Uninterruptable Power Supply (UPS)

This section deals with ‘Static’ UPS systems which use batteries as their storage medium. Rotary UPS systems will be covered within another section (at a later date). Likewise in-rack UPS systems will not be covered here.

Design Watchpoints:

  • Batteries are heavy – confirmation from structural engineer?
  • Selectivity of protective devices
  • Consider the bypass arrangement and how maintenance will be performed
  • Consider the UPS neutral conductor.
  • Is there cooling within the UPS and battery rooms?
  • Can it be moved into it’s final position (i.e. will it fit in lift or through doors?)
  • There phase in and out, or single phase out?

Purpose

A UPS system is designed to provide an immediate alternative source of power to important equipment during mains power failure. The first aspect to specifying a UPS is to have a discussion with the Client on it’s intended use:

  • What equipment is being supported?
  • How critical is it to business function?
  • Can it be taken down for maintenance?
  • Does the equipment have dual power inputs?
  • What are they looking to achieve?
  • The budget? (this usually puts the above items in perspective)

For data centres one way to assess the level of resilience and fault tolerance required is to refer to the ‘Tier’ classification by the Uptime Institute. While not a specific standard they do give specific guidance on what installations which have achieved their certification generally look like. This is usually a good start. Data Centre design will be a topic for another day.

Sizing a UPS system

To Size a UPS system you need two main pieces of information:

Battery Load – What is the anticipated load of the equipment the UPS system will support? Ensure power factor is taken into effect here as this is usually noted at 0.8.

Autonomy Time – How long should the equipment be supported for? Is it 10 minutes so the equipment can shut down, or is it 30 minutes so the equipment can operate through most power outages? Is there a generator to support the UPS for longer power outages?

Type of UPS

There are many types of UPS to consider, each with advantages / disadvantages so I’ll leave that as an exercise to the specifying Engineer. As a very simplistic overview:

‘Off-line’ – Equipment uses ‘raw’ mains until it fails, then it switches over to battery power. Efficient but messy. The switching of the mains is fast (less than 10 ms), however this may still effect critical loads, combined with no power conditioning means they are not normally used for this purpose.

‘On-line’ – Mains is routed through battery rectifier so equipment sees ‘conditioned’ power with no breaks. Clean but inefficient – best for highly-critical loads.

‘Line interactive’ – Similar to ‘off-line’ in that the mains is used to power the equipment, but with added line conditioning to condition the input power. Clean but has switching so not normally used for highly critical loads.

Battery Types

The other item often overlooked is battery technology. Lead acid batteries, through most commonly used, are not made for frequent use. If, for example, you wished to use a UPS to support a building in a part of the world with a temperamental power grid, then a lead acid battery would be a bad idea. These batteries should only be discharged once every 3 months otherwise the lifespan will be reduced. The correct choice in this scenario would be lithium batteries as they can be charged / discharged more often. They are more expensive through.

Racking – Batteries for smaller UPS systems are often located within cabinets, however once you get past a certain size then racks are recommended. Racks shouldn’t be more than 2 metres high with batteries accessible and not more than 2 deep. Racks should be made of powder coated steel and earthed regularly.

Fault Clearance and Selectivity

A UPS system doesn’t have an inherent ability to clear faults. This is because they run from a battery which doesn’t have an infinite capability to supply instant power at the levels provided to trip the breaker.

Under supply normal conditions if there is a fault the internal static switch in the UPS will quickly reach it’s limit and switch over – allowing the normal supply to clear the fault.

In the event of a fault occurring during power failure (i.e. running on UPS), the UPS cannot switch to the alternate source and therefore must be able to feed the fault until the downstream breaker disconnects. A typical UPS can only supply 150% of its rated duty for a short time, whereas a fault can be much higher, so this needs to be taken into account.

A type B MCB will trip instantaneously when a current between 3-5 times the full load current flows through it. Worst case we can choose 5x. Therefore to clear a fault on a 20A MCB we need 100A. At 150% overload this roughly translates into a minimum 50kVA (72 Amps TP + 150% = 108 Amps) UPS. For 32A MCB we would need a 80kVA UPS.

