The combined capacity of data centres across Australia and New Zealand stands at around 1.2GW . The majority of this footprint is located in and around Sydney. In comparison to other established data centre markets, growth in Australia and New Zealand is high. Data centre capacity shows an annual growth rate of around 20% from 2023 but the profile is changing. Enterprise on-prem data centres continue to decline to an estimated 170 MW by 2025 (at an average annual decline of around 3.0% from 2020) [W.Media Research estimates]. Growth in cloud and colocation facilities will more than compensate for this and reach a projected total of 950 MW in 2025 at an average annual growth of around 16% from 2020 .
Growth in New Zealand will also be high, doubling to around 350 MW by 2029 as the country moves into digitalisation, repatriation of its overseas data and infrastructure upgrades.
These trends are likely to continue as both Australia and New Zealand remain well-connected and stable host nations for data, offering high quality data centres and professional standards. AI will increase the requirement for data centres possibly by a factor of two although there are many different estimates of the likely impact of AI demand.
Data centres in Australia and New Zealand face the same challenges as those in most other places around the world – maintaining availability, meeting the challenges of accelerating advances in IT and networking, security, as well as keeping the costs of operation under control. Both Australia and New Zealand additionally face the challenge of being reasonably expensive markets for key resources – property, people, power, water and equipment. Location itself may also create supply problems – the COVID pandemic indicated the difficulties that can emerge in markets well away from where equipment is manufactured and which are smaller than the very largest global clusters of data centres.
The additional demand and supply pressures of the past few years have not so much created new stress points inside the data centre ecosystem as ramped up the pressure on existing ones. Increasing demand for services, the drive to renewable power sources and the potential for interaction between data centre power storage and the grid has changed the profile of power sourcing, distribution and protection. Higher IT densities and lower rates of data centre vacancy have also increased the pressures on cooling. The importance of cooling is emphasised by statistics that indicate that the failure of cooling/heat removal plays a role in between 13% and 20% of unplanned downtime and that cooling/heat removal can account for as much as 40% of data centre power use and for the costs that go with that.
Cooling has evolved to meet increasing workloads in a number of ways – for example, the evolution of precision cooling targeted to the points where heat removal is required rather than generally cooling a whole room or a whole rack. Cooling has also moved away from using pressurised air towards the use of liquid as a more effective means of cooling – water offers 4 times the cooling capacity of air and 25 times the heat transfer capability. The costs of cooling have led to an interest in ‘natural (free)’ cooling, usually where air from outside can be used to cool the IT equipment.
The demand growth described above is continuing and accelerating. Data increase is exponential and while advances in power and cooling technologies can flatten this rate of increase in terms the data centre environment they cannot totally compensate for it. Most improvements in data centre facility equipment have been concerned to improve an existing design rather than to rethink it. It is possible now that data centre cooling requires a strategic rethink.
Strategic thinking combined with the technological capabilities of the MHI Group have led to the development of the MHI Thermal System to ensure high-efficiency operation under a wide variety of operating conditions. The MHI Thermal System enables the integrated control of heating and cooling source systems via a centrifugal chiller once the Ene-Conductor integrated heat source control system together with a Centrifugal chiller have been installed.
The compressors, which form the heart of the chiller units, use the same technologies as in the design of commercial gas turbines. Aerodynamic analysis technologies improve the flow inside impellers and downstream of outlets to improve efficiency across a wide range of capacities. One- and two-stage vanes that adjust impeller shape and optimum refrigerant flow rate are used to efficiently compress refrigerant gases regardless of the operational point. These systems continually employ the latest analysis technologies to evolve to solution so that high performance can be demonstrated under all conditions.
MHI contends that “as data centres get older, there are limits to the PUE improvements possible through operation of existing equipment alone”. As the solution to this situation, MHI suggests that “it is important to improve the efficiency of heat source systems, which account for the largest amount of electric power in air conditioning systems”.
The efficiency of the MHI Thermal System is enhanced through the monitoring of the operating status of the chillers, and modifying the parameter settings from time to time. By having the chiller communicate with the Ene-Conductor in real-time, it is possible to automatically configure operating parameters that achieve energy efficiency based on chiller temperature settings changes around the year. By supporting the customization of high-speed chiller restarts, cooling water pools installed for the supply of cooling water at stable temperatures can be minimised. Through a high-performance microcontroller board and sensors installed within a chiller, the start-up load can be instantly tracked, enabling cooling water to be supplied at a stable temperature.