shift to DC is developing some consensus as to what DC voltage level is best for distribution, given the need to keep current levels and conductor ampacity within reason. The schools of thought range from 48 to 550VDC. These voltage levels that are under consideration fall under the National Electrical Code, ANSI/NFPA 70, which defines low voltage as any voltage up to 600 V. The grounding of DC systems is also covered in the NEC in section 250‐3 and 250‐22. Once the preferred voltage level is established, more specific codes and standards will have to be developed. Codes and Standards development is in progress overseas. The European Standard ETSI EN 300 132‐3 v 1.2.1 (2003‐08) covers DC systems up to 400V.
3.11.7 Maintenance
One of the most valuable data center maintenance tools is the Power Quality Monitor. While the monitoring of AC power quality has been done for many years, doing the same thing on DC distribution systems will require different equipment. Fortunately, there are already some instruments available that are up to the task like the PSL PQube® AC/DC Power Monitor and the Dranetz. See Chapter 9 – Power Quality for additional information.
Figure 3.10 DC Monitoring Equipment
3.11.8 Education & Training
Initially, there may be some difficulty in locating personnel with the appropriate experience to install and maintain these DC systems. With some aggressive training programs, existing personnel will develop the necessary skills. This is the same training and education that is currently underway for the photovoltaic industry, where electrical distribution systems rated up to 600VDC are key ingredients. It should also be noted that there will remain a need for AC expertise, since data center HVAC systems will no doubt continue to be powered by AC, and consequently still need the conventional AC distribution.
Figure 3.11 SmartTEAM™ mobile screenshot
(Courtesy of PMC Group One, LLC.)
3.11.9 Future Vision
Looking to the future, global work on fusion reactors has demonstrated that future power generation should be DC. For long‐distance distribution, High Voltage DC systems are less expensive and suffer lower electrical losses. High‐temperature superconductors promise to revolutionize power distribution by providing near‐lossless transmission of electrical power. The development of superconductors with transition temperatures higher than the boiling point of liquid nitrogen has made the concept of superconducting power lines commercially feasible, at least for high‐load applications. It has been estimated that the waste would be halved using this method since the necessary refrigeration equipment would consume about half the power saved by the elimination of the majority of resistive losses. Some companies, such as Consolidated Edison and American Superconductor, here in the United States, have already begun commercial production of such systems. Many people are unaware that superconductors are already being used commercially to construct the ultra‐strong magnets used in magnetic resonance imaging (MRI) scanners.
All the advantages and benefits of the DC alternative are yet to be proven over time. The industry must continue the dialogue and take steps to ensure the concept is thoroughly vetted. Small‐scale demonstrations are a good first step. Pentadyne, a leader in flywheel energy storage, assembled the first bench test at their Chatsworth, CA headquarters in 2005. An industry group sponsored by the California Energy Commission through Lawrence Berkeley National Labs and headed by EPRI Solutions and Ecos Consulting is planning more demonstrations of data center applications. Commercial proof‐of‐concept demonstrations are also being conducted in Sweden and Japan.
While retrofitting existing facilities for DC may not be cost‐effective in the near term, the DC alternative should be thoroughly investigated when designing a new data center. The increase in efficiency and the associated reduction in operating costs will be the primary driving force behind a DC revolution. And the fact that DC promises higher tier ratings with lower capital costs makes improved reliability another big attraction. As with any change that upsets the status quo, the conversion to DC will no doubt meet with opposition, but the benefits seem too irresistible to pass up.
3.12 Containerized Systems Overview
When you hear the term containerized systems, modular data centers are probably the first technology to come to mind. Initially, the primary purpose of this technology was aimed at disaster recovery operations. However, with the increasing challenges of high‐density computing, including cooling, power distribution, and the continual expansion of facilities to meet customer needs, containerized systems and modular infrastructure have emerged as an innovative solution to these problems.
Several vendors offer fully containerized data centers housed in a trailer‐like enclosure, marketed as both a rapid expansion solution, and for some companies, a complete data center architecture. The units are fitted with specially designed racks and chilled water‐cooling units. Electrical power, chilled water, and network connections are all that is required to commission a complete data center. These modules can be connected to a centralized power/cooling plant, or each can be individually equipped with their own dedicated power and cooling systems. This is proving to be a quintessential “plug and play” solution because it allows companies to forecast accurately and efficiently expand their data centers to meet computing needs. Energy efficiency is reported to be higher than traditional data centers; due in part to the compact high‐density design. As a result, companies such as Google and Microsoft are finding this technology to be invaluable.
It is important to note that modular systems can include an array of technologies not solely restricted to a shipping container. It also incorporates modularized infrastructure such as prefabricated UPS rooms, and chiller and generator containers. The benefit of these prefabricated solutions is that the quality of construction of these products is higher and more controlled than on a chaotic construction site, therefore ensuring more reliable equipment. Additional benefits include easy integration to existing sites, reduced costs of construction, and portability, and flexibility of relocation.
The offsite UPS container can be assembled while the data center building enclosure is being built, bringing data center speed to market. The typical container can house both a UPS‐A and UPS‐B system with batteries and associated input and output switchboards. The UPS container can be matched with an equivalent generator “block” and a cooling module “block,” making data center expansions relatively easy. This type of construction is typically called block load design, and blocks can be added as data centers are populated and need more power and cooling for the IT rollouts. The major benefit is that it allows valuable interior areas that would have been needed for the UPS’s and utility power transformers and switchgear to be used as white space for additional IT equipment, while the UPS’s, generators and utility power are outside, adjacent the building. Data center operators have also maximized data center white space by placing containerized pump houses and chillers on the roofs of data centers.
Figure 3.12 Open rear door of containerized Data Center
(Courtesy of Sun Corporation.)
3.13
