A lot has been done to improve cooling efficiency in data centers today, but facilities still lag far behind their claimed efficiency. Why is that?
Building a data center with all of the proper facilities is just the first step towards an operationally efficient data center. But maintaining a data center at maximum efficiency in the long-run simply relies on the quality of the maintenance performed. Proper maintenance improves the performance and longevity of the data center computer room air conditioning (CRAC) unit, which ultimately improves the cooling.
With 2015 underway, plan regular preventive maintenance with this checklist as a guide to access potential problems affecting your data center cooling environment.
Here is the essential checklist for a data center cooling system audit:
- CRAC capacity check: Make sure that the current/planned power load for your server room is within the bearable capacity of the cooling system. Remember that 1 Watt of IT power needs 1 Watt of cooling. Oversized cooling systems may lead to undesirable energy consumption. Find the cooling system capacity in the model nomenclature on or inside each CRAC unit. Capacity values are provided in the manufacturer technical data.
- CRAC unit check: Demand fighting (i.e., uncoordinated CRAC units operating in opposing modes like dehumidifying and humidifying) results in 20-30% efficiency reduction of the CRAC system, leading to insufficient cooling capacity and downtimes. CRAC units works in four modes: cooling, heating, humidification and dehumidification. Make sure that when two of these conditions (say, cooling and dehumidification) work, they are in coordination within a defined area of 4-5 adjacent units. Check and reset (if necessary) the check points for humidity and temperature to make it consistent on all CRAC units to avoid demand fighting. Ensure the fans, alarms and filters are clean and functioning normally.
- CRAC unit placement: CRAC units are placed perpendicular to the aisle on either a hot or cold aisle based on the architecture of air distribution, and they are placed at the end of hot aisles if raised floor is used for air distribution.
- Test cooling circuits, condenser circuit, and refrigerant piping: Ensure that chillers, condenser circuits, primary cooling loops, refrigerant piping, pumping systems, and evaporator coils are in proper condition. Check if all valves are working fine and that DX systems (if used) are fully charged.
- Check aisle and rack temperatures: Measure room temperatures at strategic positions in the aisle, generally centered between equipment rows and (1.5 meters) above the floor and compare with the recommended inlet temperatures of the IT equipment manufacturer. Rack temperature measuring points are at the center of air intakes at top, middle and bottom of each rack.
- Arrangement of floor & aisle tiles and air velocity from floor tiles: Uniform air velocity should be maintained across all floor grilles and perforated tiles if using a raised floor as cooling plenum. Make sure no dirt or dust below raised floors are blown into IT equipment through vented floor tiles.
- Airflow between racks: Gaps, unused rack spaces, unsealed cable openings, and excess cabling harm the cooling. Seal the gaps and openings properly.
- Check for any clogged air filters, damaged blower belts, bearings wheels and motors, leaking compressors and make sure manufacturer recommended oil type and quantity is used.
Data center cooling infrastructure audits not only provide positive impact in terms of efficiency, but they also provide insight into the present state and help plan for alternative technologies and approaches. Instead of trying to do this yourself, consider outsourcing your data center. Download our "8 Reasons Companies are Outsourcing the Primary Data Center" to learn more: