Eric Hirst and Brendan Kirby
Competitive wholesale electric power markets today can be likened to the sound of one hand clapping. They are often inefficient and not fully competitive, in part because retail-customer loads do not participate in these markets. Although electricity costs vary substantially from hour to hour, often by a factor of 10 within a single day, most customers continue to buy electricity under time-invariant prices that are set months or years ahead of actual use. Thus, consumers are fully insulated from the volatility of wholesale electricity markets.
Click here to enlarge image
Customers should have the opportunity to see electricity prices that reflect wholesale-market price variations. Permitting customers to face the underlying variability in electricity costs can improve economic efficiency, increase reliability and reduce the environmental impacts of electricity production.
Metering, communications and computing technologies are needed to support these dynamic pricing and voluntary-load-reduction programs (Figure 1). To begin, customers must have meters that record, store and communicate data on their electricity use at the hourly or subhourly level. Second, the load-serving entity (LSE, which can include load aggregators, energy service providers, energy service companies, scheduling coordinators, the marketing affiliate of a distribution utility or a municipal distribution utility) must be able to provide price and interruption-request information to the customer in a timely, usually automated, fashion. Similarly, the customer’s interval kWh-consumption data must be transmitted periodically to the LSE. The LSE may provide additional data and analysis on the customer’s electricity use and bill-saving options. Finally, based on this data and analysis, the customer may take action, either automatically or manually, to manage electricity use and costs.
The type of program determines how quickly information must be transmitted from the LSE to the customer and vice versa. For example, if the customer is participating in a day-ahead dynamic-pricing program, the LSE needs to communicate with the customer only once a day. The LSE can post the next day’s 24 hourly prices on its Web site, and customers then can access the data at their convenience. On the other hand, if the customer is providing nonspinning reserve, the system operator may issue customer-specific control signals on a minute-by-minute basis (e.g., reduce your load by 2 MW within 10 minutes).
Click here to enlarge image
The frequency with which the customer’s electricity-use data must be transmitted to the LSE also varies. Although all these programs require that the customer’s load data be measured and stored hourly or subhourly, these data need not be reported to the LSE more often than once a month. However, if the customer is participating in a load-reduction program, the LSE might want to receive data on the load reduction as it occurs, in which case the communication from the customer’s meter to the LSE needs to be much more frequent.
If the customer participates in a program that provides current information on electricity use, ways to reduce that use and current prices, the LSE will need frequent access to the meter data so it can analyze the data and provide the customer with its analysis results. Alternatively, the customer’s electric meter could communicate directly with the customer’s computer or energy management system.
Because of the factors discussed above, it is difficult to estimate the costs of these enabling technologies. Figure 2 offers a broad range, suggesting an annual cost of $30 to $250 per residential or small commercial customer. For large commercial and industrial (C&I) customers, the range in annual cost is roughly $100 to $800.
Meters for residential and small commercial customers usually measure single-phase real energy use only. Meters for large C&I customers measure three-phase demand and energy and may also measure reactive-power consumption and power-quality characteristics. The traditional residential meter records cumulative electricity use, typically read once a month by a meter reader who walks from house to house. Advanced meters record and store electricity consumption at intervals of 5, 10, 15, 30 or 60 clock-synchronized minutes.
For residential and small commercial customers, it may be cost-effective to retrofit existing meters. These retrofits include a pulse initiator that generates an electrical pulse for every revolution of the meter disc, a data recorder that records the number of pulses, and a communication interface. These packages permit the capture of electricity-use data at 15-minute intervals with communication of that data to the LSE on a daily or weekly basis. The cost of such upgrade packages is $50 to $200.
For new installations, larger customers or more sophisticated applications, a new electronic solid-state meter may be installed. Such meters range in cost from $200 to more than $3,000. The range is so large because these meters differ in the number of channels of data they record, the amount of data that can be stored within the meter, the number of communication ports and the communication medium.
