
Moderator: Mary Ann Piette, Demand Response Research Center
Background: The California Energy Commission's Public Interest Energy Research (PIER) program launched the Demand Response Research Center led by the DOE's Lawrence Berkeley National Laboratory. LBNL is hosting the Center and leading the research. A significant proportion of the research is done by researchers outside of LBNL.
Gartner: "Aggregators don't just reduce your energy bill. We write you a check." Aggregators pay a better split than the utility-direct programs, and can tailor payments to special circumstances. They also support participants through engineering and advocacy.
What differentiates aggregators? Regional versus national footprints. Ability to provide technology, such as metering and controls, to avoid doing DR manually. Sharing vs. passing on the risks of the programs to you. Amount paid, i.e., the 80/20 standard used by utilities isn't a hard and fast rule.
Godin: "The Europe market is much more progressive than the US, and much more advanced. The DR we're doing here, they've been doing for some time now. They're looking at it from a financial perspective."
Counihan: "Wellinghoff reminded me of Will Rogers, who said he wasn't a member of any organized political party, he was a Democrat." EnerNOC doesn't represent a particular power provider, it's independent. Like the other aggregators here, EnerNOC plays a role between the participant and the load-serving entity.
What makes a good DR program? Ability for aggregators to participate; stable programs for three to five years at a time; capacity payments to customers to participate whether they are called or not, to help companies make a business case for investment; penalties for non-performance, so utilities can count on DR for planning purposes; automation to remove human intervention and give real-time visibility.
Owens: "We represent the part of Silicon Valley that's on all the time." The utility has low power rates by CA standards, and has relatively low DR resources enrolled -- about 8 MW of economic and 10 MW of emergency DR, out of 1,000 MW of local system capacity. It is piloting communications technologies, including backhaul and wireless technologies for meter communications.
Q: Are there high switching costs to go from a utility to an aggregator, or the converse? Or is the equipment open enough and compatible enough to reduce that cost?
Gartner & Counihan: It depends. One year with an aggregator is generally enough time to pay for the metering and other investments. Utility programs are not as sophisticated as aggregators' programs, so the technology travels better from utilities to aggregators than in the other direction.
With that, the meeting concluded. Thanks to those of you who sent questions by e-mail or through the comments. SVLG says the presentations from today will be posted at www.svlg.net.