Search

Data centers are energy intensive buildings that have grown in size and number to meet the increasing demands of a digital economy. This paper presents a bottom-up model to estimate data center electricity demand in the United States over a 20 year period and examines observed and projected electricity use trends in the context of changing data center operations. Results indicate a rapidly increasing electricity demand at the turn of the century that has significantly subsided to a nearly steady annual electricity use of about 70 billion kWh in recent years. While data center workloads continue to grow exponentially, comparable increases in electricity demand have been avoided through the adoption of key energy efficiency measures and a shift towards large cloud-based service providers. Alternative projections from the model illustrate the wide range in potential electricity that could be consumed to support data centers, with the US data center workload demand estimated for 2020 requiring a total electricity use that varies by about 135 billion kWh, depending on the adoption rate of efficiency measures during this decade. While recent improvements in data center energy efficiency have been a success, the growth of data center electricity use beyond 2020 is uncertain, as modeled trends indicate that the efficiency measures of the past may not be enough for the data center workloads of the future. The results show that successful stabilization of data center electricity will require new innovations in data center efficiency to further decouple electricity demand from the ever-growing demand for data center services.

The growing investment by governments and electric utilities in energy efficiency programs highlights the need for simple tools to help assess and explain the size of the potential resource. One technique that is commonly used in this effort is to characterize electricity savings in terms of avoided power plants, because it is easier for people to visualize a power plant than it is to understand an abstraction such as billions of kilowatt-hours. Unfortunately, there is no standardization around the characteristics of such power plants. In this letter we define parameters for a standard avoided power plant that have physical meaning and intuitive plausibility, for use in back-of-the-envelope calculations. For the prototypical plant this article settles on a 500 MW existing coal plant operating at a 70percent capacity factor with 7percent T&D losses. Displacing such a plant for one year would save 3 billion kW h per year at the meter and reduce emissions by 3 million metric tons of CO2 per year. The proposed name for this metric is the Rosenfeld, in keeping with the tradition among scientists of naming units in honor of the person most responsible for the discovery and widespread adoption of the underlying scientific principle in question--Dr. Arthur H. Rosenfeld.

Many economic modelers believe that there is an "optimal economic path" for solving the climate problem that exists independent of human choices. This belief rests on the notion that Integrated Assessment Models can determine the path that "maximizes global welfare" and, in turn, this path should drive climate policy. This commentary focuses on an under‐appreciated problem with that belief. We argue that the existence of pervasive increasing returns to scale, network externalities, learning curves, spillovers, and other nonlinear effects puts the idea of a single optimal economic path at odds with our current understanding of the most important forces driving the development of real economic and technological systems. We further argue that this idea is detrimental to rigorous understanding of climate solutions. This article is categorized under:Assessing Impacts of Climate Change > Scenario Development and ApplicationClimate Economics > Economics and Climate ChangeClimate Economics > Economics of Mitigation [ABSTRACT FROM AUTHOR]

ENERGY STAR [registered trademark] is a voluntary labeling program designed to identify and promote energy-efficient products, buildings and practices. Operated jointly by the Environmental Protection Agency (EPA) and the U.S. Department of Energy (DOE), ENERGY STAR labels exist for more than thirty products, spanning office equipment, residential heating and cooling equipment, commercial and residential lighting, home electronics, and major appliances. This report presents savings estimates for a subset of ENERGY STAR program activities, focused primarily on labeled products. We present estimates of the energy, dollar and carbon savings achieved by the program in the year 2001, what we expect in 2002, and provide savings forecasts for two market penetration scenarios for the period 2002 to 2020. The target market penetration forecast represents our best estimate of future ENERGY STAR savings. It is based on realistic market penetration goals for each of the products. We also provide a forecast under the assumption of 100 percent market penetration; that is, we assume that all purchasers buy ENERGY STAR-compliant products instead of standard efficiency products throughout the analysis period.