Automatic price communication is required to communicate time-varying prices. Because customers are generally price takers, price signals can be broadcast to all the applicable loads simultaneously. The LSE need not send a unique message to each consumer.
The frequency with which prices need to be communicated depends on the market. It may be acceptable to take several hours to deliver day-ahead energy prices. Real-time prices (e.g., those produced by the system operator’s 5- or 10-minute markets) must be delivered within minutes. The communication system for transmitting such real-time prices must be automated and integrated with the system operator that runs the real-time market.
Several technologies are used to communicate real-time price information, including fax, telephone, e-mail, pagers, PC-to-PC communication and the Internet. E-mail is currently the most popular choice because it is inexpensive and easy to automate. In some cases, the customer determines when it wants to obtain current price data, possibly downloading from a Web site. In other cases, the supplier initiates the communication either when prices change or on a scheduled basis, usually by e-mail or telephone. In either case, the communication is with the customer’s staff, computer or end-use equipment (e.g., a thermostat), not with the electricity meter.
Communicating price signals over the Internet (i.e., posting prices on a Web site) is inexpensive, easy and provides broad access to customers. It is much simpler for an LSE to host a Web site than to establish telephone communications with each customer in a large group.
The kWh consumption data recorded by revenue meters must be transmitted to the LSE and to the customer to guide load-control actions. For most electricity consumers, meters are read manually once a month by meter readers. Automated meter reading (AMR) is required to record and communicate electricity-use data at the hourly or subhourly level. Two technologies dominate the AMR field: telephone and fixed radio networks.
Telephone line communication with meters can be initiated either by the LSE (outbound communication) or by the meter (inbound communication). In both cases, the communication can be conducted with the same phone line the customer uses; in some cases, a separate phone line is installed, which increases the cost of this option.
For inbound communication, the meter is programmed to call a phone number at certain times. The meter modems recognize when the phone is in use and will then place their phone call later. These modems also can recognize when the telephone is picked up while they are transmitting data to the LSE; they immediately terminate their use of the phone to permit the customer to use the phone and then resume data transmission later.
Outbound communication provides more flexibility for the LSE, permitting it to poll the meters whenever it wants to. On the other hand, the LSE now has to manage numerous phone numbers and needs a ringless technology (so the customer does not hear the phone ring when the LSE calls the meter).
Phone systems are appropriate for geographically dispersed customers. Where loads are densely packed, fixed radio systems may be preferred. In such systems, each meter has a transmitter/receiver attached to it. Radio transmitters/receivers are mounted on utility poles or outside building walls, situated so they can “see” the meter-mounted radio systems. The pole- or wall-mounted systems then communicate with central transmitters/ receivers, which in turn send consumption data to the LSE’s central computer.
Because radio systems do not rely on phone lines, they can communicate with the meters at any time. Communications are faster with fixed radio networks than with telephone systems. For low volumes, readings can take less than 10 seconds per meter. Such systems are used today to serve C&I customers. For high volumes, up to 100,000 meters can be read per hour, but this increases the processing time greatly. In most high-volume applications, the systems collect and deliver a daily read file for a million or more end points by 8 a.m. next business day.
Fixed radio system costs depend on the number of customers connected to the system, the geographic customer density, the area’s topology (which affects propagation of radio waves) and meter location (indoors vs. outdoors). Fixed radio networks are lowest in cost when all (or almost all) customers in a particular area are served by the same communication system. These radio systems are typically owned and operated by third-party vendors, which sell the service to utilities on a dollar-per-meter- month basis. Costs vary from about $1 to $5/meter-month, depending on the frequency of meter reading and the amount of data transferred.
Analysis, decision and response systems
Communicating price signals from the LSE to the customer and consumption data from the customer to the LSE is, technically speaking, all that is required for customer participation in dynamic-pricing programs. But the benefits of such participation are much greater if the customer responds intelligently to these price and consumption data. Based on this information, customers might decide to:
- Do nothing and continue to consume electricity as before,
- Respond to price changes by manually or automatically adjusting consumption, or
- Invest in new equipment and controls to expand the range of possible responses.
