08 April 2016

Alessandro Nalbone, Product Manager Data & IT, AEG Power Solutions

Data centres are a demanding environment for electrical power. They require the highest standards of reliability and availability, but there is constant pressure to reduce power consumption – both for cost reasons, and to reduce the demand on foot print and cooling systems. 

To achieve high availability, it is essential that data centres, as well as smaller server or data rooms, have suitable Uninterruptible Power Supply (UPS) systems in place. These systems are able to ensure secure power in all the cases of a bad quality line, especially in case of mains failure and provide electrical power from back-up batteries long enough to bridge the gap until the mains systems are restored, or power can be provided from a generator or other source, or to allow the secure shut down of critical loads. 

The UPS also ensures clean power for critical applications, and can handle incidents where the input power waveform is of low quality – such as under-voltage, over-voltage, input frequency fluctuations, harmonics, spikes or power surges. 

How can a UPS system meet the demands for ultra-high reliability, while keeping efficiency high? What other factors are important in UPS for today’s data centres, in particular with the flexibility required by the move to cloud? 

Modularity for scalability
A key approach to meeting the needs of data centres is the adoption of modular architectures. While a traditional ‘monolithic’ UPS is often still the best choice for the smallest applications, such as a small server room, the modular approach is typically the right choice for medium size servers or small networks, right up to the biggest data centres. 

A modular UPS provides protection in ‘power blocks’, which might typically be around 30 kVA, enabling protection to scale up to the hundreds of kVA in a single system. 

A modular system provides the flexibility and scalability to respond to changes in the load that requires protection. This means that companies can ‘pay as they grow’, instead of needing to invest upfront in a system that will be initially over-specified – helping to reduce capital expenditure. Similarly, if the requirement for the protected load decreases, the blocks can be removed, and then redeployed elsewhere in a company’s infrastructure, or simply “frozen”, and kept in stand-by mode inside the UPS frame, until needed by the load protected. 

This flexibility is particularly useful for data centres providing cloud capabilities. The load can vary unpredictably from day to day, month to month, so a scalable UPS can cost-effectively meet demand. 

Modularity is also a way to improve efficiency
As well as scalability, the key benefit of a modular system is how it can dramatically improve efficiency. This reduces power consumption, and hence leads to a reduction in total cost of ownership – an important factor, considering how a high percentage of data centre costs typically comes from electricity usage. 

When using redundant systems, in an N+1 configuration, a UPS may often be only protecting light loads that make up a relatively low percentage of its capacity. When a non-modular UPS is only running at 30% or 40% of load, its efficiency will tend to drop dramatically. 

To overcome this, UPS manufacturers may implement predictive management. This means that, in a modular system, the blocks will be automatically switched on and off to respond in real-time to the demands of the load on the system. The UPS may typically only maintain in an active state the power modules that are specifically necessary for the actual critical load – as well as the ‘+1’ redundant power module. 

For example, the new Protect Plus M600 UPS from AEG Power Solutions (AEG PS) has the Idle Mode function, which allows it to automatically freeze the power modules that are not currently in use to protect the load – helping it reach efficiency up to 95.5% in double conversion (VFI). To balance out usage, and hence improve reliability, the system places the power modules into Idle Mode in rotation, and periodically puts the module with the longest running time to sleep and wakes up another module to replace it. Moreover, the system can operate in ECO Mode, not involving the double conversion line, thus reaching AC/AC efficiency values up to 98.5%. 

Even with light loads, this helps improve the efficiency of the UPS – in the case of the Protect Plus M600, even at a low 40% load, efficiency is almost the same as for a full nominal load. By combining the best transformer-less IGBT technology with this flat efficiency curve, the Protect Plus M600 reduces power losses, which directly translates into financial savings. 

In operation, the controller will shut down the modules one by one every 5 minutes, until the calculated number of modules remain active in the system, and when a set load is applied the controller wakes up all the modules so protection is immediate, before recalculating requirements and putting modules back to sleep if possible. 

Another reason why the UPS may run at a low load figure is to ensure excellent resilience via redundancy. At the highest level of data centre reliability, Tier 4, expected availability needs to be 99.995% or better. To help achieve this, each UPS system will be run at a maximum of 50% loading, and paired with another system – if either fails, the remaining UPS can pick up the other’s load, and still remain below 100% loading. This means that having a UPS that runs efficiently at low loads, of below 50%, is a big advantage in reducing power consumption. 

Pure modular and block modular
The modularity we have described so far is commonly called ‘pure modular’. Systems such as AEG Protect Plus M600 have relatively small modular blocks, with a rating of 30 kVA in this case, that can be combined to support the required load. 

For larger applications, another approach may be better: the ‘block modular’ architecture, such as AEG Protect Blue. This uses larger power blocks of 250 kVA that can be combined to reach the required load capacity. 

While it retains most of the flexibility of the pure modular approach, and also the efficiency, using a block modular UPS can be a more cost-effective route for large data centres. Once the required capacity gets somewhere near 500 kVA, it’s usually worth considering block modular as an alternative. 

User-friendliness and monitoring
The actual UPS product is, of course, only one part of what is needed. When choosing a UPS, you also need to consider the user interface, any management software, and the service and support you can receive from the vendor. 

Ease of use is important. If the UPS is difficult to configure, it won’t necessarily be used in the right configuration, which can impact efficiency and reliability. In the case of AEG Protect Plus M600, we ensure that each UPS module can be monitored through the centralised LCD touch-screen display, but also individually just by opening the door of the system, with an independent LCD display for each module. We also provide the ability to monitor the system remotely via communication with common protocols.   

AEG PS also offers MoniUPS, an innovative tool that constantly monitors a customer’s entire installation. With MoniUPS, the UPS can communicate a warning to AEG PS’ technicians, who can use its remote monitoring capabilities to diagnose the problem, and then reach the site installation and restore the UPS before a failure can occur. This ensures higher availability and reliability values.

In the end, for Data Centres it always comes down to reliability and availability which are top priority when choosing a UPS.

Modular and block modular approach to UPS enables highly reliable operation, it provides the flexibility needed for the long term, and it delivers excellent efficiency.

 AEG Power Solutions’ Protect Plus M600
AEG PS will be exhibiting at Data Centre World in London, 12-13 April 2016, on Stand D64 http://www.datacentreworld.com
www.aegps.com