Technology is creating a slow revolution in the utility industry. Demand management and demand response tools are at the forefront of these changes, allowing utility customers to take charge of their electricity usage. And data analytics is at the heart of these tools.
Demand response tools are designed to help control supply and demand on the grid. They are becoming more sophisticated as renewable resources become more prominent on the transmission grid. Independent system operators will look to demand response participants to fill in the gaps when wind or solar resources wane.
To quote the California Independent System Operator (ISO), “From an operational perspective, demand response resources will contribute to the low-carbon flexible capacity needed to maintain real-time system balance and reliability, while also supporting the integration of increasing levels of renewable energy resources,” (California ISO Dec 2013).
On the other hand, demand management tools work to control supply and demand within a building. Services offered by a growing number of software companies specializing in automated building management are helping commercial and industrial companies to reduce energy costs and curtail electricity usage when kilowatt demand tariffs are extremely high or when utilities are facing emergencies.
Aiming at the Autonomous Building
Alerton provides building management solutions for heating, ventilation, and air conditioning equipment for all types and sizes of buildings. It is headquartered in Lynnwood, WA, and is part of Honeywell International Inc.’s Automation and Controls group.
Kevin Callahan, product owner and evangelist for the company’s products, owns Compass software, part of Alerton’s BACtalk Ascent product line. It is a BACnet, Ethernet-based building management system supported by desktop, mobile, and tablet browsers. BACnet is a data communication protocol for building automation and control networks.
Alerton’s other key products are Ascent Microset 4, which monitors and displays temperatures, relative humidity, fans, and CO2; and Ascent Control Module, which features multiple global controller interfaces and multiple communication networks.
Callahan has spent 15 years training software developers and customers, and continues to work with them in addition to developing close relationships with the company’s 150 local dealers—hence his evangelist title.
“It allows me to understand what customers want,” he says, like ease-of-use and the ability to find what they are looking for. “Managers don’t want workers to spend hours working on software instead of their jobs.”
Callahan’s long-term goal is an autonomous building that runs itself with minimal human interface. “With the advent of artificial intelligence and machine learning, I believe my vision will be realized in five to seven years,” he says. An autonomous building will make work more conducive and employees will be more comfortable.
As power supplies become more decentralized and wind and solar resources are becoming more popular, a building with a solar system and excess mid-day power needs to know what’s happening around it. Callahan describes a large school campus in an urban area. It has solar on all its building roofs. “I will want to move excess power across buildings. I can’t see an individual in each building managing that,” he says. A building management system would need to learn to do that, for example, by detecting an open window or lights on in an empty room detected by an occupancy sensor.
Alerton has always provided load shedding capabilities to its clients, Callahan says, in which monitors follow loads that can be shed to avoid demand charges. “But,” he cautions, “You need to understand how the building operates first. In the late 1980s, a school asked that certain circuits be designated for load shedding. Unfortunately, when those circuits were turned off, the ovens in the school cafeteria turned off, stopping food preparation. It is important to ask questions,” he says.
This is where construction and building design comes in. A net-zero energy building gets fresh air and operable windows. Callahan cites an example where control systems monitor inside and outside temperatures. To avoid windows being opened when the building is being heated or cooled, special lights were installed in hallways. When the building no longer needs to be heated, the lights turn green, telling occupants they can open windows.
It is important to study a new building, says Callahan. Let it operate for a year during the four seasons to understand where the energy is being used. Try raising heating set points and lowering cooling set points. Then determine what kinds of loads can be shed and for how long.
For example, Callahan suggests, half of the bank of elevators in an office building could be turned off between 9 a.m. and 11 a.m. and between 2 p.m. and 4 p.m. when occupants are working.
The largest costs for a business are for employees, Callahan says. An autonomous building will make work more conducive, meaning employees will be more comfortable and productive.
Roadblocks to Demand Response
For demand response to be successful, Callahan says there must be a willingness and ability to enter into an agreement with an aggregator. Older buildings may be constructed so that the cost to upgrade (or add) building management systems would be cost-prohibitive. As examples, he points to buildings in Boston or New York.
Callahan recalls that at a conference, a university building manager talked about the cost to keep up maintenance in the campus buildings: one-third were built prior to 1950 and one-half are over 100 years old. It is a challenge for them to install or improve the existing technologies.