In some cases, consumers might like to see data and analyses of their energy bills on a monthly basis. They may use this after-the-fact data to decide on operations and maintenance changes or capital investments that, on a long-term basis, will cost-effectively reduce their electricity bills. In such cases, the meter resolution is hourly, but the communications can occur infrequently.
Other customers will choose to respond closer to real time. For these customers, LSEs may provide data and analysis tools that help customers decide what to do and when to do it. Such systems can be especially helpful to customers who decide to take manual-control actions only during the few hundred highest-priced hours during the year. Systems such as Silicon Energy’s Enerscape software (and similar packages from Itron and Active Energy Management) can facilitate such decision making.
If the customer plans to respond routinely to price changes, automatic response is more practical. Johnson Controls and Honeywell market systems that integrate electricity-price data with their automated building-management systems. In general, these systems use the Internet to communicate with customers, although pagers are also used. The Honeywell system was tested at a few large facilities and demonstrated savings of 10 percent to 20 percent through automated response to dynamic prices.
Carrier Electronics developed a thermostat for residential application that, upon receipt of a control or price signal from the LSE, can adjust air-conditioning or space-heating temperature settings up or down. Puget Sound Energy conducted a small Home Comfort Control pilot program during the 1999/2000 winter season. The program featured a two-way radio-communication system from Schlumberger, programmable thermostats from Carrier Electronics, and energy management and Internet services from Silicon Energy. The program tested customer comfort and acceptance of 2 and 4 degrees Fahrenheit decreases in temperature setting for two hours at a time, periodically invoked by the utility. The pilot program demonstrated the reliable operations and integration of the technologies and very high customer satisfaction.
Apogee Interactive operates an Internet-based demand exchange that permits customers to bid load reductions to their LSE. Several utilities, including Portland General Electric, GPU Energy and the Bonneville Power Administration use the Apogee system. Apogee has about 2,800 MW of load reduction under contract, involving more than 500 customers of all types.
The Federal Energy Regulatory Commission recently noted that “lack of price responsive demand is a major impediment to the competitiveness of electricity markets.” Clearly, customer participation in bulk-power markets is necessary; otherwise these markets will not and cannot be truly competitive, and the expected benefits of competition will not be realized.
Although dynamic pricing’s potential benefits are large, so too are the barriers to widespread adoption. One barrier is the limited deployment to date of the metering, communication, computing and control technologies needed to realize these benefits. These technology barriers likely will fall as more dynamic-pricing programs are offered throughout the country and as these functions are standardized, allowing equipment and software manufacturers to harness the benefits of mass-produced systems.
Standardization can help lower the costs of advanced meters and communication systems. To realize cost reductions requires mass production, which should not be a problem, given the existence of 125 million electric meters in the United States. Such mass production, however, requires standardization of functions, communication protocols and markets so that inexpensive devices can be designed and manufactured.
Ultimately, the convergence of retail competition, wholesale competition and improved technologies should greatly expand the type and magnitude of price-responsive demand. Permitting and encouraging retail customers to respond to dynamic prices will improve economic efficiency, discipline market power, improve reliability, and reduce the need to build new generation and transmission facilities.
This article is excerpted and adapted from a report titled “Retail-Load Participation in Competitive Wholesale Electricity Markets,” originally prepared for the Edison Electric Institute.
Eric Hirst is an independent consultant who focuses on issues related to electric-industry restructuring, including the characteristics, requirements, costs, market rules and prices for ancillary services; bulk-power reliability; market structures for regional transmission organizations; and the role of price-responsive demand in wholesale markets.
Brendan Kirby is director of the Power Systems Research Program at Oak Ridge National Laboratory and a private consultant. He has 25 years of electric utility experience and has been working on restructuring since 1994.