The calculation above only refers to the UPS inverter size. Generally the batteries can be sized for the anticipated load so these don’t need to match. Ensure you consult the manufacturer as there is also a lower end to this arrangement as too few batteries won’t have the peak power output to clear a fault.

Maintenance Bypass

Regular maintenance of the UPS will likely be done by switching over to the internal bypass. From there you can maintain the batteries etc. More extensive maintenance will need the UPS isolated entirely. This is where a complete bypass is required. Careful consideration is required here however as dropping the load is not usually an option .

UPS Bypass panel and Isolation Transformer

Above is a typical detail of a 3 breaker bypass panel. The principles are:

Step 1 – The UPS is switched over into internal bypass by the static switch

Step 2 – The middle breaker is closed to parallel the UPS with the mains. This breaker will not close to parallel the mains with the UPS until the UPS has switched on to it’s internal bypass. This is why the control interlock is required.

Step 3 – The input and output breakers are then opened to allow the UPS to be fully isolated.

Room requirements

Mechanical cooling – Lead acid batteries operate best around 20 degrees C. Lower temperatures reduce the performance, higher temperatures reduce the lifespan of the batteries. Often N+1 or N+N on the cooling system will be specified too.

Ventilation – Lead acid batteries when charging give off hydrogen. Even VRLA (Valve Regulated Lead Acid) batteries give off some hydrogen. Ventilation is required to ensure this doesn’t build up to a harmful concentration. Normally this will be small (e.g. a trickle vent on the door) however a calculation will be required.

Fire rating – consider what the UPS is supporting and how this / adjacent areas would be affected in a fire. If the UPS has been utilised for life safety systems then this will be a key concern.

Eye wash & Spill kits – The contents of the batteries are harmful. Consider what happens if they leak / spill. An eye wash kit stuck to the wall is recommended.

Warning signage – When specifying a UPS consider not just the room signage, but also the electrical safety signage – e.g. “This device is fed from multiple sources”.

The general room makeup should also be considered, such as Anti-static matting, painting the room, dust sealing etc

Structural conditions

Batteries are heavy. It seems quite obvious, however this is something which is frequently over-looked. Once you have an initial size of battery rack / UPS make sure this information is passed along to the structural engineer. This could effect where you put this equipment.

The second part to this is that battery racks have a higher point load as the weight is spread out on to smaller feet. Spreader beams may be required to distribute the weight.

Neutral Earth Transformer

A neutral earth transformer is required if the UPS transformer-less and has the possibility of having the neutral switched.

An example of where this can occur is when a UPS is fed from multiple sources (e.g. mains and generator) and there is a transition between two supplies utilising 4-pole breakers. Depending on how this switching takes place (e.g. overlapping neutrals or open-transition), or even if the supply itself is a 4 pole device, then there is the possibility the UPS could be left without a neutral – earth reference. This could leave the UPS effectively ‘floating’.

It should be noted that a neutral earth transformer is not strictly required in all scenarios. Losing a reference for a second will not mean the phase to neutral voltage will go spiralling out of control and short term breaks can usually be accommodated.

If this UPS is to be paralleled with the mains supply (such as in the example above utilising the maintenance bypass), Regular transformers have a phase shift as the electricity is passed through them so a zero-sequence (delta zig-zag) transformer will be required.

Careful consideration should be taken on where to position this transformer so it’s effective in all scenarios. In the example above the transformer is positioned on the output of the UPS. This was chosen specifically in-case the internal input of the UPS was isolated it prevents the UPS running with no neutral – earth reference.

Use of UPS for life safety applications

In many technical seminars UPS’s have been proposed for life safety equipment, however when it comes to actually specifying this I have always hit the following roadblocks. For a generator you specify how much fuel at full load is required. A three hour tank supporting a fire-fighting lift will typically run for 10 hours or more as the lift is often in-frequently used – thus the generator idles. When specifying a UPS, 3 hours at full load would result in acres of batteries at great expense.

So the question always comes back to “how often will the lift be used within that 3 hour window?” I don’t have that answer, and have never received that answer when asked. Thus a diesel generator is specified.

Let me know if you have had more successful projects and I’ll update here.