UPS – let’s make it modular

07 March 2024

Stuart Dealing, sales, service and project execution leader, ABB Power Protection

When choosing a power protection scheme for a data centre, the first thing to consider is the total cost of ownership (TCO).

Operators typically want a 20-year lifetime for a UPS nowadays. So, investing in a more energy efficient model will result in significant overall cost and emissions savings, especially with energy prices surging. Similarly, specifying a UPS with better quality components, that only need replacing once rather than two or three times over the lifetime of the UPS, drastically reduces TCO.

Some data centres still rely on traditional monolithic UPS configurations – where all the components are built into individual blocks. So, the capacity cannot be adjusted without replacing the entire system or adding another UPS.

Alternatively, a modular UPS consists of multiple smaller, swappable, autonomous modules that work together within a single frame. For example, a 1MW modular UPS can be built from four parallel 250kW modules.

While the upfront cost of a monolithic UPS is often lower than a modular one of the same capacity, the TCO and availability benefits of going modular almost always outweigh a monolithic setup.

This is because modular UPS have more in-built redundancy. Operators can simply specify an additional module rather than an entire monolithic UPS block. For example, to achieve redundancy on a 1MW capacity data centre, you can save 30% costs by specifying a 1.5MW modular UPS frame to achieve N+2 redundancy compared with N+N from two 1MW monolithic blocks. This can also produce a lighter UPS system that can more easily be installed on the raised floors or building’s roof, for example.

In addition, a modular setup reduces the need for additional power equipment, like cables and switchgear. In the above case, two parallel monolithic UPSs would have two input switches and two output switches, where a modular setup would only have one on the input and one on the output.

Furthermore, modular UPS systems are more energy efficient. For example, if load demand is low, modern modular setups can automatically switch modules to sleep or economy mode to reduce energy consumption.

Operators managing 30+MW data centres are deploying medium-voltage (MV) equipment. The most obvious advantage of an MV UPS is that it offers lower TCO for several reasons.

Since the MV structures are able to cater for hundreds of megawatts, alternative power protection schemes for the entire facility can be proposed with ease, leading to lower capital expenditure.

The lower currents at the MV level enable smaller cable sizes, in turn reducing installation costs and space, allowing for the UPS to be placed further away from the load and maximize the given land plot.

Switching to MV also lowers overall footprint compared to using multiple LV UPS. Not just because of the size of the UPS itself, but also because less electrical equipment is required in the energy centre. Not only is it easier to maintain fewer UPS, but it also frees up space for server racks, maximizing available revenue from the same footprint. Since an MV system can handle the UPS function of the whole data centre, rather than selected server racks, it can often sit nearer the grid connection rather than in the server room. This saves more space and protects the entire facility through resilient architectures, such as distributed redundant and similar.

They are also more energy efficient than LV UPS as they suffer lower electrical losses because they run at lower currents. Compared to a rotary UPS, an MV UPS can save up to 4.2GWh of energy. This is equivalent to 1,245 tonnes of CO2 emissions over a 15-year lifetime.

According to the Uptime Institute’s Annual Outages Analysis 2023, 70% of events causing data centre downtime cost $100,000 or more, with 25% costing more than $1 million. With downtime costs this high, operators are looking for ways to ensure data continuity and reliability.

Remote monitoring supports this through a proactive approach to maintenance that means UPS issues can be addressed before they result in faults or downtime, heavily reducing TCO.

Moreover, remote monitoring avoids unnecessary emergency site visits from service professionals – reducing costs and the carbon emissions associated with their travel. In fact, it’s possible to troubleshoot UPS alarms remotely without the need for a call-out, which is especially valuable for more inaccessible data centres.

Data centres often hold very sensitive data, in which case, operators can opt for a remote monitoring system with one-way communication coming from the UPS, with no signals or commands going back to the UPS – it’s more of a ‘read-only’ connection that doesn’t allow for remote access but provides alarms.

The most important priorities when specifying a UPS are efficiency, footprint, and weight - modular and MV solutions meet all these criteria. With these strategies, data centres will significantly reduce their overall costs, minimize the risk of outages, and gear themselves for future growth.