|
||||||||||||
Analysis of comprehensive smart grid technology applications at 200 of the largest U.S. utilities shows potential smart grid savings of 115,145 Megawatts (MW) with avoided costs of more than $120 billion and net savings after smart grid costs of $48 billion. This study is the first to apply individual utility customer end-use hourly electric loads to evaluate smart grid costs and benefits. Data for more than 800,000 residential and commercial utility customers in the 200 largest U.S. utilities were applied in the study. These utilities represent slightly over 70 percent of all residential and commercial electricity use. (The commercial sector is defined to include commercial, institutional and government utility customers.)
Studies to date, including a recent analysis released by FERC, rely on assumptions about elasticities and electricity pricing to estimate changes in total utility hourly loads or broad customer class aggregate hourly loads.
Instead of applying the elasticity/aggregate load approach of previous studies, this study applies load control and pricing program impacts directly to individual customer end-use loads such as air conditioning, water heating and so on to determine utility-level impacts.
Results indicate that more than 10 percent of the 200 utility applications are close enough to break-even that more targeted strategies are appropriate. More than one-third of the utility applications reveal benefit cost ratios less than 1.5, suggesting careful development of smart grid strategies to ensure that economic benefits exceed costs. Detailed customer analysis shows, however, that all utilities can significantly improve returns on smart grid investments by targeting individual market segments with specific smart grid strategies.
Insights provided by "bottom-up" utility customer end-use hourly load data like that applied in this study will be an essential component in development and evaluation of smart grid deployment at individual utilities.
Smart Grids
Smart grids reflect the integration of state-of-the-art communication and control technologies applied to every aspect of the generation, transmission, distribution of electricity including remote and onsite control by utilities and utility customers of equipment in homes, commercial, institutional and government buildings and industrial processes. In its most comprehensive application, smart grid applications allow utility customers to both reduce and shift equipment electricity use to off-peak periods depending on price signals provided by the utility and as a result of customer commitments to "sell" electric loads back to the utility system.
The smart grid concept also includes traditional load control programs used by many utilities to cycle water heating and air conditioning equipment during peak demand periods. Results of several dozen utility experiments show that a combination of equipment control and price feedback technologies often doubles the impacts of load controls. Most residential and commercial programs consist of a relatively small number of participants, relative to utility populations, making it difficult to apply elasticities estimated from these programs in other utility service areas and states.
The regulated nature of the utility industry (distribution utilities are still regulated in competitive states) requires that estimated costs and benefits of smart grid programs be quantified in advance to justify rate increases required to cover the cost of implementing these systems. These evaluations are important as comprehensive smart grid technology implementation could potentially top $70 billion.
Smart Grid Analysis Methodology
The objective of the current study is to identify maximum potential peak load reductions that can be achieved for both residential and commercial utility customers on a sustained basis during peak seasons. The results of this study should be viewed as a potential that can be achieved with smart grid technologies rather than a forecast based on an assumed penetration of smart grid technology applications.
Residential customers in summer peaking utilities typically contribute most heavily to peak periods in late afternoon when occupants return home, turn up air conditioners, turn on appliances including TVs and stoves and ovens. Dwelling units with electric water heating contribute to peak electricity demand with sink/faucet use and with tub/shower uses. Electric clothes dryer and washing cycles (when accompanied by electric water heating) can also contribute to residential electricity use in peak periods.
Smart grid technologies can provide traditional load control functions, cycling air conditioning, water heating and swimming pool pumps, while more advanced options provide individual equipment control to these and additional end uses through programmable devices accessible by both the utility and the utility customer. These technologies permit households to respond to a price signal to lower thermostat setting and schedule major electric appliances. Our analysis applies load controls and scheduling of major electric appliances in households to smooth out loads from these end uses during peak periods.
Customer behavior is also enabled with smart grid technologies either through programmed responses or simply as a response to information on real-time prices. Related behavior includes reduction of summer thermostat settings, use of cold water in washing and air drying for clothes, use of microwave rather than stovetop burners, reduced shower/tub water temperatures, turning off lights and unwatched TVs and other actions. A variety of studies indicate that the behavioral responses are approximately equal to behavioral load shape reductions achieved with load control and scheduling activities. This approach effectively distinguishes between different customer behavioral responses as a function of electric appliance holdings with smaller behavioral impacts for dwelling units with natural gas water heating, cooking, and so on.
The fact that smart grid impacts are achieved at the individual customer level means that smart grid impact analysis can be provided for any customer dimension or combinations of customer dimensions in our databases. For instance, results can be presented by income, dwelling unit type and size, demographic characteristics, business type and size and other factors. This information can be important in developing smart grid strategies and evaluating smart grid program achievements.
Compared to residential customers, commercial, institutional and government utility customers reflect considerably less potential for smart grid peak demand savings. Increased summer thermostat settings and pre-cooling provide the greatest demand reduction opportunity. Additional savings can be achieved by turning lights off or through lighting adjustments accomplished with energy management and control systems. Some reduction can be achieved with office equipment shutdown, limitations on elevator use and other behavioral responses.
