Villanova University’s College of Engineering is part of a National Science Foundation (NSF) Industry/University Cooperative Research Center (I/UCRC) in Energy-Smart Electronic Systems (ES2), partnering with Binghamton University and the University of Texas at Arlington to develop methodologies, tools and systems to maximize energy efficiency for the design and operation of data centers.
A high growth and energy-intensive sector of the U.S. economy, data centers are responsible for 2.8% of total U.S. energy consumption. Worldwide, they use about 30 billion watts of electricity, roughly equivalent to the output of 30 nuclear power plants.
As consumers of such large amounts of electricity and energy, the data center industry is starting to pay more attention to the carbon footprint associated with data centers. The industry has adopted a metric called the carbon usage effectiveness, or CUE, that dictates the amount of greenhouse gas emissions generated in a data center based on the electricity used for the computing equipment.
Villanova’s ES2 research group is currently exploring the balance of economics, location, and environmental burden on data centers to develop strategies for minimizing the carbon usage effectiveness (CUE) as much as possible. Their end goal is a set of recommendations that would provide economic and environmental benefits for the industry, potentially reducing the carbon footprint of data centers nationally by a significant margin.
Data centers consume large amounts of power, with large data centers using several megawatts of electricity. This large electric demand calls for the development of new electric power plants near where new large data centers are proposed. Some in the data center industry have even called for dedicated power plants to be located next to data centers since this solution would eliminate the losses experienced as electricity travels through the grid. The location chosen for new large data center construction then would dictate an associated required growth of electric power generation, meaning that various regional policies and available resources would dictate the type of new power plants that would be developed.
These new power plants could have an influence on the carbon footprint of the data center, since the carbon footprint of the data center would essentially be equivalent to that associated with the approach used for additional power generation in the municipality. For example, designing a large data center in a region that relies primarily on coal-based electric generation would result in a larger carbon footprint than a region that has a power generation portfolio largely based in nuclear energy and renewables. This assumes that the data center owners choose to supply electricity in the most economically advantageous way to the data center, as opposed to paying more for electricity from renewable energy sources. Fortunately, for many data center applications the location does not have a critical impact on performance, meaning that there is much flexibility in where the data center could be built.
“Data center owners are starting to look at CUE as a factor in data center design to mitigate any negative public perception of new data center construction,” said Dr. Aaron Wemhoff, Site Director of Villanova’s ES2 Research Center. “If data center owners can show that the CUE for their proposed new data center is lower than the industry average, it may garner public support for the construction.”
The CUE is difficult to calculate in general, Wemhoff noted, so the most straightforward way is to estimate the CUE based on the anticipated data center power needs and to look at the municipal electric generation portfolio, which available for each state by the federal government’s Energy Information Administration. According to this database, Washington, South Dakota, and Idaho provide the lowest carbon energy production in the nation.
The Villanova research team is also examining the economic factors which play a role in the strategy used to lower a data center’s CUE. The two most straightforward ways to do so, Wemhoff says, are by locating the data center in a low-carbon municipality or by purchasing more expensive power from zero-carbon sources. Another way to lower a CUE is to provide on-site alternative energy sources such as solar photovoltaics or wind energy production, and a fourth approach is to reduce the overall power needs by the data center through a variety of cooling technologies such as taking advantage of a cold climate to remove the heat from the data center. Villanova researchers believe that a combination of these methods can significantly reduce CUE without causing large economic burden.
For more on the Center for Energy-Smart Electronic Systems (ES2), visit: