Are resilience and sustainability within data centre design mutually exclusive?

Data Centre Management – Winter Edition 2021

In recent years the data centre market has laid the groundwork for a greener, more resilient growth path despite a surging upwards trajectory in demand for data and the inevitable increase in speed of deployment that accompanies it. Many data centre clients are now asking for contractual commitments from their data centre service providers to minimise environmental impact and improve sustainability, whilst meeting their growing need for resilient and reliable server space. But are resilience and sustainability within data centre design mutually exclusive, or can operators achieve the best of both worlds for the sake of their clients and the environment?

One recent study confirmed that while data centres’ computing output jumped six-fold from 2010 to 2018, data centre energy consumption rose by only six percent.¹ Whilst this is a clear demonstration of the sustainable approach to growth now widely adopted in the marketplace, let’s not forget that the total installed base of Internet of Things (IoT) connected devices is projected to amount to 30.9 billion units by 2025, a sharp jump from the 13.8 billion units that are expected in 2021.² Statistics like this indicate that data centre operators must continue to find increasingly sustainable ways to meet demand if they are to remain competitive in a rapidly evolving and highly competitive marketplace.

With cooling and ventilation systems being so integral to resilience (offering a minimum resilience of N+1 or even N+N on state-of-the-art cooling technologies), whilst typically accounting for 25-45% of energy consumed by a data centre, it is an obvious place to start when trying to deliver both sustainability and resilience.

Not only can the right cooling equipment maintain a reliable and stable internal data centre environment but selecting cooling plant can make the difference between an energy efficiency ratio (EER) as low as 4 or as high as 100+.

Delivering resilience and sustainability with accurate cooling plant specification

There are a number of variables to consider when selecting cooling plant that could transform how a data centre delivers resilience and sustainability:

Geographical location and climate

Warmer climates, especially those with high humidity, are still more likely to rely on mechanical cooling and, as a result, will have a lower EER. Locations with a cooler climate can take advantage of ambient based cooling equipment such as dry coolers and free cooling that can have a dramatic and positive impact on efficiency with higher EER and lower power usage effectiveness (PUE).

Internal cooling plant

CRACs, CRAHs, Fan Walls, RDHx, and immersion cooling systems all have different cooling water temperature requirements, and this will determine the required performance of your external cooling infrastructure.

IT equipment

Despite the presence of published generalised temperature and humidity bands for IT equipment, many manufacturers are happy to allow equipment to run outside of them for periods of the year. For example, if a data centre service level agreement (SLA) stipulates that an internal air temperature of 25°C is required year-round, it is likely that mechanical cooling will be required to achieve this in the height of summer. But if the IT equipment can operate safely at air temperatures up to 4°C higher than this for 150 hours per year, then it may remove the need for mechanical cooling, reducing costs and energy consumption.

Maintain performance within operational bands

Coolers

The operational bands on fluid flowrate are narrow. If flowrate is too high the tubes in the heat exchange coil will erode and if it is too low, the velocity through the tubes will go into laminar flow and performance will reduce dramatically.

Pumps

Pump efficiency reduces quickly when operated below 60% so it may make sense to have a greater number of smaller pumps rather than one large single pump. However, stray too far from the suggested bands and you can end up with cavitation and early pump failure.

Case in point

Looking at a 1,000kW adiabatic cooler with design conditions of cooling water/glycol from 35°C down to 30°C, with a 35°C dry bulb and 20°C wet bulb, we can put all of the above points into a real-world situation:

Tim Bound, Director at Transtherm Cooling Industries, comments:

“Statistics often only cover point of use data, and it would be disingenuous not to point out the obvious increased cost to the environment when manufacturing additional standby equipment.  Fortunately, 95%+ of the materials used in the construction of our products are readily recycled at the end of a typical 20-year lifespan and, crucially, point of use energy consumption can be significantly reduced when operating all run and standby equipment simultaneously.

Since all of our cooling equipment is offered with the latest variable speed EC fans, we can take advantage of the cube rule energy saving law, which states that the power consumption of a fan is relative to the cube of its speed. Therefore, a fan running at 50% speed has a power consumption of 12.5% of its maximum. When run and standby coolers are operated together maximum fan power consumption can be reduced.

Furthermore, because the hydraulic pressure drop in the coolers and pipe work is subject to a square law – pressure drop reduces by the square of the reduction in flow, so when flow rate is halved, the pressure drop is reduced to a quarter. System pumps have a much lower pressure drop to overcome when total design flow rate is spread across all coolers, helping to reduce pump power consumption as well.”