Andy Starkey from data centre design, construction and maintenance specialists, TKR, discusses how a holistic approach lies at the heart of power and cooling efficiency

Technology advances quickly.  Year on year we see developments that promise better performance, greater efficiency and reduced power consumption.  Indeed, these promises are often perfectly valid and capital expenditure on newer, more efficient cooling and power systems may not only reduce energy consumption and optimise efficiency, it could also help to future-proof a data centre’s capabilities.  But cooling and power management is about much more than ensuring an installation has the latest bit of kit.  Any investment in new cooling technology or power equipment must be balanced against the returns it can offer and sometimes an intelligent, holistic and contextual approach to managing cooling and power can yield dramatic results – with or without a system upgrade.

There can be no doubt that newer UPS systems have significantly improved efficiency across a wider range of operating loads, reducing a data centre’s energy consumption and the amount of heat generated.  However, it is still as important as ever to consider efficiency across the whole load range when specifying a UPS, as the common practice of considering only the full load efficiency as the key parameter still does not give a true picture of the data centre’s energy profile.  If required, it is also important to consider high efficiency power transformers – again taking into account the full load range - and as the type and extent of losses vary with load.

Meanwhile, cooling systems have also advanced considerably to include latest generation CRAC (Computer Room Air Conditioning) units, free cooling chillers and adiabatic coolers to name but a few.  Notwithstanding these advancements, there is a wide array of common sense improvements that can be made to any data centre to maximise cooling efficiency.   Shorter and unobstructed air paths require less power and it is advisable to fit variable speed fans to CRAC units to enable fan speeds to be adjusted to meet the actual rather than the maximum requirement as this approach can result in considerable energy savings.  Configuring the room layout to maximise efficiency can also pay dividends: something as simple as arranging equipment in a hot aisle/cool aisle configuration with suitably positioned CRAC units can help to eliminate  ‘hot spots’ and reduce power consumption whilst maintaining the required environmental conditions.

Indeed, it is advisable to commission a thorough engineering assessment from a specialist like TKR to diagnose any issues with power and cooling efficiency so that these can be rectified. Many data centres, for example, have multiple CRAC units that actually work against each other – the antithesis of energy-efficient cooling!  This may be due to a BMS system that is not sufficiently ‘intelligent’, resulting in units that are being instructed to do one thing by the BMS and another by their integral controls!  By upgrading to a BMS system like that developed by TKR, which provides continual reporting / trending information with built in failsafe control, data centres can become much more efficient.

Other data centres fall victim to more basic housekeeping issues such as obstacles preventing the movement of cooled air within the floor plenums, which, in many cases, could be easily rectified by simply removing redundant cables.  Similarly, sealing holes to ensure that cooled air is retained where it is needed and installing a suitable number of floor grilles in appropriate locations will all help to ensure that the cooling capabilities of the air conditioning system are being used effectively, reducing energy wastage.

One of the greatest challenges of managing the balance between the need to protect business critical equipment and the drive to reduce power consumption is reconciling current and potential requirements.  Whilst future-proofing is an essential element of managing costs, it is also often the root cause of over-sizing, which can increase both capital outlay and energy consumption.  The answer to this dilemma is to use modular, scalable power and cooling systems that meet today’s requirements without limiting future options or wasting resources on just-in-case engineering.  In just the same way, data centres should be managed at the temperature actually required to safeguard the installation rather than allowing that just-in-case culture to dictate the required temperature range. Raising the temperature by just one degree can significantly reduce power consumption and in most cases this can be done safely whilst still allowing ample tolerance.

Certainly, new generation power and cooling systems can provide efficiencies but they will not deliver their maximum unless they are part of a holistic approach to power and cooling management.  Only by properly assessing cooling loads, airflows and temperature and taking a 3600 approach to managing power and cooling effectiveness can data centres truly be run efficiently.