Callahan also describes an experience a university had with a demand response contract with Pacific Gas & Electric (PG&E). The University chose to curtail after 5 p.m. when students in the dormitories were working on their laptops. If they didn’t have backup power (this was the early 2000s when laptops were more expensive), they would lose papers. Their parents complained to the University— who chose to buy generators and operate them during curtailment periods. Callahan says the University did not save any money
The moral, says Callahan, is that the problem was created by a design issue and the university should have planned better. “It’s really about planning,” he says. Furthermore, demand response is often overridden because comfort and security trumps energy savings, Callahan says. Facility engineers want to avoid complaints.
Making Virtual Meters Possible
CopperTree Analytics, headquartered in Vancouver, BC, was incorporated in 2012 after developing its software for several years.
The company’s software product, Kaizen, a cloud-based analytics energy management system, allowed the contractor, Automated Temperature Control (ATC)—which installed the system at a Nevada State prison hospital, Lakes Crossing Center, in Sparks, NV—to identify a boiler cycling on and off over 25 times in one hour. Resetting controls prevented unnecessary wear and tear on the boiler and helped avoid potential capital costs.
Keith La Rose, director of business development at CopperTree Analytics, a partner with ATC, explains that once installed, Kaizen takes data from the hospital’s building automation system, plus forecasted weather data from NOAA and Weather Underground, and stores it in the cloud using the facility’s BACnet data protocols.
Running the data through algorithms, Kaizen then predicts data points and alerts and programs the equipment to start, turn off, or delay HVAC equipment, the chiller, and boiler automatically—and most critically, to avoid peak demand, La Rose says. The chiller was programmed to precondition the hospital during the early part of the day and shut it down during peak periods. This simple adjustment reduced the hospital’s peak demand by 5 kW per day and reduced energy costs during the most expensive time of the day. La Rose predicted annual savings at the small Nevada hospital to be about $3,500 from this one initiative alone.
Kaizen is able to communicate with the building’s management system through the Copper Cube, a data logger similar to a small industrial PC, which searches the building management system to locate all BACnet trend logs and archives those trends in its internal database, providing redundant and long-term storage of the building’s information. It then sends the management system trend-log data to Kaizen in the cloud, says La Rose.
Kaizen also contains an automated fault detection and diagnostics system, says La Rose. It’s like having all the experts in the industry constantly checking a system’s operations. It will detect systems not running when they are supposed to, for example. One air handler unit might not be handling enough cooling, and a report back to facility’s staff will recommend actions to fix the problem.
Kaizen, aided by Copper Cube, also enables virtual meters throughout the building at a much lower cost than a physical meter, explains La Rose. These meters use points in the building management system, like the temperature sensors on the return and supply units, or a reading of pump status horsepower or amperage to calculate energy consumption. Kaizen is then able to create historical energy data from trend logs to use in energy management plans.
CopperTree’s system is scalable by starting small, creating a baseline of all the building management system settings—including schedules, set points, and manual mode settings for every BACnet object and associated mechanical system. It can then grow the system, eventually adding the automated fault detection and diagnostics to the building’s systems. Its graphical, Logic Builder interface customizes Kaizen’s analytics package to fit the parameters that a customer wants. Key performance indicators, metrics, and charts can be displayed on a single screen.
La Rose says CopperTree’s energy management system, where installed, is saving an average 18% of its buildings’ energy annually and maintaining those savings. At the Nevada State Library, where the Kaizen system was also installed, the chiller electrical consumption was reduced by 71%.
Reducing Demand Faster than a Peaker
Enbala Power Networks, headquartered in Vancouver, British Columbia, provides demand response and frequency regulation to PJM, the Eastern grid operator, the Independent Electric Service Operator in Ontario, Canada, and soon to be Public Service Company of New Mexico. It also has an aggregation contract with PG&E.
Enbala Power Networks’ software platform aggregates, controls, optimizes, and dispatches distributed energy in real time, 24 hours a day, seven days a week. Mark Mosher, Enbala Power Networks’ manager of Grid Services, calls the software platform, Symphony, a virtual power plant.
Mosher explains that demand response is a service where customers reduce loads when a grid operator sees extremely high demand for power and sends out emergency emails to reduce load to aggregators like Enbala, which combine customers’ pledged reductions. He says this is much more efficient than waiting for a peaker generating plant, designed to start up within eight to 10 minutes when called on. Peaker plants are an inherent part of the grid, whether they are independent or installed behind the meter at a customer site to reduce peak demand loads, he says.
