Data centres Energy efficiency
>
Average number of hours per annum that ambient air temperature is below 13C = 7 x 30 x 24 = 5,040 hours Approximate compressor power consumption for 1000kW cooling load = 400kW Unit cost of electricity = ¤0.11/kWh Annual cost saving = 400kW x 5,040hrs x ¤0.11/ kWh = approx ¤221,760
Chilled water storage The provision of CHW storage is not necessarily the only way to ensure that adequate cooling is always available. Multiple diesel generators, UPS systems and chillers configured to give adequate system redundancy and resilience will go some way to allay fears of complete system failure. CHW storage does, however, offer the client a guaranteed source of cooling supply at all times in the event of mains and/ or generator failure. It’s always there, ready to be used in the event of a power outage. Its provision is justified, particularly in cases where no generator and chiller redundancy is provided. In the event of a complete power outage (mains and
Server-rack densities are unlikely to exceed 35kW in the foreseeable future – which means that in-row cooling units are likely to be the most effective solution
generator), the primary CHW generation system and plant will cease to operate. In this scenario, a properly sized UPS system must be designed to provide secondary side-cooling delivery to the data hall. This will include UPS power to secondary CHW pumps, CRAC units and associated building management system (BMS) control panels as well as the IT load itself to allow the data centre to be progressively shut down in a planned manner. The CHW buffer vessel ideally should be sized to
match the UPS autonomy in the event of mains and/ or generator failure. Clearly, having a UPS 10-minute autonomy makes no sense if the cooling autonomy
can’t match it, and vice versa. Some inertia in the data centre building fabric may also be taken into account. Thermal modelling predictions show that about two minutes after complete cessation of cooling delivery, air temperatures inside the data hall will rise to levels above the recommended server rack operating temperatures. This time delay will vary depending on the thermal mass of the building fabric – such as concrete floors/ceilings – exposed to the data centre space. For design purposes, thermal inertia should be ignored. An excess of primary CHW flow rate over the
secondary flow rate of about 5% is required to maintain a constant CHW reserve in the buffer vessel and avoid depletion. A further enhancement of CHW buffer vessel temperature control can be achieved by selectively diverting CHW to the top or bottom of the buffer vessel. On chiller start-up (warm) CHW can be delivered to the top of the buffer vessel, while (cold) CHW can be delivered to the bottom of the vessel. This is simply achieved by automatic valve arrangements on the CHW primary flow pipework leaving the chiller. As an indicator of buffer vessel size, a cooling load
of, say 1,000 kW would require a 25 cu m buffer vessel to match a UPS autonomy of 10 minutes, calculated as follows:
UPS Autonomy = CHW buffer vessel autonomy = 10 x 60 = 600s (ignoring thermal inertia) Assuming CHW F&R DT = 6K CHW specific heat capacity Cp = 4 kJ/kg K (allowing for say 10% ethylene glycol additive, which reduces Cp from its normal value of 4.2 kJ/kg K) CHW mass flow rate m, is given by 1,000/(4 x 6) = 41.7 kg/s To ensure continuous operation for 600s, requires a CHW reserve of 41.7 x 600 = 25,020 litres
A CHW buffer vessel used as a cooling reserve
should always be installed in the vertical configuration so as to enable proper CHW stratification temperature management by the BMS in the data centre.
Other solutions Alternative design solutions for high-density rack configurations are mainly variations on the same theme. These include an enlarged rack dimension that effectively houses the hot aisle (plenum) and cold aisle (plenum) within the rack itself. In this case the system boundary is not the room but the rack enclosure. Cold air is delivered by in-rack cooling modules or local cooling panels (LCPs) into a cold plenum. It then moves through the server rack, where it gains heat, and then delivers this to the hot plenum, from where it is returned to the LCP for cooling. This arrangement offers the advantage of close proximity precision-targeted cooling without any risk of undesirable thermal bypass leaks – a scenario that is more difficult to achieve where the room is the system boundary. This is because the quality control in a
58 CIBSE Journal November 2010
www.cibsejournal.com
>
Shutterstock
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80