A recent (1996-2000) acceleration of declines in energy and carbon intensity in the U.S. remains largely unexplained. This study uses Divisia decomposition and regression to test two candidate explanations - fuel mix and weather. The Divisia method demonstrates that fuel mix does not explain the declines in carbon intensity. The fuel mix, both overall and for electricity generation, became slightly more carbon intensive over the study period (though the slight trend reversed before the end of the period). A regression-based correction to the Divisia indices, accounting for variation in heating- and cooling-degree-days, indicates that warmer weather accounts for about 30 percent of the total declines. This leaves declines of more than 2 percent per year (and an acceleration of more than 1 percent over previous decade) remaining to be explained.

Peak power demand issues have come to the fore recently because of the California electricity crisis. Uncertainties surrounding the reliability of electric power systems in restructured markets as well as security worries are the latest reasons for such concerns, but the issues surrounding peak demand are as old as the electric utility system itself. The long lead times associated with building new capacity, the lack of price response in the face of time-varying costs, the large difference between peak demand and average demand, and the necessity for real-time delivery of electricity all make the connection between system peak demand and system reliability an important driver of public policy in the electric utility sector. This exploratory option paper was written at the request of Jerry Dion at the U.S.Department of Energy (DOE). It is one of several white papers commissioned in 2002 exploring key issues of relevance to DOE. This paper explores policy-relevant issues surrounding peak demand, to help guide DOE's research efforts in this area. The findings of this paper are as follows. In the short run, DOE funding of deployment activities on peak demand can help society achieve a more economically efficient balance between investments in supply and demand-side technologies. DOE policies can promote implementation of key technologies to ameliorate peak demand, through government purchasing, technology demonstrations, and improvements in test procedures, efficiency standards, and labeling programs. In the long run, R&D is probably the most important single leverage point for DOE to influence the peak demand issue. Technologies for time-varying price response hold great potential for radically altering the way people use electricity in buildings, but are decades away from widespread use, so DOE R&D and expertise can make a real difference here.

Our research was conducted in support of the EPA ENERGY STAR Office Equipment program, whose goal is to reduce the amount of electricity consumed by office equipment in the U.S. The most energy-efficient models in each office equipment category are eligible for the ENERGY STAR label, which consumers can use to identify and select efficient products. As the efficiency of each category improves over time, the ENERGY STAR criteria need to be revised accordingly. The purpose of this study was to provide reliable data on the energy consumption of the newest personal computers and monitors that the EPA can use to evaluate revisions to current ENERGY STAR criteria as well as to improve the accuracy of ENERGY STAR program savings estimates. We report the results of measuring the power consumption and power management capabilities of a sample of new monitors and computers. These results will be used to improve estimates of program energy savings and carbon emission reductions, and to inform revisions of the ENERGY STAR criteria for these products. Our sample consists of 35 monitors and 26 computers manufactured between July 2000 and October 2001; it includes cathode ray tube (CRT) and liquid crystal display (LCD) monitors, Macintosh and Intel-architecture computers, desktop and laptop computers, and integrated computer systems, in which power consumption of the computer and monitor cannot be measured separately. For each machine we measured power consumption when off, on, and in each low-power level. We identify trends in and opportunities to reduce power consumption in new personal computers and monitors. Our results include a trend among monitor manufacturers to provide a single very low low-power level, well below the current ENERGY STAR criteria for sleep power consumption. These very low sleep power results mean that energy consumed when monitors are off or in active use has become more important in terms of contribution to the overall unit energy consumption (UEC). Current ENERGY STAR monitor and computer criteria do not specify off or on power, but our results suggest opportunities for saving energy in these modes. Also, significant differences between CRT and LCD technology, and between field-measured and manufacturer-reported power levels reveal the need for standard methods and metrics for measuring and comparing monitor power consumption.

Electronic office equipment has proliferated rapidly over the last twenty years and is projected to continue growing in the future. Efforts to reduce the growth in office equipment energy use have focused on power management to reduce power consumption of electronic devices when not being used for their primary purpose. The EPA ENERGY STAR[registered trademark] program has been instrumental in gaining widespread support for power management in office equipment, and accurate information about the energy used by office equipment in all power levels is important to improving program design and evaluation. This paper presents the results of a field study conducted during 2001 to measure the power levels of new monitors and personal computers. We measured off, on, and low-power levels in about 60 units manufactured since July 2000. The paper summarizes power data collected, explores differences within the sample (e.g., between CRT and LCD monitors), and discusses some issues that arise in metering office equipment. We also present conclusions to help improve the success of future power management programs.Our findings include a trend among monitor manufacturers to provide a single very low low-power level, and the need to standardize methods for measuring monitor on power, to more accurately estimate the annual energy consumption of office equipment, as well as actual and potential energy savings from power management.