Mosher adds that customer pumps that only turn off and on are designated for demand response; and those that have variable frequency drives, which can change speeds, can be used for frequency regulation.
Participating energy users may also provide frequency regulation, via aggregators, to a system operator. This need occurs when a gap between power generation and demand on the grid causes the grid frequency of 60 Hertz to move away from its nominal value. The system operator calculates the area control error with the goal of maintaining grid stability and sends the desired kilowatts or frequency regulation setpoint to the customer. The participating energy user must respond within two seconds. Currently, Enbala’s customers provide frequency regulation only to PJM.
One of Enbala’s customers that provides frequency regulation is Pennsylvania American Water. It sells water and wastewater services to approximately 2.2 million people. Its Shire Oaks Pump Station has a peak demand of 1,650 kW. The processes of pumping and treating water and wastewater contain inherent flexibility, allowing the utility to deliver grid balance to PJM, generate a new revenue stream, and help offset high electricity costs for itself and its ratepayers.
Mosher explains this pump operation moderates between 75 and 100% power depending on need. Enbala can peel off the targeted amount of power that PJM needs for regulation when the pumps operate below 100%. Enbala’s software knows which of its assets in the network, based on their constraints, are available to respond at any given time. No asset owner is ever penalized for not being able to accept a grid balance request.
What Can Clients Earn?
Mosher says there are different models for revenues. For example, Enbala reports that Pennsylvania American Water is paid 2 to 3% of the site’s energy bill based on its pump’s response to grid balance requests. Annual payments range from $35,000 to $50,000 per MW year.
Enbala’s customers sign shared performance contracts for demand response services. “Historically, we split revenues from PJM with customers,” he says. “It’s what Enbala will be using for Public Service Company of New Mexico, where we have been awarded contracts which will start in 2018.
“Right now we are working in several regions with battery storage where storage can provide load shifting. Not all regions have rules for demand response participation,” says Mosher.
Mosher agrees that it is important to gain insight and understand the capabilities of the equipment that provides the services for demand response. With a new customer, it takes about three months to plan, walk through the site, learn about and install the necessary control equipment, learn the metering requirements, commission the equipment, and do onsite testing. Commissioning is an ongoing process, Mosher says, since new equipment is installed periodically.
The Big Guy in the Field
EnerNOC describes itself as a leading provider of demand response solutions and energy intelligence software. It was started in 2001 by David Brewster, president, and a friend when both were college students. Their intent was to provide clean generation resources such as Capstone microturbines and a data analysis platform to connect customers with metering and analytical diagnoses.
EnerNOC was the first demand response supplier with ISO New England in March 2003, he says. The company now operates in 12 countries around the world and has over 6,000 MW of active demand response with 14,000 large industrial and commercial companies.
“We’re the largest demand response aggregator in the world,” says Brewster. “It’s a very exciting time for the industry.” With renewables taking prominent roles on the transmission grid, intermittency is increasing and demand response can take on the role of smoothing out power supply. Furthermore, companies are evaluating diversifying their electricity supplies, he says, adding “and with the addition of energy storage, we are well positioned” to fill those roles.
On June 25 this year, EnerNOC announced it was being acquired by the Enel Group, a multinational power utility headquartered in France. The transaction values the company at over $300 million. Brewster says it will provide EnerNOC with Enel Group’s huge platform of 30 countries in which it can grow its services.
EnerNOC’s relationships with customers go beyond demand response. Its energy intelligence software provides customers with energy matrix analytics, Brewster says. It pulls all fuel and electricity data into one place to “help the customer to know how to buy electricity, how to use it, and when to use it,” he says. The final leg is demand response and demand management, he adds.
The energy intelligence software zeros in on operational best practices by making sure buildings are operating as they should. For example, the data may show high energy use on weekends when few staff should be there. EnerNOC staff, which keeps tabs on the data, sends alerts to the building operator to adjust time clocks or sensors to turn equipment off.
EnerNOC’s platform for buying electricity for its customers is designed as a reverse declining price auction which brings sellers to bid against each other. “We have 500 different suppliers of gas and electricity,” says Brewster. “The auction creates transparency.”
Brewster says the company has collected data for the past 15 years—it streams data for each building every five minutes—and uses that data to compare electricity usage across buildings. It can then use that data to show customers how it is doing compared to peers.
