Data centers’ intricate design: Sustainability
Tim Chadwick, PE, LEED AP, President, AlfaTech Consulting Engineers, San Jose, Calif.
Robert C. Eichelman, PE, LEED AP, ATD, DCEP, Technical Director, EYP Architecture & Engineering, Albany, N.Y.
Barton Hogge, PE, ATD, LEED AP, Principal, Affiliated Engineers Inc., Chapel Hill, N.C.
Bill Kosik, PE, CEM, LEED AP, BEMP, Building Energy Technologist, Chicago
Keith Lane, PE, RCDD, NTS, RTPM, LC, LEED AP BD+C, President/Chief Engineer, Lane Coburn & Associates LLC, Seattle
Robert Sty, PE, SCPM, LEDC AP, Principal, Technologies Studio Leader, SmithGroupJJR, Phoenix
Debra Vieira, PE, ATD, LEEP AP, Senior Electrical Engineer, CH2M, Portland, Ore.
CSE: Energy efficiency and sustainability are often a request from building owners and CIOs. What net zero energy and/or high-performance systems have you recently specified on a data center building?
Chadwick: Economization strategies that take advantage of ambient air or ambient wet-bulb temperatures are critical to an efficient data center strategy. Another significant efficiency-improvement tool is to take advantage of redundant HVAC equipment. By operating all fans and pumps and other cooling equipment at reduced speeds (as opposed to operating a portion of the equipment at full speed), you can take advantage of fan-affinity laws. With these benefits, significant reduction in energy consumption can be achieved. We have also completed several recent projects where renewable energy sources have been acquired by the tenants to offset the power demands of the facility.
Eichelman: The strategy to reach net zero energy first requires that the IT equipment and the supporting infrastructure use as little energy as possible. High-efficiency UPS and distribution topologies that result in high-load factors, such as distributed and block-redundant systems, will drive down energy use. Economizers, high-temperature chilled water, containment systems, and containerized data center solutions will also help to increase efficiency. Efficiency gains in the IT system, such as through virtualization, offer the biggest opportunity to reach net zero, as these will not only reduce the power requirements of the IT equipment, but of the supporting infrastructure as well. To reach net zero, the remaining power requirement must be offset by renewable energy sources. These can consist of solar photovoltaics (PV), wind, or fuel cell systems, although these can be costly.
Sty: The NREL ESIF HPC data center used direct water-cooled cabinet technology to perform at an extremely low PUE ratio (PUE< 1.06). Beyond the specialized IT cabinets, it’s important to note that the bulk of the data center cooling-system infrastructure was designed without specialized equipment or components-standard cooling towers, pumps, and piping that we use in everyday design. It was a function of higher operating temperatures, being in the right climate zone, and innovative thinking on the part of our client. Highly efficient and effective solutions do not always require complex equipment.
CSE: Many aspects of sustainability (power, HVAC, etc.) require the IT team to follow certain practices to be effective. What, if anything, can an engineer do to help increase chances of success in this area?
Hogge: Effective sustainable data center solutions require an education in the IT hardware, comparing start-up and running amp loads, as well as acceptable thermal environments for the equipment. This can be challenging if the infrastructure is being designed without the specific network architecture plan and hardware specifications. Historically, the easy response is to be conservative, using boilerplate specifications and requesting power to accommodate the maximum starting electric demand, and keeping inlet air at as cold of a temperature as can be attained. This is no longer reasonable as the community moves toward a more energy-conscious mentality. The engineer can assist by promoting IT peer networking with other clients and equipment vendors, as well as hands-on metering of existing installations if possible, to provide education that will validate smaller electrical equipment sizing and higher air temperatures.
Sty: Keep the systems and strategies simple and straightforward. Complexity is the enemy and will increase the chances of human error, which is the biggest cause of data center downtime. There is a quote by our favorite fictional engineer from the Starship Enterprise in "Star Trek III." After Scotty disabled a rival, more sophisticated ship by pulling one part, he said, "The more complicated the plumbing, the easier it is to stop up the drain!"
Kosik: Designing a novel cooling solution for a data center is an iterative process where the facilities engineering team works closely with the IT equipment planner and computer manufacturer, among others. The planning of the data center IT equipment layout will be influenced by the technical and economic realities of the performance of the cooling and electrical systems. As an example, if the customer is looking to a computer approach that results in very high-density servers, the cooling engineer will need to be at the table early in the process to ensure the appropriate design solution is put in place. This approach becomes increasingly important in a retrofit situation, where there are often many more constraints.
Eichelman: Data center design should first and foremost be responsive to the nature of the IT equipment in terms of what is possible in the short and long term. Many data centers are designed with a high degree of flexibility. Typical approaches include a raised floor, which can provide a pressurized plenum for cooling air; a chilled-water piping loop (including a high-temperature loop) in the computer room, which could be run under the raised floor to support connections to computer room air handlers, in-row coolers, and water-cooled IT cabinets; and placement of overhead cable-tray and busway systems in a coordinated manner that facilitates cabinet or aisle-level containment systems.
Chadwick: Through proper definition of the design strategies and training of operations personnel, the benefits of maintaining containment and maintaining system operations can be shown. In addition, measurement and verification and retro-commissioning activities can restore original design intent and remind operators of the importance of maintaining the systems in peak operating conditions. In particular, when you can demonstrate the benefits of maintenance by showing improvements in reliability, you can gain future buy-in from operations staff whose main priorities may be system uptime.
CSE: What types of renewable energy systems have you recently specified to provide power for a data center project?
Vieira: A recent project in the Middle East had a series of PV panels mounted on top of a parking structure, which was integrated into the data center power system via an inverter. The PV system was projected to generate 250 kW of usable solar power. The system did not incorporate batteries to store unused power. The design intent used the solar power to offset overall power consumption from the utility, rather than designating the solar power for a specific IT or HVAC load. This approach allowed the facility to use all the power generated and alleviated concerns of poor load performance due to the cyclical nature of solar power generation.
Chadwick: We have completed some projects with onsite fuel cell and solar (PV) electrical-generation systems. We have also completed a recent project where, as part of the LEED certification of the site, offsite wind power was funded to offset the local power consumption.
Eichelman: Recently, I’ve been involved in data center projects that have used solar PV. I’m acquainted with other projects that have used biogas generators and fuel-cell technologies.