This paper provides a general overview of electricity consumption and peak load in California, by both sector and end-use. We examine the growth in electricity demand between 1980 and 2000, as well as the composition of electricity end-uses in 1999. One of the main conclusions from this analysis is that electricity use in California in the 1990s did not grow explosively, nor was the amount of growth unanticipated. In both absolute and relative terms, growth in electricity use was greater in the 1980s than the 1990s. During the1990s, most of the growth in electricity use has been in the buildings sector, particularly commercial buildings. In 2000, the building sector accounted for 2/3 of annual electricity consumption and 3/4 of the summer peak load.

ENERGY STAR(Registered Trademark) is a voluntary labeling program designed to identify and promote energy-efficient products, buildings and practices. Operated jointly by the Environmental Protection Agency (EPA) and the U.S. Department of Energy (DOE), ENERGY STAR labels exist for more than thirty products, spanning office equipment, residential heating and cooling equipment, commercial and residential lighting, home electronics, and major appliances. This report presents savings estimates for a subset of ENERGY STAR program activities, focused primarily on labeled products. We present estimates of the energy, dollar and carbon savings achieved by the program in the year 2000, what we expect in 2001, and provide savings forecasts for two market penetration scenarios for the period 2001 to 2020. The target market penetration forecast represents our best estimate of future ENERGY STAR savings. It is based on realistic market penetration goals for each of the products. We also provide a forecast under the assumption of 100 percent market penetration; that is, we assume that all purchasers buy ENERGY STAR-compliant products instead of standard efficiency products throughout the analysis period.

This paper presents the results of 11 after-hours walk-throughs of offices in the San Francisco CA and Washington D.C. areas. The primary purpose of these walk-throughs was to collect data on turn-off rates for various types of office equipment (computers, monitors, printers, fax machines, copiers, and multifunction products). Each piece of equipment observed was recorded and its power status noted (e.g. on, off, low power). Whenever possible, we also recorded whether power management was enabled on the equipment. The floor area audited was recorded as well, which allowed us to calculate equipment densities. We found that only 44 percent of computers, 32 percent of monitors, and 25 percent of printers were turned off at night. Based on our observations we estimate success rates of 56 percent for monitor power management and 96 percent for enabling of power management on printers.

In spite of the recent explosive growth in the use of office and network equipment, there has been no recent study that estimates in detail how much electricity is consumed by that equipment in the United States. In this study, we examined energy use by office equipment and network equipment at the end of 1999. We classified office equipment into 11 types; for each type we estimated annual energy consumption for residential, commercial, and industrial use by combining estimates of stock, power requirements, usage, and saturation of power management. We also classified network equipment into six types and estimated the annual energy consumption for each type. We found that total direct power use by office and network equipment is about 74 TWh per year, which is about 2 percent of total electricity use in the U.S. When electricity used by telecommunications equipment and electronics manufacturing is included, that figure rises to 3 percent of all electricity use (Koomey 2000). More than 70 percent of the 74 TWh/year is dedicated to office equipment for commercial use. We also found that power management currently saves 23 TWh/year, and complete saturation and proper functioning of power management would achieve additional savings of 17 TWh/year. Furthermore, complete saturation of night shutdown for equipment not required to operate at night would reduce power use by an additional 7 TWh/year. Finally, we compared our current estimater with our 1995 forecast for 1999. We found that the total difference between our current estimte and the previous forecast is less than 15 percent and identified the factors that led to inaccuracies in the previous forecast. We also conducted a sensitivity analysis of the uncertainties in our current forecast and identified the data sets that have the largest impact on our current estimate of energy use.

Each year, the U.S. Department of Energy's Energy Information Administration (EIA) publishes a forecast of the domestic energy economy in the Annual Energy Outlook (AEO). During the forecast period of the AEO (currently through 2020), renewable energy technologies have typically not achieved significant growth. The contribution of renewable technologies as electric generators becomes more important, however, in scenarios analyzing greenhouse gas emissions reductions or significant technological advancements. We examined the economic assumptions about wind power used for producing forecasts with the National Energy Modeling System (NEMS) to determine their influence on the projected capacity expansion of this technology. This analysis should help illustrate to policymakers what types of issues may affect wind development, and improve the general understanding of the NEMS model itself. Figure 1 illustrates the model structure and factors relevant to wind deployment. We found that NEMS uses various cost multipliers and constraints to represent potential physical and economic limitations to growth in wind capacity, such as resource depletion, costs associated with rapid manufacturing expansion, and grid stability with high levels of capacity from intermittent resources. The model's flexibility allows the user to make alternative assumptions about the magnitude of these factors. While these assumptions have little effect on the Reference Case forecast for the 1999 edition of the AEO, they can make a dramatic difference when wind is more attractive, uch as under a carbon permit trading system. With $100/ton carbon permits, the wind capacity projection for 2020 ranges from 15 GW in the unaltered model (AEO99 Reference Case) to 168 GW in the extreme case when all the multipliers and constraints examined in this study are removed. Furthermore, if modifications are made to the model allowing inter-regional transmission of electricity, wind capacity is forecast to reach 214 GW when all limitations are removed. The figures in the upper end of these ranges are not intended to be viewed as reasonable projections, but their magnitude illustrates the importance of the parameters governing the growth of wind capacity and resource availability in forecasts using NEMS. In addition, many uncertainties exist regarding these assumptions that potentially affect the growth of wind power. We suggest several areas in which to focus future research in order to better model the potential development of this resource. Because many of the assumptions related to wind in the model are also used for other renewable technologies, these suggestions could be applied to other renewable resources as well.

In 1992 the U.S. Environmental Protection Agency (EPA) introduced ENERGY STAR (registered trademark), a voluntary labeling program designed to identify and promote energy-efficient products, in order to reduce carbon dioxide emissions. Since then EPA, now in partnership with the U.S. Department of Energy (DOE), has expanded the program to cover nearly the entire buildings sector, spanning new homes, commercial buildings, residential heating and cooling equipment, major appliances, office equipment, commercial and residential lighting, and home electronics. This paper provides a snapshot of the ENERGY STAR program in the year 2000, including a general overview of the program, its accomplishments, and the possibilities for future development. First, we describe the products that are currently eligible for the ENERGY STAR label and the program mechanisms that EPA and DOE are using to promote these products. Second, we illustrate selected milestones achieved in some markets, and ways that EPA and DOE are responding to challenges or changes in certain markets. Third, we discuss the evolving ENERGY STAR brand strategy. Next, we explore ways in which ENERGY STAR interacts with and enhances other policies, such as appliance standards and regional market transformation collaboratives. We then discuss evaluation studies that EPA and DOE are undertaking to quantify the impact of the ENERGY STAR program. Finally, we discuss future areas of expansion for the ENERGY STAR program, including labeling of new products and integrated programs for commercial and existing residential buildings.

Historically, residential energy and carbon saving efforts have targeted conventional end uses such as water heating, lighting and refrigeration. The emergence of new household appliances has transformed energy use from a few large and easily identifiable end uses into a broad array of "miscellaneous" energy services. This group of socalled miscellaneous appliances has been a major contributor to growth in electricity demand in the past two decades. We use industry shipment data, lifetimes, and wattage and usage estimates of over 90 individual products to construct a bottom-up end use model (1976-2010). The model is then used to analyze historical and forecasted growth trends, and to identify the largest individual products within the miscellaneous end use. We also use the end use model to identify and analyze policy priorities. Our forecast projects that over the period 1996 to 2010, miscellaneous consumption will increase 115 TWh, accounting for over 90 percent of future residential electricity growth. A large portion of this growth will be due to halogen torchiere lamps and consumer electronics, making these two components of miscellaneous electricity a particularly fertile area for efficiency programs. Approximately 20 percent (40 TWh) of residential miscellaneous electricity is "leaking electricity" or energy consumed by appliances when they are not performing their principal function. If the standby power of all appliances with a standby mode is reduced to one watt, the potential energy savings equal 21 TWh/yr, saving roughly $1-2 billion annually.