This analysis applies a cycling load control strategy to commercial, institutional and government air conditioning systems to smooth out loads and a reduction of 15 percent to reflect increased thermostat settings and some pre-cooling. This estimate is consistent with the limited information on commercial customer pre-cooling and a change in building thermostat setting to 80 degrees. A 10 percent reduction in lighting and other loads is assumed. Reduced waste heat from the 10 percent reduction also cycles back as an additional reduction in air conditioning loads of approximately five percent.
Commercial buildings with energy management and control systems (ECMS) have considerably more demand management potential than described above; however, since these systems are still in relatively limited use and reflect a considerable investment to install in existing buildings, the impact of smart grid EMCS interactions is not addressed in this study.
Utility Benefits and Costs
One objective of this study was to determine the economic feasibility of instituting comprehensive smart grid initiatives at the customer level, that is, deploying advanced smart grid technologies to all utility customers. Benefits are calculated as avoided costs of new generation, transmission and distribution. This analysis used the most recent Department of Energy combustion plant cost of $670/kW. The same figures for transmission and distribution avoided costs used in the Rand study were applied here ($125/kW and $250/kW respectively) for total avoided cost benefits of $1045/kW. While $250/kW is likely to be a reasonable average cost of distribution across the U.S. (though the cost estimates are now somewhat low since the data were developed a decade ago), distribution costs vary widely across utilities. For instance, the distribution cost for Con Edison from the original study is more than $1,500/kW. Consequently, the average $250/kW is applied here to all utilities to reflect average utility costs. Since costs vary around all of these averages for individual utilities, only the total cost results are reported here.
Costs of comprehensive meter programs are assumed to be $500 for each residential and small commercial customer, $5,000 for medium commercial customers (between 20 and 200 peak kW) and $20,000 for large commercial customers. The benefit-cost results change only slightly when medium and large commercial customer costs are modified reflecting the fact that medium and large commercial customers are not more than several percent of total residential and commercial customers in most utilities. These costs include utility costs of developing the infrastructure required to support real time pricing, load control and other programs along with equipment and installation costs.
Utility analysis shows that nearly all utilities will save enough in avoided costs with a comprehensive smart grid deployment to at least cover smart grid development and deployment costs; however, many utilities barely break even. For example, about 10 percent of utilities achieve benefit/cost ratios less than 1.2 and should be considered at risk of incurring costs that are greater than benefits given uncertainties concerning actual deployment costs. More than one-third reflect benefit cost ratios less than 1.5. Customer-detailed analysis shows, however, that all utilities can significantly improve returns on smart grid investments by targeting individual market segments with specific technologies.
Strategies to Support Utility Smart Grid Program Development and Evaluation
This study illustrates the intuitive and straightforward application of end-use (air conditioning, water heating, etc.) utility customer hourly load databases to evaluate smart grid program development.
Utility customer database detail supports direct representations of actual load control and smart grid technologies to individual end-use hourly loads providing a true bottom up approach to program development and evaluation. The alternative is to apply elasticities to customer class hourly loads and load profiles to drive a top-down aggregate representation; however, this approach provides little insight on program development and evaluation issues and is of questionable accuracy when applied in individual utility applications.
This study illustrates the process of evaluating costs and benefits of smart grid initiatives. In addition to applying customer and end-use detailed representative hourly load databases, information on generation, transmission and distribution avoided costs, technology costs relevant for the utility size and deployment and utility infrastructure support costs must be developed. These data vary significantly by utility and even by geographic location within the utility service area.
By focusing on end-use hourly electricity use within individual utility customer buildings, the utility customer database approach provides a consistent accounting of impacts of specific technologies and their impacts at the customer, customer segment, customer class and utility level. This information provides a basis for developing a resource planning strategy for utility smart grid deployment that insures economic benefits will exceed program costs.
Utility Customer Database information can be used to forecast benefits and costs over time using penetration assumptions and forecast utility customer growth. All standard economic cost tests (Participant Test, Ratepayer Impact Measure Test, Utility Cost Test, Total Resource Cost Test and Society Cost Test) can be applied to the results of a Utility Customer Hourly Load based analysis.
This framework also permits extensive scenario analysis, risk analysis and customer segment evaluations where segments can be defined by any variable in the database (income, business type, etc.).
Finally, a Utility Customer Hourly Database analysis framework provides an easily maintained system that can be used continuously to evaluate ongoing smart grid activities and to adjust strategies to maximize utility smart grid benefits.
References
Baer, Walter S., Fulton, Brent . Mahnovski, Sergej, "Estimating the Benefits of the GridWise Initiative," Phaase I Report, Rand Science and Technology Technical Report Prepared for the Pacific Northwest National Laboratory, May, 2004
Faruqui, Ahmad and Sanem Sergici, "Household response to dynamic pricing of electricity: A survey of the experimental evidence," January 10, 2009.
Federal Energy Regulatory Commission, "A National Assessment of Demand Response Potential," June, 2009.