When considering cooling a data centre, current and future power, cooling and energy demands that IT equipment will place on current or prospective data centre infrastructure need to be factored in for more efficient power and cooling decision-making. Consequently, a knowledge of best practices for architecting a customised and balanced plan to keep power levels of your IT infrastructure within the power constraints of your data centre is crucial.
There are several ways of cooling a data centre, ranging from hot-aisle and cold-aisle layout, populating the rack correctly, hot-aisle containment, and efficient power cable management. A study by Emerson Network Power indicates that a 10 degree F increase in return air temperature can typically result in a 30-38 percent increase in cooling unit capacity. A hot-aisle and cold-aisle rack arrangement, containing the cold aisle to prevent mixing of air, and implementing perimeter cooling systems supported by row and rack cooling are all able to improve cooling unit performance by reducing mixing of hot air with cold air and so enabling higher return air temperatures. Implementing an intelligent control system for understanding, predicting and adjusting cooling capacity and airflow based on conditions within the data centre is another possibility. Intelligent controls facilitate individual cooling units to work in harness as well as supporting more accurate air flow control based on server inlet and return air temperatures.
Another best practice is implementing state-of-the-art cooling designs that reduce energy consumption. Improving fan efficiency is a possibility – on chilled water cooling units, for example, fans are the largest consumer of energy. Variable frequency drives, that enable a fan’s speed to be adjusted based on operating conditions, can represent a significant improvement over fixed-speed fans. Electronically commutated (EC) plug fans are inherently more efficient than more traditional centrifugal fans since they eliminate belt losses. Data centre infrastructure management and monitoring to improve capacity, efficiency and availability should also be enabled. For example, installing a network of temperature sensors across the data centre can be a valuable addition to the supply and return air temperature data supplied by the cooling units. By sensing temperature at different locations, airflow and cooling capacity can be more precisely controlled to ensure a more efficient data centre operation.
With today’s rapid advances in computing power, data centre facilities are now expected to deliver increased computing capacity whilst at the same time improving efficiency. More traditional technologies are now being combined with newer technologies to help support these increases in capacity and efficiency in the data centre cooling system. Raising the return air temperature is one way of achieving this, along with intelligent controls and high-efficiency components – such as variable speed/EC plug fans, microchannel condenser coils and proper economizers –that allow airflow and cooling capacity to be matched with dynamic IT loads.
http://www.datacentrecooling.net/