“From a customer’s point of view, the economics have to make sense,” says Brewster. “Economics outweigh opportunity costs,” he adds. “Do as much automatically as possible, like data collection and setting limits. Our job is to make it simple for the customer. As aggregators, we are shielding customers from penalty risk” at the ISO. “Clarity of expectations are set up front. The big part of the work is education and clarity,” says Brewster.
Will DSOs Maximize DR Benefits?
One of the innovative concepts being discussed in the utility industry is that of a distributed system operator (DSO) where utilities create an operator for their own distribution system, modeled after the transmission system operators like Cal ISO and PJM. The DSO would dispatch and control the renewable systems in their service territory to be compatible with the grid power coming from the ISO.
Brewster supports the concept, as long as the DSO maximizes benefits for customers. As an example, he points to Consolidated Edison in New York where, under the statewide Reforming Energy Vision (REV) program, the utility is trying to create a market that benefits distribution at the wholesale level.
Brewster described the Brooklyn Queens Demand Management project. ConEdison needed to upgrade a substation in the area at an estimated cost of $1.2 billion. The New York Public Service Commission asked the utility to come up with non-traditional solutions. ConEdison produced a demand response project with energy storage behind the meter for an estimated $200 million. “A DSO should have economic incentive to do that at other sites,” he says.
Another Roadblock to DR
The major roadblock to DR is regulatory, says Brewster. It is not a technology or customer issue. “We need electricity markets that compete with traditional utility generation,” and to do that, the market rules need to be rewritten and made current, he argues.
We have been using traditional generation resources for 100 years but now that we have DER and with energy storage coming online, we need to rewrite the grid rules so they are technology agnostic,” he says. “One of the key things is you’re talking about a lot of small resources which have one second to communicate data with the ISO.”
Brewster points to the Midwest ISO where there is great potential for demand resources, but the market design has not facilitated its growth.
How DR Works at One ISO
Created in 1927, the Pennsylvania, New Jersey, and Maryland utilities realized the benefits and efficiencies of operating as a continuing power pool. In 1993, they transitioned to an independent, neutral organization as the PJM Interconnection Association. It now covers 13 Eastern states and the District of Columbia and serves 61 million people.
PJM offers multiple demand response programs where retail customers have the opportunity to participate in its energy, capacity, and other markets and receive payments for the demand reductions they make. Participants have the choice of signing up for the day-ahead option where they agree to reduce load at a specific time the next day, or they can agree to reduce loads given a signal in the real-time market the same day.
Participants are required to sign contracts with curtailment service providers (CSPs)—called aggregators at other regional transmission organizations like Enbala or EnerNOC. CSPs do the scheduling with both the participants and PJM and they must provide a minimum of 100 kW. The average amount of demand reduction supplied by CSPs varies widely, says Pete Langbein, manager of demand response operations at PJM. CSPs respond to and are paid a locational market price offer posted by PJM. CSPs then decide whether they can deliver their aggregated reduction at the minimum posted price of $50 per MWh, for example. If they answer yes, their offer will be picked up.
Another demand resource opportunity lies with the synchronized reserve, frequency regulation, and day-ahead scheduling reserves markets. Again, participants work through CSPs. A participant must reduce power within 10 minutes and hold that reduction for 30 minutes, says Langbein. For day-ahead scheduling, loads must be reduced within 30 minutes of receiving the notification and hold that reduction longer than 30 minutes.
What does a participant do to drop load for frequency regulation? They can use variable frequency drives on pumps, energy storage batteries, or a behind-the-meter generator, says Langbein. They must make a two- to eight-second response.
Participants, through CSPs, may also participate in the Reliability Pricing Model’s capacity market which operates differently than the demand response energy market. There are two separate opportunities for participants in this market: reduce load during any entire year, or reduce load during emergency conditions, says Langbein.
To participate year-round, CSPs bid into a competitive auction for their contracted customers looking to serve a market three years in advance. The winning bids set the market-clearing price for capacity based on the ability of the system to deliver electricity into the area where the additional capacity is needed. This design allows for new demand resources to compete with existing generation.
In the emergency market, retail customers will be paid when they voluntarily reduce their usage whenever PJM declares an emergency condition during summer months.
In the May 2017 auction, PJM procured 165,109 MW of capacity for the 2020/2021 delivery year. It included 2,350 MW of new gas-fired generation and 7,532 MW of demand response resources committed to year-round availability and higher performance standards. Solar, wind, and energy efficiency resources also cleared the auction.
“Demand Response and Energy Efficiency Roadmap: Maximizing Preferred Resources.” California ISO, December 2013.