Environmental Energy Technologies Division News

Environmental Energy Technologies Division News
  • EETD News Home
  • Back Issues
  • Subscribe to EETD News
  • Print

Energy Efficient Data Center Retrofit — Wireless Sensor Network

Data-Center Rack Doors

Passive rear door heat exchanger devices help optimize energy efficiency in a data center at Lawrence Berkeley National Laboratory.

Over the decades, the energy efficiency technologies and processes researched and developed at Lawrence Berkeley National Laboratory (Berkeley Lab) have supported energy-efficient technologies throughout the world. From time to time, however, Berkeley Lab has the opportunity to apply its solutions right on its own grounds.

In 2007, Berkeley Lab's 40-year-old data center had HVAC problems and was running out of cooling capacity. In response, it instigated a project to address those challenges while showcasing energy-saving data center technologies. The project was a partnership of the Lab's IT Division and Environmental Energy Technologies Division (EETD). Berkeley Lab engineers worked with Taylor Engineering, ANCIS Corporation, and SynapSense to reconfigure the data center's HVAC system and install a wireless sensor network (WSN) to collect real-time data and enable operators to visualize system performance. The WSN provides real-time feedback so that they can adjust temperature and humidity optimally, to minimize energy use.

In its initial facility review, Taylor Engineering found inefficiencies with airflow and cooling. Although the center had more than enough cooling capacity to meet server cooling needs (a 20-ton cooling system and water-cooled computer room air conditioner [CRAC] units for underfloor cooling), the cooling was unbalanced, and there were a number of hot spots throughout the facility. To get the most out of the existing cooling systems, and to help rebalance the airflow, ANCIS Corporation and SynapSense were brought in.

Photo of Person with Rack

The opened back of a server rack with rear door heat exchanger devices installed.

ANCIS conducted computational fluid dynamics (CFD) modeling to better understand the airflow patterns and to evaluate efficient airflow before reconfiguring the conditioning systems. Wireless sensors provided by SynapSense were used to help tune the existing system. The CFD modeling and wireless sensors helped engineers design and reconfigure the system more quickly and at less cost, as well as help ensure that the resulting system configurations would perform more effectively and more efficiently than its predecessor. As a result of these efforts, the center was able to increase its IT load, greatly reduce the number of hot spots, and do so with fewer CRAC units than when the project started.

In 2009, to help mitigate the heat from an increasing load, data center rack cooling was improved with rear-door heat exchangers (RDHx). These units were added to cool the hot air coming from the servers before it was released into the data center. By fall of that year, it was clear that all of the changes at the center had really paid off. The modifications enabled the total IT load to reach 500 kilowatts (kW)—more than 30 percent higher than it had been when Taylor Engineering, ANCIS Corporation, and SynapSense had first assessed the facility.

Data-Center Rack Diagram

Wireless sensors provide the data to help operators maximize data center energy efficiency.

However, because Berkeley Lab's need for research computers is constantly growing, it once again asked Taylor Engineering to see if it could determine a way to modify the cooling system so that it could accommodate an additional 207kW server load.

The end result is a data center that can now support the increased server load at a higher efficiency, with better temperature control and no increase in installed cooling equipment. The WSN data collection and data visualization systems continue to be used to optimize the facility's operation. Payback for the WSN modifications is estimated to be 1.5 years.

This overall retrofit process also identified some key strategies that can be used by other data centers interested in making their facilities more efficient:

  • Gather air temperatures and develop trend information with an energy monitoring and control system (EMCS), building automation system (BAS), or other monitoring system before installing additional cooling.
  • Meter and monitor equipment to provide energy use data to support design and optimize performance.
  • Optimize airflow management before installing additional cooling.

—Mark Wilson

For more information, contact:

  • Dale Sartor
  • (510) 486-5988
  • Geoffrey Bell
  • (510) 486-4626

Data Center Rack Cooling with Rear-door Heat Exchanger
(Technology Case-Study Bulletin, June 2010) [PDF]

Wireless Sensors Improve Data Center Energy Efficiency
(Technology Case-Study Bulletin, September 2010) [PDF]

Data Center Airflow Management Retrofit
(Technology Case-Study Bulletin, September 2010) [PDF]

This research is funded by the California Energy Commission's Public Interest Energy Research program.

↑ home | ← previous article | next article →