Decision making around the type and scope of an uninterruptible power source (UPS) may at first seem overwhelming. However, asking yourself the following six questions will help you start the process to determine the UPS solution that best meets your needs:

1. Do you need three-phase feed?

If you're protecting three-phase equipment – generally in an industrial or business setting with larger critical loads – then you'll need three-phase UPSs. If the project involves single-phase equipment such as heating, lighting, smaller IT systems or devices run from a standard mains plug, you can opt for either a single- or three-phase UPS solution.

2. How much power do you require? 

To calculate your total load, list all the equipment you're going to protect with the UPS and combine their power requirements. You may also want to take into account the need for any additional load capacity you may need in the future – for example were you to add more server racks in your data centre later on. Depending on your budget, you can build in a safety margin to go beyond the minimum UPS capacity required to handle the maximum potential load.

3. Do you want a centralised UPS or distributed architecture?

While centralised architectures power the complete building or area, de-centralised UPSs are used to protect only a single load or piece of equipment such as a machine or rack. A centralised UPS system can be built resiliently by adding redundancy. This can be done by adding power modules in modular systems and/or more UPSs in parallel. This approach also enables the servicing of equipment one power module or UPS at a time, without the need to shut the system down completely. The same applies to scaling up the system – you can add power by adding more power modules or UPSs in parallel. 

UPS systems require status monitoring through an aggregated dashboard view that can be enabled using Eaton or third-party systems. They can also feature sensors for environmental monitoring that show how the UPS is behaving relative to local conditions including air temperature, water leakage, humidity etc. Ensuring appropriate cybersecurity is especially important to prevent any 'back door' system entry given the data flows generated by monitoring processes, as well as ensuring business continuity. 

Other considerations should you opt for a distributed architecture include the service package required to cover multiple systems – which may involve additional complexity and servicing time etc. 

A range of environmental factors will influence the design and materials used for UPS hardware. Is the UPS correctly designed and certified for marine or other harsh environments? Or is it appropriately constructed and rated for power protection in an industrial setting? 

4. What are your runtime requirements on batteries or other energy storage technologies? 

The starting point here is determining the storage capacity you need to deliver the runtime required for controlled shutdown or transfer etc – which, in general, can be delivered using VRLA, Lithium-Ion or supercapacitor technology. However, your choice here may be influenced by location and other practicalities. For example, the higher capex for using Lithium-Ion would be easily offset from a Total Cost of Ownership perspective in a remote location such as an offshore wind farm, where more frequent replacement of cheaper VRLA batteries would be expensive, as well as challenging. 

A further factor here is Energy Aware grid support – whether you wish to be able to generate revenue by selling unused stored power back into the grid on a 'demand response' basis, also supporting frequency harmonisation.

5. What are your redundancy needs?

Depending on the criticality of the load you are protecting with a UPS, you might want to add more redundancy to make the installation even more fault tolerant. You can do this throughout the power chain, starting by using dual feeds at the UPS's input or even dual feeds for the actual load (e.g. server). 
 
Looking at the UPS equipment specifically, redundancy can be built in either by paralleling multiple UPSs or by using modular UPSs. It all comes down to the question of how critical your load is – and whether you can afford downtime in the event of a failure in the power chain. Ultimately, it’s a balance between uptime and the CAPEX needed at installation.

6. Would you like to leverage data for further automation beyond the UPS and protection it provides? 

The power chain status data derived from a UPS can help you protect the IT infrastructure that's running underneath it – integrating both to maximise business continuity. For example, if the UPS knows that it's running on batteries with say 75% capacity remaining, IPM software can trigger actions at the IT load level. This might then involve moving virtual resources from one rack to another – or from one site to elsewhere – to keep the application running. 

Smart power distribution units (PDUs) come with very sophisticated sensors and a range of different power outlets. So if, for example, you're renting rack space to a tenant in a data centre on a monthly basis, you could bill them based on PDU power consumption data. 

Integrating with third-party monitoring applications is made easy through the use of public protocols such as SNMP/MQTT..

With cybersecurity a dynamic, constantly changing picture, choosing a supplier that delivers firmware updates is key to helping you stay ahead of the threat.