Why airtightness is key in planning data centre success
Thu 10 Aug 2017 | Calum Maclean
Calum Maclean, senior research engineer at BSRIA, discusses data centre airtightness and energy efficiency
Airtightness is sometimes a forgotten hero of energy efficiency and because of the high energy demands in data centres, any savings will be magnified significantly. The airtightness of many components and the ensemble is important in data centres or stand-alone racks.
Firstly, the reliability of the equipment and power supplies is paramount in the uptime of a system. Any risk of overheating or contamination in the system can affect equipment lifetime and consequently the uptime of the data centre.
Secondly, the target is to have the capacity required with enough redundancy to cover the highest server demands. There is also the requirement of achieving cost efficiency through either energy efficiency or better utilisation of equipment allowing smaller data centres or a combination of both.
Both operators and users also require ease of use and accessibility of the data centres for updates, upgrades or data usage.
The requirements of specific data centres are vastly different and depend on the application of the data centre. For example, a data centre for a bank, used mainly for financial transactions, needs uptime for the bank and the account holders, a capacity covering the busiest periods and cost efficiency at the forefront of the data centre use.
Whereas an office’s data server storage, outsourced, requires cost efficiency for the operator, ease of access and capacity for the user and uptime for both.
Improving and controlling the air distribution and movement in any data centre type will help achieve energy efficiency, and, therefore, cost effectiveness. Any air movement outside of the designed air paths (leakage) will affect the distribution of the cooled air and may also introduce contaminated air into the data centre.
The air paths are different for the different applications, from CRAC-cooled, raised-floor data centres to individual water-cooled racks with internal fans passing air through a water/glycol cooled coil. Airtightness is vital in maintaining the energy efficiency of any building. Any cost savings can accumulate into a significant saving in all of the examples above, however, in the case of a search operator the savings are magnified by the number of data centres utilised.
Airtightness can be applied in a wide range of situations in data centres and data centre equipment, including whole building, floor voids, contained aisles, container, standalone racks and even some individual components (CRAC units etc).
Significant savings can be made by planning the airtightness from the start and committing to following the design throughout the build
Whole building airtightness is important so that refrigerated or conditioned air inside cannot escape but also so that untreated outside air cannot get in. The effect of wind on buildings can be dramatic in a building with poor airtightness and lead to significant airflow path changes internally and possible cooling inconsistencies.
Floor void leakage can have a significant impact on the distribution of cold air. Floor tiles should be sealed prior to and during installation and the air paths identified during design and minimised during construction. Cold air from the void can leak into the cold aisles, or other chilled areas prior to the servers, which may be useful in terms of cooling, but cooling will be inefficiently distributed. On the other hand, if the cold air leaks into the hot aisles or return air paths, this will lead to cooling losses and inefficiency.
When contained aisle construction is utilised, the airtightness of the containing structure is paramount in delivering the chilled air onto the racks removing the hot air in cold/hot containment, respectively. Any loss in the pressure differences due to leakage can affect the overall air distribution in the racks, putting equipment at risk of overheating or causing rack temperature to be non-uniform and/or outside the recommended limits.
The systems utilising container style and stand-alone racks should incorporate airtightness criteria in their design in order to minimise the ingress and egress of air into and out of the systems. The container systems are at particular risk as they are able to be placed outside and, therefore, the wind effects can be significant as with the whole building setup.
The elephant in the room is can you eliminate air leakage entirely? The simple answer is not easily or cost effectively, however, significant savings can be made by planning the airtightness from the start and committing to following the design throughout the build. Any costs associated with extra design and build of a more airtight envelope can be made up by the energy savings achieved during the use of a data centre.
Airtightness methodology can be applied to a full range of building types, ensembles and equipment (e.g. floor void or contained cold aisles). The basic principle is to pressurise (or depressurise) a building and measure the air flow required to achieve that pressure differential. The amount of air introduced is effectively the air leakage at the test pressure.
The ATTMA TSL2 standard for testing buildings and structures is the generally accepted methodology for airtightness quantification. As well as the airtightness tests, prior to the installation of any data centre equipment, a smoke test can also be conducted to look visually for any large leakage paths. A smoke test is easy to conduct and can show any major discrepancies between design and build.
Airtightness tests can be applied to data centre components, such as floor void or contained cold aisles, with slight modifications from the ATTMA TSL2 standard, for example, BSRIA has recently published Floor Plenum Airtightness – Guidance and Testing Methodology (BG65/2016) or DW/144 Sheet Metal Ductwork (BESA, 2016).
Reducing the risk of equipment degradation and maximising the cooling capacity by good air distribution will ultimately improve both uptime and cost efficiency
Any standard test can allow for the comparison of different designs or manufacturers equipment prior to or after install in a building. Manufacturers of standalone equipment or containment systems can run these tests on any number of configurations so a proven level of airtightness can be achieved every time on site.
The design and build phases of any new data centre, or even the upgrade of an existing one presents the ideal opportunity for investigating and minimising the air leakage paths from any data centre type.
Low air leakage levels in different data centre applications can minimise the amount of active cooling being lost to the surrounding spaces or into a hot aisle, reducing the amount of cooling required and associated costs.
Improving airtightness will also reduce the ingress of untreated, potentially contaminated air (e.g. dust particles) which could cause system failures.
Lower air leakage levels will further improve the air distribution in the data centre, providing the right temperature and relative humidity conditions, therefore, reducing the risk of degradation of the equipment. Reducing the risk of equipment degradation and maximising the cooling capacity by good air distribution will ultimately improve both uptime and cost efficiency.
Airtightness should be thought about for any project and any piece of equipment throughout the design and build phases, to plan for success.